THE OXFORD HANDBOOK OF

PHILOSOPHICAL THEOLOGY
 THE OXFORD HANDBOOK OF

PHILOSOPHICAL THEOLOGY

Edited by
THOMAS P. FLINT
and
MICHAEL C. REA
 Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research,
scholarship,
and education by publishing worldwide in

Oxford New York

Athens Auckland Bangkok Bogotá Buenos Aires Cape Town

Chennai Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi
Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai
Nairobi
Paris São Paulo Shanghai Singapore Taipei Tokyo Toronto Warsaw
with associated companies in Berlin Ibadan

Oxford is a registered trade mark of Oxford University Press

in the UK and in certain other countries

Published in the United States

by Oxford University Press Inc., New York

© Oxford University Press, 2009

The moral rights of the authors have been asserted

Database right Oxford University Press (maker)

First published 2009

All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any
means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the
appropriate
reprographics rights organization. Enquiries concerning
reproduction
outside the scope of the above should be sent to the Rights
Department,
Oxford University Press, at the address above

You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer

British Library Cataloguing in Publication Data

Data available

Library of Congress Cataloging in Publication Data

The Oxford handbook of philosophical theology / edited by Thomas

P. Flint.
p. cm. - (Oxford handbooks in religion and theology)
Includes index.
ISBN 978–0–19–928920–2 (alk. paper)
1. Philosophical theology. I. Flint, Thomas P.
BT40.094 2008
210–dc22 2008036033
Typeset by SPI Publisher Services, Pondicherry, India
Printed in Great Britain
on acid-free paper by
CPI Antony Rowe

ISBN 978–0–19–928920–2

1 3 5 7 9 10 8 6 4 2
To Michael J. Loux
 CONTENTS

List of Contributors

Introduction
THOMAS P. FLINT AND MICHAEL C. REA

PART I THEOLOGICAL PROLEGOMENA

1. Authority of Scripture, Tradition, and the Church

RICHARD SWINBURNE

2. Revelation and Inspiration

STEPHEN T. DAVIS

3. Science and Religion

DEL RATZSCH

4. Theology and Mystery

WILLIAM J. WAINWRIGHT

PART II DIVINE ATTRIBUTES

5. Simplicity and Aseity

JEFFREY E. BROWER

6. Omniscience
EDWARD WIERENGA
 7. Divine Eternity
WILLIAM LANE CRAIG

8. Omnipotence
BRIAN LEFTOW

9. Omnipresence
HUD HUDSON

10. Moral Perfection

LAURA GARCIA

PART III GOD AND CREATION

11. Divine Action and Evolution

ROBIN COLLINS

12. Divine Providence

THOMAS P. FLINT

13. Petitionary Prayer

SCOTT A. DAVISON

14. Morality and Divine Authority

MARK C. MURPHY

15. The Problem of Evil

PAUL DRAPER

16. Theodicy
MICHAEL J. MURRAY
17. Skeptical Theism and the Problem of Evil
MICHAEL BERGMANN

PART IV TOPICS IN CHRISTIAN PHILOSOPHICAL THEOLOGY

18. The Trinity

MICHAEL C. REA

19. Original Sin and Atonement

OLIVER D. CRISP

20. The Incarnation

RICHARD CROSS

21. The Resurrection of the Body

TRENTON MERRICKS

22. Heaven and Hell

JERRY L. WALLS

23. The Eucharist: Real Presence and Real Absence

ALEXANDER R. PRUSS

PART V NON-CHRISTIAN PHILOSOPHICAL THEOLOGY

24. Jewish Philosophical Theology

DANIEL H. FRANK

25. Islamic Philosophical Theology

OLIVER LEAMAN
26. Chinese (Confucian) Philosophical Theology
JOHN H. BERTHRONG

Index
 LIST OF CONTRIBUTORS

Michael Bergmann is Professor of Philosophy at Purdue University.

John H. Berthrong is Associate Professor of Comparative Theology
at the Boston University School of Theology.
Je rey E. Brower is Associate Professor of Philosophy at Purdue
University.
Robin Collins is Professor of Philosophy at Messiah College.
William Lane Craig is Research Professor of Philosophy at Talbot
School of Theology.
Oliver D. Crisp is Lecturer in Theology at Bristol University.
Richard Cross is Reverend John A. O’Brien Professor of Philosophy
at the University of Notre Dame.
Stephen T. Davis is Russell K. Pitzer Professor of Philosophy at
Claremont McKenna College.
Scott A. Davison is Professor of Philosophy at Morehead State
University.
Paul Draper is Professor of Philosophy at Purdue University.
Thomas P. Flint is Professor of Philosophy at the University of
Notre Dame.
Daniel H. Frank is Professor of Philosophy at Purdue University.
Laura Garcia is Adjunct Assistant Professor of Philosophy at Boston
College.
Hud Hudson is Professor of Philosophy at Western Washington
University.
Oliver Leaman is Professor of Philosophy and Zantker Professor of
Judaic Studies at the University of Kentucky.
Brian Leftow is the Nolloth Professor of the Philosophy of the
Christian Religion at Oxford University.
Trenton Merricks is Cavaliers’ Distinguished Teaching Professor of
Philosophy at the University of Virginia.
Mark C. Murphy is Fr. Joseph T. Durkin, S. J. Professor of
Philosophy at Georgetown University.
Michael J. Murray is Arthur and Katherine Shadek Professor in the
Humanities and Philosophy at Franklin and Marshall College.
Alexander R. Pruss is Associate Professor of Philosophy at Baylor
University.
Michael C. Rea is Professor of Philosophy at the University of Notre
Dame.
Del Ratzsch is Professor of Philosophy at Calvin College.
Richard Swinburne is the Emeritus Nolloth Professor of the
Philosophy of the Christian Religion at Oxford University.
William J. Wainwright is Distinguished Professor Emeritus in the
Philosophy Department at the University of Wisconsin, Milwaukee.
Jerry L. Walls is Professor of Philosophy of Religion at Asbury
Theological Seminary.
Edward Wierenga is Professor of Religion and of Philosophy at the
University of Rochester.
 INTRODUCTION

THOMAS P. FLINT MICHAEL C. REA

THE rst half of the twentieth century was a dark time for
philosophical theology. Sharp divisions were developing among
philosophers over the proper aims and ambitions for philosophical
theorizing and the proper methods for approaching philosophical
problems. But many philosophers were united in thinking, for
di erent reasons, that the methods of philosophy are incapable of
putting us in touch with theoretically interesting truths about God.
To be sure, doubts of this sort never gained a sure foothold in
Catholic universities, which maintained the theological focus
evident from their founding. But, for a variety of reasons, the
scholasticism practiced in these institutions went on in virtual
isolation from the philosophical trends dominant at the great secular
universities of Europe and America. There, doubt reigned about the
possibility of fruitful interaction between philosophy and religion.
Since philosophical theology (as we understand it) is aimed
primarily at theoretical understanding of the nature and attributes
of God, and God’s relationship to the world and things in the world,
the prevailing skepticism about our ability to learn about God
through philosophical reasoning left philosophical theology on the
wane.

A bit of history is needed to understand the genesis of this

skepticism. It is common now to see the eld of academic
philosophy as divided broadly into two camps—‘analytic’ and
‘Continental’—and to locate the origins of the division somewhere
in the rst half of the twentieth century. The two camps elude
precise de nition and cannot plausibly be seen as encompassing all
philosophical work. Furthermore, it is misleading at best to treat
them either as wholly discrete from one another or as anything
more than very loosely uni ed within themselves. Still, at the risk of
dramatically oversimplifying, we o er the following
characterizations. The analytic tradition has, by and large, treated
philosophy as an explanatory enterprise aimed at analyzing
fundamental concepts (‘person’, ‘action’, ‘law’, etc.), and at using
this analytic method to clarify and extend the theoretical work
being done in the natural sciences. The Continental tradition, on the
other hand, has viewed philosophy as an autonomous discipline
aimed, more or less, at exploring and promoting our understanding
of the human condition in creative and decidedly non-scienti c (and
not even mostly explanatory-theoretical) ways.

As the division between the two camps was developing, the

aboriginal gures of the analytic tradition leaned strongly in an
empiricist direction. By and large, they thought, in the words of
Wilfrid Sellars, that ‘science is the measure of all things: of what is
that it is, and of what is not that it is not.’1 Indeed, the logical
empiricists— gures such as Otto Neurath, Moritz Schlick, Rudolf
Carnap, and, a bit later, A. J. Ayer—went so far as to say that
statements that do not admit of empirical veri cation (i.e.
statements where no observations would be su cient to determine
their truth or falsity) are entirely meaningless. Carnap likened
metaphysicians to ‘musicians without musical ability’:2 not only are
the noises they produce utterly devoid of propositional content, but
the noises don’t even manage to sound beautiful. Thus, since
philosophical theology is a mostly non-empirical enterprise, the
heyday of logical empiricism found philosophical theologians
beating a hasty retreat.3 Moreover, even those in the analytic
tradition who were unwilling to endorse anything resembling a
veri ability criterion of meaning were nevertheless very suspicious
of anything that looked like theory-building that wasn’t somehow
grounded in the natural sciences. In short, the analytic tradition
generally proved to be an inhospitable climate for religious
theorizing.4

Matters were not much better on the Continental side either.

Among Continental philosophers, at least two strains of thought
tended to choke out philosophical theology. Some were so gripped
by the transcendence of God that they came to think that God was
beyond all human categories, even Being itself. In the eyes of these
thinkers, the project of trying to arrive at theoretical understanding
of God is just hopeless. Indeed, more than hopeless—it might even
be idolatrous, since what we would inevitably end up talking about
in trying to discuss God in terms of human concepts would be a
‘simulacrum’ of human creation rather than God himself.5 On a
closely related note, some were gripped more by human limitations,
and came to despair of the possibility of arriving via philosophical
methods at general, universally valid theoretical understanding of
anything at all. The idea, roughly, was that our belief systems are so
inevitably tied to our own very limited perspectives—perspectives
conditioned by our biological make-up, our sociopolitical
circumstances, our own particular experiences in life, and the like—
that it is ridiculous (at best) to think that we might ever attain to
any kind of undistorted ‘absolute’ knowledge or understanding that
would be valid for all rational creatures from all points of view for
all of eternity.6 Given the prevalence of both these strands of
thought throughout the Continental tradition, it is no surprise that
philosophical theology did not ourish there either.

In the latter half of the twentieth century, however, there was a

great revival of interest in the philosophy of religion in general and,
in its wake, in philosophical theology in particular. It is common to
locate the origin of this revival in the publication of Alasdair
MacIntyre and Anthony Flew’s landmark anthology, New Essays in
Philosophical Theology.7 In that volume, the main issues on the table
were concerns about the meaningfulness of religious discourse and
questions about the rationality of religious belief. In other words,
the volume was oriented mainly toward objections to religious
belief and discourse arising primarily out of the analytic tradition.
These topics constituted a large proportion of the agenda for
subsequent work in philosophy of religion for the next two or three
decades.

In a recent essay, Nicholas Wolterstor has argued persuasively

that the revival just described was made possible within the analytic
tradition by three major developments.8 First, there was the death of
logical empiricism during the 1960s. Logical empiricism was
responsible for much of the anti-metaphysical bias within the
analytic tradition. Thus, with the death of logical empiricism came a
revival of interest in metaphysics more generally, and a
corresponding openness to the theoretical investigation of religious
topics.9

Moreover, according to Wolterstor , the demise of logical

empiricism also brought about a loss of interest in general questions
about the origins of our concepts and the limits of human thought
and judgement. This was the second major development. Whereas
the Continental tradition remained (like the modern period through
Kant, and like the logical empiricists) rather preoccupied with the
idea that human limitations might entirely close o certain avenues
of inquiry or render impossible meaningful thought or discourse
about certain kinds of topics, the analytic tradition seems to have
left such concerns behind, thereby opening the door even wider to
all sorts of metaphysical inquiry, theological and otherwise.
 Finally, the third development was the owering of meta-
epistemology—explicit re ection on and evaluation of alternative
theories of knowledge. One of the main developments within meta-
epistemology was the rejection of classical foundationalism (the
view, roughly, that a belief is justi ed only if it is indubitable,
incorrigible, evident to the senses, or deducible from beliefs that are
indubitable, incorrigible, or evident to the senses). Classical
foundationalism was an epistemological theory that many modern
philosophers implicitly took for granted; and it has been shown to
lie at the heart of objections against the rationality of religious
belief leveled by a variety of thinkers, including Hume, Freud, Marx,
W. K. Cli ord, and others. The collapse of classical foundationalism
made room for the ourishing of alternative epistemological
theories, including some that were much more friendly to the idea
that religious belief might be perfectly rational.10

In sum, then, the analytic tradition seems to have moved beyond

several important biases that placed obstacles in the way of the
growth of philosophical theology. As Wolterstor notes, however,
the same sort of thing has not happened within the Continental
tradition. Thus, though there has surely been some measure of
philosophical theology done within that tradition, the eld of
philosophical theology has been dominated by gures writing
within the analytic tradition. This fact goes a long way toward
explaining why the present volume is oriented in that direction as
well.
 We said earlier that the agenda set for philosophers of religion
for a couple of decades placed heavy emphasis on discussion of the
epistemology of religious belief and the meaningfulness of religious
discourse. There was also quite a bit of discussion of the divine
attributes (omnipotence, omniscience, perfect goodness, and the
like), of traditional arguments for the existence of God, and of the
most widely discussed argument against the existence of God—
namely, the problem of evil. Over the past twenty years, however,
philosophers of religion have begun to focus more of their attention
on theological doctrines apart from those concerning the nature,
rationality, and meaningfulness of theistic belief. Thus, for example,
a great deal of attention has been devoted recently to philosophical
problems arising out of the Christian doctrines of the Trinity, the
Incarnation, and the Atonement; there has been an explosion of
work on questions about the nature of divine providence and its
implications for human freedom; and a fair bit of recent work has
also been done on questions about the metaphysical possibility of
the resurrection of the dead. Other topics are still ripe for
discussion. For example, there is a (relatively) very small literature
on the topic of divine revelation and the inspiration of Scripture,
only a handful of works on the topics of prayer, original sin, and the
nature of heaven and hell, and virtually nothing on the Christian
doctrine of the Eucharist.

In the present handbook we have tried to provide articles

covering most of the above topics. However, we have tried to avoid
covering topics that have already been discussed in the Oxford
Handbook of Philosophy of Religion. The most notable exceptions are
in Parts II and III where we include chapters on omniscience,
omnipotence, moral perfection, and the problem of evil. We include
these topics for two reasons. First, we believe that readers of a
handbook in philosophical theology would quite naturally expect to
see these sorts of issues covered. Second, and perhaps more
importantly, we believe that each of these topics deserves more
extended and detailed treatment than it could sensibly have
received in a more general philosophy of religion handbook. To take
just one example: there is a vast literature on the problem of evil;
but the Oxford Handbook of Philosophy of Religion devotes (sensibly)
only one chapter to that issue. For our more narrow purposes here,
however, we have seen t to include distinct chapters on (a) the
di erent versions and instances of the problem of evil (e.g. the
logical problem, the evidential problem, the problem of divine
hiddenness, and perhaps others), (b) the so-called ‘skeptical theist’
strategy for responding to the problem of evil, and (c) questions
about the prospects for ‘theodicy’ (i.e. a response to the problem of
evil that o ers a complete story about why God in fact permits evil).
Similar reasoning explains our decisions to include the few other
topics in this handbook that have already received coverage in the
Oxford Handbook of Philosophy of Religion.

The chapters that follow are divided into ve parts covering ve

general topics:
 I. Theological Prolegomena

II. Divine Attributes

III. God and Creation

IV. Topics in Christian Philosophical Theology

V. Non-Christian Philosophical Theology

The chapters in the rst part treat questions about the authority
of scripture, tradition, and the church; the nature and mechanisms
of divine revelation; and the nature of theology. We also include a
chapter on theology and mystery, a topic that we think has not yet
received its due in the analytic tradition.

Chapters in the second part focus on philosophical problems

connected with the central divine attributes: aseity, omnipotence,
omniscience, and the like. These have traditionally been among the
most widely discussed topics in philosophical theology.

In the third part, we take up questions about God’s relationship

to creation. The chapters in this part explore theories of divine
action and divine providence; questions about the purpose and
e cacy of petitionary prayer; problems about divine authority and
God’s relationship to morality and moral standards; and, nally,
various formulations of and responses to the problem of evil.

In the fourth part, we turn to topics in speci cally Christian

philosophy. In recent years there has been a surge of interest in
philosophical problems that arise in connection with the Christian
doctrines of the Trinity, the Incarnation, the Atonement, original
sin, and resurrection. Other topics, the doctrine of Christ’s real
presence in the Eucharist, or the nature of heaven, are ripe for
exploration but have not, as yet, received their due. Part IV provides
coverage of all these issues.

Finally, in the fth part, we have included three chapters on

non-Christian philosophical theology—Jewish, Islamic, and
Confucian. The vast majority of philosophers (in the English-
speaking world, anyway) devoting their attention to philosophical
theology tend to focus either on distinctively Christian doctrines or
on topics that are common to all the theistic traditions. Thus,
relatively few pay much, if any, attention to topics in philosophical
theology that fall squarely outside the Christian tradition. Since the
target audience of our handbook comprises English-speaking
philosophers of religion mostly in the analytic tradition, it is
appropriate that our handbook emphasize, as it does, topics that
have occupied and will probably continue to occupy center stage in
the major journals in the philosophy of religion. But we believe that
it is both important and valuable more widely to expose precisely
that group of philosophers to the work that is being done in
philosophical theology outside the Christian tradition. Providing
such exposure is the goal of our fth part.11

NOTES
1. Wilfrid Sellars, ‘Empiricism and the Philosophy of Mind’, in his
Science, Perception, and Reality (London: Routledge & Kegan
Paul), 127-96, at 173.

2. From ‘The Elimination of Metaphysics Through Logical

Analysis of Language’, English translation in S. Sarkar (ed.),
Logical Empiricism at its Peak: Schlick, Carnap, and Neurath
(London: Routledge, 1996), 30.

3. We say only ‘mostly’ for two reasons. First, not everything that
would count as philosophical theology is non-empirical. G. K.
Chesterton famously remarked that the doctrine of original sin
is the one doctrine of Christianity that admits of direct
empirical veri cation. Or, more seriously, consider e.g.
Richard Swinburne’s defense of belief in the resurrection of
Jesus in The Resurrection of God Incarnate (Oxford: Oxford
University Press, 2003). Second, because even outside the
aforementioned Catholic institutions, a number of scholars
kept alive (though more in the popular than in the academic
realm) the great perennial questions in philosophical theology,
C. S. Lewis being the most prominent example.

4. Fuller versions of the brief account we have given here of the

relationship between philosophical theology and analytic
philosophy in the mid-twentieth century and before can be
found in Alvin Plantinga’s ‘Advice to Christian Philosophers’,
Faith and Philosophy 1 (1984): 253–71 and, more recently,
Nicholas Wolterstor ’s ‘How Philosophical Theology became
 Possible within the Analytic Tradition of Philosophy’, in Oliver
D. Crisp and Michael Rea (eds.), Analytic Theology: New Essays
in the Philosophy of Theology (Oxford: Oxford University Press,
2009).

5. Heidegger is one locus of such thinking; though here, we are

e ectively reading Heidegger through the eyes of such
contemporary philosophers as Merold Westphal and Jean-Luc
Marion. See e.g. Westphal’s ‘Appropriating Post-Modernism’,
ARC, The Journal of the Faculty of Religious Studies, McGill
University 25 (1997): 73-84, and ‘Overcoming Onto-theology’,
in J. D. Caputo and M. J. Scanlon (eds.), God, The Gift, and
Postmodernism (Bloomington: Indiana University Press, 1999),
146-69, both reprinted in Westphal, Overcoming Onto-Theology:
Toward a Postmodern Christian Faith (New York: Fordham
University Press, 2001), and Marion’s God Without Being, trans.
Thomas A. Carlson (Chicago: University of Chicago Press,
1991); ‘Metaphysics and Phenomenology: A Relief for
Theology’, trans. Thomas A. Carlson, Critical Inquiry 20 (1994):
572-59; and ‘The Idea of God’, in D. Garber and M. Ayers
(eds.), The Cambridge History of Seventeenth-Century Philosophy
(Cambridge: Cambridge University Press, 1998), i. 265-304.
This strand of thinking has also in uenced twentieth-century
systematic theology (which, historically, has signi cantly
overlapped what we are calling philosophical theology). See
 e.g. Paul Tillich, Systematic Theology (Chicago: University of
Chicago Press, 1951), i., esp. 235 .

6. For fuller articulation and development of this line of thinking,

see K. J. Vanhoozer (ed.), The Cambridge Companion to
Postmodern Theology (Cambridge: Cambridge University Press,
2003), M. Westphal, ‘Taking Plantinga Seriously: Advice to
Christian Philosophers’, Faith and Philosophy 16/2 (1999);
‘Appropriating Postmodernism’; and ‘Father Abraham and his
Feuding Sons’, in Overcoming Onto-Theology: Toward a
Postmodern Christian Faith (New York: Fordham University
Press, 2001). A very useful collection of relevant primary
sources is L. E. Cahoone (ed.), From Modernism to
Postmodernism, 2nd edn. (Oxford: Wiley-Blackwell, 2003).

7. London: SCM, 1955. Again, cf. Plantinga, ‘Advice to Christian

Philosophers’.

8. ‘How Philosophical Theology Became Possible within the

Analytic Tradition of Philosophy’.

9. For a discussion of this revival, see the Introduction to Michael

J. Loux and Dean W. Zimmerman (eds.), The Oxford Handbook
of Metaphysics (Oxford: Oxford University Press, 2001).

10. Notoriously, postmodern philosophers have drawn very

di erent lessons from the collapse of classical foundationalism.
On this, see e.g. Richard Rorty, Philosophy and the Mirror of
Nature (Princeton: Princeton University Press 1979), or, more
 recently and with speci cally theological applications,
Vanhoozer (ed.), Cambridge Companion to Postmodern Theology,
and Stanley Grenz and John Franke, Beyond Foundationalism
(Louisville: Westminster/John Knox, 2001).

11. The editors would like to thank Claire Brown for her excellent
work on the index to this volume. Her endeavours were
supported by a grant from the Institute for Scholarship in the
Liberal Arts, College of Arts and Letters, University of Notre
Dame.
 PART I

THEOLOGICAL PROLEGOMENA
 CHAPTER 1

AUTHORITY OF SCRIPTURE, TRADITION,

AND THE CHURCH

RICHARD SWINBURNE

CHRISTIANITY, Islam, and Judaism all claim that God has given
humans a revelation. Divine revelation may be either of God, or by
God of propositional truth. Traditionally Christianity has claimed
that the Christian revelation has involved both of these. God
revealed himself in his acts in history; for example in the miracles
by which he preserved the people of ancient Israel, and above all by
becoming incarnate (that is human) as Jesus Christ, who was
cruci ed and rose from the dead. And God also revealed to us
propositional truths by the teaching of Jesus and his church. Some
modern theologians have denied that Christianity involves any
propositional revelation, but there can be little doubt that from the
second century (and in my view from the rst century) until the
eighteenth century, Christians and non-Christians were virtually
unanimous in supposing that it claimed to have such a revelation,
and so it is worthwhile investigating its traditional claim. It is in any
case very hard to see how it would be of great use to us for God to
reveal himself in history (e.g. in the Exodus, or in the life, death,
and resurrection of Jesus) unless we could understand the cosmic
signi cance of what happened—e.g. that Jesus was God incarnate
and that his life and death constituted an atonement for our sins.
And how are we to know that unless with the history God provides
its interpretation?

My concern in this chapter is with the Christian claim to have a

propositional revelation. In the rst section I shall describe the
process by which Christians of past centuries have come to believe
that certain propositions have been revealed. (My account of this
process is a summary of the far fuller account given in my book
Revelation,1 where it is backed up by far more historical references
than there is space to provide here.) Then in the second part of the
chapter I shall assess alternative philosophical accounts of what
constitutes a belief that such-and-such propositions have been
revealed, being a ‘justi ed’ belief (or a ‘warranted’ or ‘rational’ one).

Although almost all Christians (until the eighteenth century) agreed

that God has revealed propositional truths to various individuals
and groups at various times, they also agreed that there was a major
public revelation through the teaching of Jesus and his apostles,
which included his (quali ed) authentication of a revelation by God
to ancient Israel. From the earliest days of the Christian church,
Christians recognized what we now call the Old Testament as
revealed truth—despite one major denial of this (by Marcion—see
later), and despite di erences about exactly which books belonged
to the Old Testament. By the middle of the second century, AD most
Christians also agreed that there were sacred books that contained
the record of the new revelation given by Jesus and his apostles;
and that these books included the four Gospels, the Acts of the
Apostles, and most of the letters attributed to St Paul in today’s
Bibles. They disagreed, however, for some further centuries about
which other books should be included in what we now call the New
Testament. In his book The Canon of the New Testament Bruce
Metzger2 analyses the three criteria that led church bodies to
recognize some book as New Testament Scripture—its conformity
with basic Christian tradition, its apostolicity (being written by an
apostle, or someone closely connected with an apostle), and its
widespread acceptance by the church at large. So a prior
understanding of the general nature of the Christian revelation (of
central Christian doctrines), evidence of historical reliability, and
church authority operated to determine which books constituted the
Bible, Old and New Testaments together.

Christians have disagreed about whether there was more to the

‘deposit of faith’, the record of God’s revelation, than the Bible. For
Protestants the ‘deposit of faith’ just is the Bible, and many Catholics
and Orthodox today also hold this view. But in the past, as also
today, many Catholics and Orthodox hold that the deposit of faith
also includes ‘unwritten traditions’, traditions of the apostolic
church (that is the church in the years immediately after its
foundation) which were not written down at that time in any book.
These traditions have been thought to include not merely traditions
as to how to worship (e.g. how to celebrate the Eucharist), but
doctrinal truths some of which might be implicit in the rules for
worship.3

Nevertheless, despite these di erences about the extent of the

deposit of faith, Christians agreed that taken together the biblical
books constituted God’s ‘word’. While the New Testament seems to
make a fairly modest claim for the authority of the Old Testament,
that ‘all scripture is inspired by God and pro table for teaching, for
reproof, for correction and for training in righteousness’ (2 Timothy
3: 16), much stronger claims soon came to be made for the
authority of Scripture. Gregory the Great wrote that Holy Scripture
is ‘a letter of God Almighty to his creature’.4 Theologians did not
wish to deny that these books were written by di erent human
authors at di erent times, as was evident by the fact that so often
there are references to ‘I’ who did certain things, which can only be
construed as a reference to such an author (e.g. St Paul), and by the
discrepancies of style of the di erent books. But theologians claimed
that these books were divinely inspired, and divinely authorized—
that is, God was the ultimate author who authorized these books to
be published as his revelation.

The main disagreements between Christians concern the criteria

for determining the meaning of biblical passages and for deriving
doctrines from them; and (partly in consequence of the former
disagreement) about what various passages mean and which
doctrines can be derived. The disagreements about the meaning of
biblical passages often turn on which ones should be taken in their
most natural literal sense, which should be taken in a possible literal
sense although not the most natural one, and which should be
understood metaphorically. It might seem that this is something to
be sorted out by careful historico-literary investigation; and
Protestant scholars of the last four centuries have had this ideal of
discovering by normal methods of historical inquiry the ‘original
meaning’ of the text, as rst written down by a human author.
Hermeneutics consisted in discovering this original meaning; the
text with this meaning—although originally directed at ancient
Israelites and early Christians—was also ‘God’s word’ to all humans
ever.

This project, however, soon shows that, so understood, the Bible

contains many obvious inconsistencies, major and minor, between
and within biblical books. The most obvious of these is the apparent
inconsistency between Israelite leaders (apparently on God’s behalf)
commending the extermination of the Canaanites, men, women, and
children, and burning their cities by re, as described in the Book of
Judges; and Jesus’s (apparently on God’s behalf) commending non-
violence in the Gospels. There are also many biblical passages taken
in the sense most probably intended by their human authors that are
manifestly inconsistent with established conclusions of science and
history. If we treat the book of Genesis and the subsequent
‘historical’ books with the dates and numbers of years recorded
there as literally true history, we can deduce that the period
between the creation of Adam and the birth of Jesus lasted
something of the order of 4,000 years. So it seems to be claiming
that the human race began to exist about 4,000 BC. Yet archaeology
has shown us human civilizations much older than that. Genesis
records a ood that covered the whole earth; but while geology
certainly provides evidence of extensive oods, there is not the
slightest record of a universal ood after the appearance of humans.
And so on.

These evident facts led many liberal-minded theologians of the

twentieth century to cease to talk of the Bible being ‘true’, but to
speak rather of it being ‘useful’ or ‘insightful’ if read in accord with
some rule or other of interpretation; and there have evolved as
many ways of interpreting as there have been theologians to do the
interpreting. And saying this sort of thing about the Bible hardly
gives it a special status—the same could be said of any great work
of literature. A general fog settled over ‘hermeneutics’.5

The church of the early centuries was, however, well aware of

these problems. It was well aware that if certain passages of the Old
Testament were understood in the most natural way, they
represented God as commending much immoral conduct, and
contained many false assertions about matters of science and
history. The biblical picture of a at earth covered by a dome was
inconsistent with the picture of the world provided by the
contemporary Greek science of a spherical earth, centred in the
middle of a hollow sphere that rotated daily around the earth. And
many of the educated theologians (the ‘Fathers’) believed what the
Greeks had to say about science. And while they didn’t have any
knowledge of archaeology and geology that would show the falsity
of the historical claims in Genesis and other biblical books taken in
their most natural senses, they still rejected some of them on the
theological grounds that they depicted too anthropomorphic a God.

The Marcionite controversy made theologians more acutely

aware of the need to develop a theory about how to interpret the
Bible. In the late second century Marcion, a priest at Rome
convinced of the immorality of much of the conduct apparently
commended in the Old Testament, campaigned for a Bible that had
no Old Testament, and a New Testament consisting merely of an
expurgated version of Luke’s Gospel and ten Letters of St Paul. The
orthodox reassertion of the Old Testament was led by St Irenaeus,
who claimed that much of the Old Testament had merely temporary
and symbolic signi cance. It was, if read in its most natural sense, a
record of the growth of understanding of God by the Israelites; and
also God’s revelation to Christians, if read metaphorically. At the
beginning of the third century Origen claimed that some of the Bible
(New Testament as well as Old) was not literally true, but all of it
contained deep metaphorical truth. This view, developed in the next
century by Gregory of Nyssa and accepted by Augustine, became a
standard view. Augustine’s basic rule was the same as that of Origen
and Gregory: ‘We must show the way to nd out whether a phrase is
literal or gurative. And the way is certainly as follows: whatever
there is in the word of God that cannot, when taken literally, be
referred either to purity of life or soundness of doctrine, you may set
down as metaphorical.’6

In other words, the way to nd out what the Bible meant was
not to use secular historical methods to discover the ‘original
meaning’ of the text as written of its human author. Since God was
the ultimate author of the Bible, it was what he meant the text to
mean that was its meaning as a revealed text. The Fathers did
normally suppose that the human author also understood this, but
they allowed that he might be writing something that he did not
understand. To understand what God meant by the written word,
we need to take into account God’s beliefs. We know, the Fathers
claimed, two kinds of beliefs that God has—God believes central
Christian doctrines, and God believes all truths about history and
science. (Secular science and history, when very well established,
provides us with strong evidence about God’s beliefs about the
latter.) So if anything in Scripture seems to contradict these beliefs,
they argued, it must be understood in a metaphorical way. This
result follows from normal rules for understanding a sentence of
ordinary human discourse. If someone utters some sentence which,
taken literally, they know to be false, and they know that (much of)
their audience knows to be false, then they cannot mean it to be
taken literally. If I say to others about our human friend Larry,
‘Larry is an elephant’, this must be understood metaphorically. And
so if God is the ultimate author of Scripture and its intended
audience is not merely its original hearers and readers, but the
church of many future centuries, then it must be interpreted in the
light not merely of Christian doctrine, but of future scienti c
knowledge. But the Fathers would have said, if such a possibility
had occurred to them, that in the case of a con ict as to how to
interpret a Scriptural passage between an interpretation guided by
their prior understanding of central Christian doctrines and an
interpretation guided by current established ‘scienti c knowledge’,
that the former took precedence. For God had revealed central
Christian doctrines to his church before ever the New Testament
was written; purported scienti c knowledge was fallible. But how
most scriptural passages should be interpreted was not regarded as
xed by already recognized central Christian doctrines.

One verse that all the Fathers refused to take literally on the
ground of its incompatibility with Christian moral teaching was
Psalm 137: 9 which says of the Babylon which held the Jews in
captivity, ‘Happy shall they be who take your little ones and dash
them against the rock.’ A verse from the Psalms which some of the
Fathers refused to take literally on scienti c grounds was Psalm 136:
6 which gives thanks to God ‘who spread out the earth on the
waters’ on the grounds that science showed that earth did not oat
on water.7 And in a wide sense of ‘science’ it was on scienti c
grounds that some of the Fathers rejected the literal understanding
of the Genesis 1 claim that God made the world in six days. For
there could not, they claimed, be days before the sun existed, and
Genesis claimed that the sun was created on the fourth day.
Augustine’s basic rule deriving from Origen was applied throughout
the Middle Ages. Beryl Smalley has written that ‘to write a history
of Origenist in uence on the West would be tantamount to writing a
history of Western exegesis’,8 and she could have added Eastern
exegesis as well.

When any sentence is to be interpreted metaphorically, the

normal (secular) rules for interpreting it are: take the words of the
passage (apart from words that link it to its context) in their literal
senses. Consider the objects or properties normally designated by
these words, and objects or properties commonly associated with
them. Interpret the words as designating some of these latter objects
or properties instead. Take as the true interpretation that which
interprets the words as designating objects or properties closely
rather than remotely connected with their normal designata, in so
far as this can be done in a way which makes the sentence
appropriate to the context. In so far as there is more than one
equally plausible interpretation, the passage is to be read as having
both (or all) of these interpretations. Origen recognized three
di erent kinds of sense possessed by most passages of Scripture, and
Augustine recognized four.9 The Christian Fathers had available to
them a whole set of objects or properties commonly associated with
the people, places, and actions referred to in the Old Testament,
which provided symbolic meanings for the words that normally
designated the latter. ‘Moses’ can refer to Jesus, since Jesus was the
new Moses who was leading his people out of slavery to sin, in the
way that Moses led them out of slavery to the Egyptians. And
‘Joshua’ can also refer to Jesus, since Jesus was the New Joshua (the
Hebrew word for ‘Jesus’) who was leading his people to his
kingdom, as Joshua led them to the promised land. Jesus Christ is
also the ‘rock’ to which we can cling successfully when storms
attack us. And since the Jews became enslaved in Babylon, ‘Babylon’
comes to represent evil generally.

Hence the blessing confessed in Psalm 137: 9 on those who take

the ‘little ones’ of Babylon and dash them against the rock was
naturally interpreted as a blessing on those who take the o spring
of evil (Babylon), which are our evil inclinations, and destroy them
through the power of Christ (‘the rock’).10 And the command to the
Israelites invading Canaan to exterminate the Canaanites, and some
details of the invasion, may be interpreted in the same spirit.11

When you begin to interpret passages metaphorically, con icting

interpretations multiply, some of which are more plausible than
others. To my mind the metaphorical interpretation nearest to the
literal under which Genesis 1’s talk of ‘days’ comes out as plausibly
true, and so the preferred metaphorical interpretation, is derived by
treating ‘days’ as long periods of time and the exact order of
creation as not relevant to the truth of the chapter. Then the passage
tells us that gradually God brought about the various facets of
creation over long periods of time, no doubt through ‘secondary
causes’ (i.e. the normal operation of scienti c laws that God
sustains), as Genesis 2 suggests. Augustine also took the passage
metaphorically, but his interpretation is by comparison a very far-
fetched one. He claimed that all the things described in Genesis 1 as
created by God were created simultaneously, and that talk about
‘days’ is to be interpreted as talk about stages in the knowledge of
creation possessed by the angels.12 But ‘days’ are more like billennia
than they are like logical stages in the growth of knowledge. My
metaphorical interpretation is, therefore, by normal criteria for
metaphorical interpretation, better than his.

Augustine’s De Genesi ad Litteram contains so many di erent

interpretations of passages (most but not all of them literal
interpretations, but not perhaps the most natural literal
interpretations) that he faced the objection: what have you achieved
except to show that we cannot know what these passages of
Scripture mean? His answer was: I have shown that Genesis is
compatible with whatever physical science might show.13 But he
would have found it natural to add: so long as what physical science
shows is not incompatible with already recognized central Christian
doctrines.

As I have just noted, a crucial constraint on the interpretation of

Scripture was that it should be interpreted consistently with the
church’s teaching. All the Fathers were well aware that without this
constraint, Scripture was capable of a thousand di erent (and often
incompatible) interpretations. But ‘every word’ of Scripture ‘shall
seem consistent’ to someone, wrote Irenaeus, ‘if he for his part
diligently read the Scriptures in company with those who are
presbyters in the church, among whom is the apostolic doctrine’.14
And Tertullian comments that disputes between orthodox and
heretics could not be settled by appeal to Scripture, since the
meaning of Scripture was uncertain; the teaching of the church must
rst be identi ed and that will determine how Scripture should be
interpreted.15 The Reformation claim that Scripture should be
interpreted without any external constraint constituted an enormous
innovation in Christian thought. Because of the innumerable
di erent ways in which Scripture can be interpreted, the
Westminster Confession’s assertion that ‘the infallible rule of
interpretation of Scripture is the Scripture itself’16 seems to me quite
hopeless. The unavoidability for the above reason of a ‘theological
perspective’ for the interpretation of biblical texts is developed at
some length in Jorge Gracia’s book How Can We Know What God
Means?17

But all Christians shared another view, which seems opposed to

what I have just described—that all doctrine must be derived from
Scripture (perhaps with the addition of unwritten traditions). Some
doctrines of the Creed (e.g. the doctrines that Jesus ‘was cruci ed …
buried …and rose up on the third day’) simply repeat Scripture,
which—it was agreed—must be understood literally unless there is
reason to deny the literal sense. But the doctrines of the Trinity and
the Incarnation as stated in the Creed and as de ned by the
Ecumenical Councils of the fourth and fth centuries clearly do not
repeat Scripture. Nor can they be deduced from Scripture without
the aid of some de nitions of the philosophical terms that occur in
the Creed, such as (substance), (individual),
(nature), and so on. Even so, they can be rendered consistent with
certain scriptural passages only if those passages are read in a none
too natural way. And while everyone accepted that Scripture should
be read in a way constrained by established doctrine, that was of no
help when it was still disputed what the doctrine was. And that was
the case in the Arian controversy. The Catholics said that Jesus
Christ was (of the same substance) as the Father, in other
words fully divine; and the Arians said that Jesus Christ was
(of similar substance) to the Father, in other words semi-
divine. Both parties could quote biblical passages in their support. St
John’s Gospel, speaking of Jesus, claims that ‘the word was God’.18
But it also quotes Jesus as saying ‘The Father is greater than I’,19
and how could any being be greater than God? And various other
quotations could be assembled on each side.

What was at issue was surely which position ts best the overall
pattern of the teaching of the New Testament, whose content was
itself determined by a prior understanding of the nature of central
Christian doctrines. And the inference from scriptural passages to a
doctrinal de nition that ts best their overall pattern is not a
deductive inference. It is a probabilistic inference of a kind used by
historians of thought in their systematizations of the work of some
thinker which t most of what he is alleged to have a rmed (taken
in its most natural sense) better than any other simple account, but
which may not t all of what he is alleged to have a rmed. The
historian then deals with any such recalcitrant sentences in one of
three ways. First, he may claim that the thinker must have
understood the recalcitrant sentences in a way other than a normal
way. Secondly, he may claim that the thinker did not write these
sentences—the record is inaccurate. Or thirdly, the historian may
claim that his suggested principles t almost everything the thinker
wrote so well, that the thinker would have come to deny the
statements expressed by the recalcitrant sentences if their
incompatibility with his general viewpoint was pointed out to him.
But when dealing with the Bible, believed by con icting interpreters
to be a true record of the teaching of God, only the rst way is
available. A sentence, such as the saying of Jesus ‘the Father is
greater than I’, incompatible with one such systematization if
understood in the most natural way, was interpreted in a less
natural way than the most natural way—for example as reporting
Jesus saying that the Father was greater than Jesus in so far as Jesus
was human (that is, in his human nature).

But of course it is often none too obvious which systematization

of biblical passages ts the overall message of the text best. And so
the church used a further criterion for judging between plausible
competing interpretations: a proposed doctrinal de nition must be
compatible with the church’s own past decisions (even if the issue
was still disputed). Newman characterized the derivation of a new
doctrinal de nition as involving the perception of its ‘congruity’
with Scripture and with the ‘intuitive sense of the Church’.20 There
were competing theories about how the church could reach such a
binding conclusion—by the decision of an Ecumenical Council (itself
recognized as ‘Ecumenical’ by its subsequent widespread acceptance
in the church), by the decision of a council rati ed by the Pope, or
even by the Pope acting on his own. But the suggestion that
Christians could start again to discover the meaning of biblical
passages without being constrained by the way those passages had
been interpreted in the past is wildly out of line with the tradition of
Christianity’s rst 1,400 years.

And how was the church that had this vital role in establishing
true doctrine to be identi ed? While certainly the fact that some
ecclesial body gave the most plausible overall interpretation of the
‘deposit of faith’ was used as a criterion for determining the identity
of the church and so for which new doctrine should be adopted, the
criterion of that ecclesial body’s past approval of some plausibly
derived doctrine could not be used as the sole test for the identity of
the church without hopeless circularity. There was a further
criterion recognized by all. An ecclesial body was the church in so
far as it had continuity of organization with the apostolic church:
that is, in so far as its church leaders were commissioned by other
church leaders and so on back to the apostles. Again there were
di erences about the kind of commissioning required. (Did bishops
need to be ordained by bishops? Was the recognition of a bishop’s
authority by the Pope required for that authority to be legitimate?)
But all agreed that some body was only part of the church in so far
as it had some continuity with the apostolic church.

So, to summarize this historical section, for Christianity’s rst

1,400 years Christians have recognized as revealed truths those
properly derived from Scripture (and perhaps also unwritten
tradition). Scripture consisted of books recognized as such by the
church. The meaning of scriptural passages was determined in part
by the church’s prior teaching, and the proper derivation of
doctrines from Scripture was also determined in part by the church’s
recognition of that derivation. Despite all the many di erences of
doctrine and organization between ecclesial bodies, it is a signi cant
fact that for the thousand years between the ninth and nineteenth
centuries virtually all ecclesial bodies (Catholic, Orthodox,
Protestant, and Oriental Orthodox—that is monophysites and
‘Nestorians’) with any case for continuity of organization with the
apostolic church were committed to the Nicene Creed as containing
the central doctrines of Christianity. (By the ninth century Arian
opposition to the Creed had virtually died out, while after the
nineteenth century some ecclesial bodies began to adopt liberal
interpretations of credal doctrines very di erent from earlier
interpretations.) If the church is the source of our knowledge of a
revelation from God through Jesus, at the very least the doctrines of
the Nicene Creed must be true. And without the church’s selection
and interpretation of biblical texts we would have no knowledge of
the content of that revelation.

II

Christian doctrines about what God is like and how he has acted
have come to us through the church’s tradition of interpreting
Scripture, and without the historical process I have described in
section I there would be no Christian doctrines to assess. But why
should we believe these doctrines to be true? No doubt in the case
of some of them there are considerations of pure reason that make it
to some extent probable that the doctrines are true. For example we
might expect a God who saw us su ering to become incarnate
(become a human) in order to show solidarity with us by sharing
our su ering. But such a priori reasoning seldom seems conclusive;
and it will not in any case tell us where and when any divine actions
were done, for example in which human God became incarnate. God
needs to reveal these truths to us. So the question remains why, if at
all, we are justi ed in believing that Scripture interpreted in a
certain way is God’s revelation. For anyone’s belief in Christian
doctrines depends in large measure on such a belief.

Modern philosophical theories of what makes a belief ‘justi ed’

(or ‘rational’ or ‘warranted’) divide into internalist and externalist
theories. There is no space in this chapter to discuss adequately the
merits of these two kinds of theory, or of all their consequences for
the justi cation of Christian doctrines. But it will be appropriate to
outline and comment on examples of recent attempts to apply
theories of one or the other kind to assessing the justi cation for
believing that God has revealed truths through Jesus Christ that are
contained in Scripture, and for regarding some particular
interpretation of Scripture as the correct one. The most common
kind of internalist view holds that a belief is justi ed if (and only if)
it is rendered probable by all a believer’s basic beliefs taken
together, basic beliefs being ones that seem obviously true to the
believer but not for the reason that they are rendered probable by
other beliefs. They include beliefs about what the believer seems to
be perceiving, or clearly remembers, or what other people tell him
that they have perceived, or that everyone knowledgeable agrees
about. If you have a basic belief that you have observed so-and-so,
but also a basic belief that renders the former belief improbable (for
example that someone else claims that they were at the relevant
place at the relevant time and observed that so-and-so did not
happen, or that they saw you at a di erent place at the relevant
time), then you must discover what is made most probable by the
balance of all your basic beliefs taken together, each basic belief
being weighted by the strength of your initial degree of conviction
that it is true. Many of your basic beliefs will be ones shared by
most other people, and so constitute public evidence. Virtually no
basic belief can be regarded as infallible; but some few basic beliefs,
for example beliefs about what you were doing two minutes ago, are
so strong (that is, seem so obviously true), as to deserve to be
believed despite almost any conceivable counter-evidence. My own
view is an internalist view of this kind.

It follows from this internalist view that if you have had an

overwhelming religious experience on reading Scripture that God is
‘three persons of one substance’ you are justi ed in believing it. So
too if everyone tells you that everyone who knows anything about
the subject agrees that everything written in Scripture is true and
that Scripture says that God is three persons of one substance.
Calvin wrote that ‘Scripture exhibits fully as clear evidence of its
own truth as white and black things do of their colour, or sweet and
bitter things do of their taste,’21 and many Protestant theologians
and confessions have said much the same: that Scripture wears its
truth on its face. And since Calvin also held that it was in general
obvious (at any rate to those who were both educated and
unprejudiced) what Scripture meant—we did not need church
councils to tell us this—he held that the truths of Christian doctrine
could be read from the pages of Scripture.

But in my view in the twenty- rst century, while quite a number

of people have religious experiences of the presence of God,
relatively few people reading Scripture have experiences of the truth
of Christian doctrines; and those that do often have di erent
understandings of those doctrines. But it must very seldom be the
case that any individual has had an experience of this kind so strong
that it needs no further support from publicly available evidence.
And in the twenty- rst century there must be very few communities
in which everyone assures you and no one denies that everything
written in Scripture (understood in a certain way) is true. Among
the basic beliefs of all of us will be basic beliefs that many people
who read Scripture nd it highly implausible, and that many of
those who believe that it contains much truth interpret it in
di erent ways from ourselves. So we need to assess what is shown
by the balance of our basic beliefs, and—in order to get a belief as
probably true as we can get—we need to seek more relevant basic
beliefs; and that means primarily getting to know about the large
amount of relevant public evidence. Even those who have had some
relevant, though not overwhelmingly strong, religious experience
need also some public evidence for the truth of Christian doctrines if
their belief that those doctrines are true is to be justi ed in the
internalist sense of ‘justi ed’.

I have argued (e.g. in The Existence of God22) that there is

substantial general public evidence (in the form of the existence of
the universe, its conformity to simple general laws, these laws being
such as together with the boundary conditions of the universe to
give rise to human beings, these human beings being conscious, etc.
etc.) which makes it probable that there is a God (omnipotent,
omniscient, perfectly free, and perfectly good). I argued (e.g. in
Revelation) that God had good reason to become incarnate in order
to identify with our su ering, make atonement for our sins, and
reveal important truths to us about what God is like and how we
ought to live in order to enjoy life with God in Heaven forever. If he
did this, he would have to lead a life of a certain kind (e.g. a perfect
human life under di cult conditions in which he showed that he
believed that he was divine and was making atonement for our sins,
and founded a church which continued his work). He would also
need to have his actions and teaching culminated by a great miracle,
a setting aside of the laws of nature that God alone can do; this
miracle would constitute God’s signature of approval on those
actions and that teaching. All this background evidence makes it
quite probable that God would become incarnate in a prophet who
would live a life of a certain kind, which would be culminated by a
great miracle.

So we need only a modest amount of detailed historical evidence

in the form of witness testimony that one prophet did lead the right
kind of life that was culminated by a great miracle (together with
evidence that this was true of no other prophet) in order to make it
probable that that prophet was God incarnate. I argued further in
Revelation that that evidence is available; there is only one prophet
in human history about whom there is a modest amount of evidence
of a kind to be expected if he lived the right kind of life, and also
only one prophet about whom there is a modest amount of evidence
that his actions and teaching were culminated by a great miracle. It
is unlikely that there would be this combination of evidence unless
it was produced by God, who is most unlikely to have produced it
unless the prophet was indeed God incarnate. The unique prophet of
whom both these things were true was Jesus (the miracle
culminating his work about which there is evidence being his
resurrection). He founded a church to continue his work; and if God
revealed truths through Jesus, he must be providing that church
with continuing divine guidance to ensure that it teaches what he
taught and what is implicit in that. Otherwise God’s revelation
through Jesus would have been largely pointless, for except for a
few immediate disciples, no one would know what Jesus had taught.
Given that the teaching of Jesus and the resulting teaching of the
Christian church is not too improbable on other grounds (e.g. it does
not teach that there is nothing wrong with rape and pillage), then
the total evidence (the general public evidence for the existence of
God together with detailed historical evidence about Jesus) makes it
probable that Jesus was God incarnate who rose from the dead, and
so that God authenticated his teaching and the teaching that the
church derived from his teaching. (In this process the historical
evidence is con ned to that obtained by treating the New Testament
like any normal historical source.)

It follows from all this that the fact that a doctrine has been
developed by the church through the process of derivation from
Scripture described in section I, which the church recognized as the
proper process for this, is substantial reason for supposing that that
doctrine is true. So Scripture turns out to have an authority far
greater than that of a normal historical source. (I believe that the
apparatus of the calculus of probability can be used to elucidate the
structure of the above historical argument, and gives us reason to
suppose that that argument gives a high probability to the truth of
Christian doctrines. In Warranted Christian Belief23 Alvin Plantinga
produced an argument from ‘dwindling probabilities’ purporting to
show that a historical argument of the type I have produced will
give only a fairly low probability to Christian doctrines. In an
Appendix to the second edition of Revelation, I claim that Plantinga’s
argument based on his criticism of the rst edition of that book
misunderstands my argument, and is not cogent.)

An approach somewhat similar to mine, emphasizing the

cumulative e ect of di erent kinds of evidence in supporting what
he calls ‘canonical theism’ (theism as developed in the Christian
church) characterizes the work of William Abraham—see, most
recently, his Crossing the Threshold of Divine Revelation.24

For an externalist, however, a belief is ‘justi ed’ (or whatever) if

the belief is caused by a process of the right kind, whether or not
the believer is aware of what that process is or that it is of the right
kind. The simplest form of externalism is reliabilism: a belief is
justi ed if (and only if) it is produced by a reliable process, a
reliable process being one that produces mostly true beliefs. Thus it
may be said: my belief that I am perceiving my desk is justi ed if it
is produced by perception and perception is a generally reliable
process. There is an enormous problem (the ‘generality problem’)
about how the type of process to which the token process (the
particular process that operated on a particular occasion) belongs is
to be selected. For suppose that I am a retired professor who has
recently taken the drug LSD. Then is what makes my belief that I am
now perceiving a desk justi ed or unjusti ed, the fact that this
token process belongs to the type of perception by anyone, or the
type of perception by me, or the type of perception by a retired
professor, or the type of perception by a retired professor who has
recently taken LSD, or what? For the success rate of each of these
types will vary. Maybe perception as such is 90 per cent reliable,
whereas perception by retired professors who have recently taken
LSD is only 30 per cent reliable. According to the type to which the
token belief is referred, so di erent consequences follow for whether
the belief is justi ed or not. I do not think that pure reliabilism has
any principled solution to this problem.

But there are forms of externalism that incorporate a reliabilist

element that do have such a solution. The best-known form of
externalism that has been applied to assessing the ‘justi cation’ (or,
as he calls it, the ‘warrant’) of central Christian doctrines (what he
calls ‘the great things of the Gospel’) is Alvin Plantinga’s theory of
warrant. For Plantinga,25 a belief has warrant (subject to defeaters)
if and only if: (1) it is produced by cognitive faculties functioning
properly, (2) in a cognitive environment su ciently similar to that
for which the faculties were designed, (3) according to a design plan
aimed at the production of true beliefs, when (4) there is a high
statistical probability of such beliefs being true (that is, the process
of production is highly reliable). By someone’s cognitive faculties
‘functioning properly’, Plantinga understands them functioning ‘in
the way their creator (whoever or whatever that was) meant them
to function’. If God made us, our faculties function properly if they
function in the way God designed them to function; whereas if
Evolution (uncaused by God) made us (the main alternative that
Plantinga considers), then our faculties function properly if they
function in the way that Evolution (in some sense) designed them to
function. Now if God, who is an intentional agent, created us, there
will be clear content to all four criteria. God’s intentional actions
will have caused the operation of a particular type of process that
caused the token belief that there is a God. If he designed it to
operate in the environment in which it is currently operating so as
to lead to true beliefs, and if the process is reliable (which it will be
if God designed it for that purpose), the belief will be warranted.
The type to which a token belief should be referred to assess its
reliability is xed by God’s intention to produce true beliefs by a
process of that type. The normal type of process that according to
Plantinga God uses to bring about true beliefs in the ‘great things of
the Gospel’ is the operation of the Holy Spirit on our hearts
‘directly’, that is, in response to hearing these ‘great things’
preached or reading them in Scripture. He does however allow that
in some people the Holy Spirit might operate indirectly, making
them recognize the cogency of some argument for the truth of these
great things.

If, however, God (or any other intentional agent) did not create
us (and Plantinga is seeking an account of warrant that does not
presuppose that God created us), then there is a big problem about
how (1) is to be understood. For Evolution (or any other inanimate
cause) is not an agent who has intentions (despite the incautious
talk by some biologists about our organs having ‘design plans’). The
only sense I can give to claims that Evolution ‘meant’ something to
function in a certain way, or that it ‘designed’ it to function in a
certain environment or to produce certain beliefs, is that it caused it
to function in that way in that environment so as to produce those
beliefs. So cognitive faculties function properly if they function in
the way Evolution causes them to function; their design plan is
‘aimed’ at the production of true beliefs only in so far as it does
produce true beliefs in that environment; and an environment is
‘su ciently similar’ to that for which the faculties were ‘designed’
only in so far as they produce true beliefs equally well in that new
environment. And so the whole edi ce collapses on to (4): a belief is
warranted if it is produced by cognitive faculties in an environment
in which they produce true beliefs as frequently as in their original
environment in which they produced mostly true beliefs. This is a
form of simple reliabilism—and the normal form of reliabilism that
is open to the generality problem, for there is no creator’s ‘intention’
by which the type of ‘cognitive faculty’ and the ‘environment’ in
which it was originally reliable can be selected from the whole
range of other types of cognitive faculty and types of environment
to which the token process and the token environment in which it
originally operated belong. (Is the cognitive faculty ‘perception’ or
‘vision’? Is the environment the earth, or merely a particular human
community? And so on.)

Plantinga, however, thinks that there is a God, and thus that

Christian beliefs are warranted if they are produced by a God-
created process designed to produce true beliefs, and that will
normally be the operation of the Holy Spirit on our hearts producing
those beliefs, often in response to reading Scripture. Plantinga
allows that there could in principle be ‘defeaters’ to show that,
despite the initial warrant of the belief, it is not really warranted
after all. Such defeaters to the warrant of belief in Christian
doctrines might include the results of New Testament historical
scholarship, or the conclusions of postmodernist philosophy; but
Plantinga does not consider that any such defeaters are very strong.

William Alston has in the past also advocated a mainly

externalist theory—in his case a straightforwardly reliabilist theory
—of the justi cation of belief in general, and applied it to religious
belief (to include belief in Christian doctrines), in particular where
it is produced by religious experience resulting from what he calls
‘Christian mystical practice’.26 He claims that the process of
believing beliefs produced in this way is a reliable process (that is,
produces mostly true beliefs) and so makes us ‘justi ed’ in believing
any belief so produced (e.g. about the authority or interpretation of
Scripture). He allows that religious experience may need support (to
make it overall probable) from arguments from public evidence
(presumably also to be justi ed on a reliabilist account). Alston’s
theory is open to the generality problem common to all reliabilist
theories. Stephen Evans applies an externalism derived from
Plantinga and Alston to the justi cation of Christian doctrines in his
The Historical Christ and the Jesus of Faith.27 Evans emphasizes the
point that Plantinga allows as an occasional possibility that the Holy
Spirit may operate on humans to produce a justi ed belief that
Christian doctrines are true not in a direct way or in showing us that
they are well supported by public evidence. Both Evans and Alston
allow the possibility of defeaters to belief in Christian doctrines that
might need to be taken seriously. A more hardline externalism is put
forward by John Lamont in his Divine Faith.28 For Lamont, divine
‘faith’ is belief in the testimony of God spoken to us through the
teaching of the church; response to that testimony is the only way to
Christian faith. He holds that ‘a belief that is based on the motives
of credibility [that is, arguments from public evidence] is…not
Christian faith’; and once you have that faith, ‘it is irrational …to
take objections to faith seriously’.29

So for all these writers what makes a belief in the authority of

Scripture justi ed is that our belief in that authority and so in the
doctrines that Scripture contains is produced by the right God-
created process, and not (as such) by any argument from historical
evidence or even from some overwhelming experience of the truth
of that authority. (Such an argument or experience should be
trusted on an externalist view only if it was produced by the right
process.) These di erent writers have di erent views about which
Christian doctrines are contained in Scripture, and they must
maintain that the justi cation not merely of their belief in the
authority of Scripture but of their belief in their particular
interpretation of Scripture depends on how those beliefs were
produced. For the internalist, however, a believer’s interpretation is
justi ed in so far as it is rendered probable by evidence supporting
one interpretation against another, and that may include evidence of
the extent to which di erent interpretations have been
authenticated by the church.

Some of the writers whom I have been discussing use ‘faith’

interchangeably with ‘belief’; for Lamont, for example, Christian
faith is used to mean a belief that certain central Christian doctrines
are true. For others, Christian faith is the belief that these doctrines
are true accompanied by some a ective component (such as a love
of God or hatred of one’s sins), so long as it is produced in the right
way (e.g. by the operation of the Holy Spirit). But, as theses about
the kind of faith that on traditional Christian views is necessary for
salvation, these positions need argument of a kind not fully
developed in the writings I discuss. For there is a di erent view that
Christian faith is a matter of acting on the assumption that (or
trusting that) Christian doctrines are true, and perhaps also
believing that they are true or perhaps without that belief. In my
view the Christian tradition has had no clear view of the nature of
faith. This is partly because the same Greek word and the same
Latin word des can be translated either as ‘belief’ or as ‘trust’. A
distinction between ‘belief’ and ‘trust’ can only be brought out by
di erent ways in which the corresponding verb is used: (or
credo ut) means ‘I believe that, whereas Images (or credo in) or
the verb followed by a dative noun means ‘I trust in’. In this
context30 I am treating all theories merely as theories about what
makes a belief that Christian doctrines are true ‘justi ed’ (or
‘warranted’).

So—to return to the main theme—we all want to have true

beliefs and no false beliefs, on all matters and above all on religious
matters. We can seldom be absolutely certain that our beliefs
(especially in the twenty- rst century our beliefs about religious
matters) are true, but we can have beliefs that are probably true,
and we can look for and assess further evidence in order to get
beliefs that are more probably true than the ones we now have.
What people who want beliefs which are ‘warranted’ or ‘justi ed’
are, I suggest, normally looking for is beliefs that are probably true
—on the evidence. But we have access to no other evidence than
our own (though we can add to that). So in seeking such beliefs we
are seeking beliefs that are probably true on our evidence (that is,
our own basic beliefs taken together). It is more sensible to be
guided in our conduct by beliefs that are probably true than by ones
that are not probably true, for the reason that if we rely on the
former we are more likely to attain our goals (e.g. the forgiveness of
our sins, and making ourselves and our fellow humans tted for the
worship of God in Heaven) than if we rely on the latter. And the
more probably true our beliefs are, the more sensible it is to rely on
them. So it is clearly a good thing to have beliefs about Christian
doctrine that are justi ed in the sense of being rendered probable by
our evidence, that is, ‘justi ed’ in an internalist sense.

It might also be a good thing to have beliefs about Christian

doctrines that are ‘warranted’ in Plantinga’s or some similar
externalist sense. But it would be foolish to rely on such beliefs
unless we have reason to suppose that they are true. Of course if
they are warranted in a Plantinga-type sense, they will be true. But
again before we rely on them for this reason, we need reason to
suppose that they are appropriately warranted; and that means
having a belief that they are probably (on our evidence) thus
warranted. That is, before we rely on them for the stated reason, we
need an internally justi ed belief that our Christian beliefs are
caused in the way in which Plantinga claims that they are—that is,
by the ‘internal instigation of the Holy Spirit’ leading to our having
relevant true beliefs. That will involve, to begin with, an (internally)
justi ed belief that there is a God who produces some of our beliefs.
If there is a God, he is responsible in large measure (though humans
can in uence these processes) for the operation of all natural
processes including those which produce beliefs in us, and so he has
a considerable responsibility both for the operation of processes that
produce in some of us beliefs that Christian doctrines are false and
for the operation of processes that produce in others of us beliefs
that Christian doctrines are true. So in order to have a probably true
belief that our Christian beliefs have been produced by the right
process, we must see what our evidence indicates about this. And
the only evidence we can have that the beliefs have been produced
by the right process will be any evidence we have that makes it
probable that they are true. How else could one distinguish beliefs
produced by the Holy Spirit according to the design plan aimed at
producing true beliefs about the content of God’s revelation, except
as those beliefs in this area that are (on our evidence) probably
true? That evidence may, I repeat, be for some of us merely the
intuitive obviousness of such beliefs, or the unanimous testimony of
those known to us that every knowledgeable person holds these
beliefs. But surely for almost all of us today public evidence of the
kind I summarized earlier must play a signi cant role. Without
evidence that Christian doctrines are probably true, and so ‘justi ed’
in an internalist sense, we can have no reason to believe that they
have warrant in a Plantinga-type sense. Similar criticisms apply to
other externalist theories of the justi cation of religious belief.
Plantinga, as also Lamont, identi es the process of the production of
our Christian beliefs in such a way that, if they are indeed produced
in that way, they are very probably true. If the Holy Spirit causes us
to believe that Scripture is true, then very probably Scripture is true.
So we need evidence that the beliefs are produced by the relevant
process. Alston by contrast describes the process of production of
Christian beliefs in a non-question-begging way; ‘Christian mystical
practice’ can still produce beliefs even if there is no God. What he
assumes is that this process is reliable; and what we need is
evidence that the process is reliable. Only then will it be probably
true (on our evidence) that a belief it yields is justi ed. Yet in either
case if we have an internally justi ed belief (however produced)
that Christian doctrines are true, there seems no obvious need for a
further belief that they were caused in us by a special type of
reliable process (unlike that by which any other true beliefs are
caused). More generally, I suggest, an externalist warrant for
Christian beliefs has no role to play in our religious lives; but we
certainly need an internalist justi cation for them if it is to be
sensible to use them to guide our lives.

NOTES

1. Second edn. (Oxford: Clarendon, 2007). I am grateful to the

Oxford University Press for permission to use material from
this book in this chapter.

2. (Oxford: Clarendon, 1987).

3. Thus St Basil of Caesarea wrote that ‘time would fail me if I

attempt to recount the unwritten mysteries of the church’ (On
the Holy Spirit 27. 67). Although the sample list of mysteries
that he does recount are all mysteries of liturgical practice
(e.g. signing a catechumen with the sign of the Cross, and
blessing the water of baptism and the oil of chrismation before
their use), he argues that doctrine can be derived from
liturgical practice. He makes this assertion in the course of
arguing that if the divinity of the Holy Spirit were not
 explicitly stated in Scripture, that would not su ce to dismiss
it as not an essential part of the Christian faith. For the
divinity of the Spirit, he argues, is clearly implicit in the words
used by the church when catechumens are baptized. The
second Council of Nicaea, the seventh Ecumenical Council
(recognized by both Orthodox and Roman Catholics),
anathematized anyone who ‘rejects any written or unwritten
tradition of the church’.

4. Epistolae 4. 31.

5. For descriptions of di erent modern ways of interpreting the

Old Testament, see John Barton, Reading the Old Testament.
Barton defends explicitly the thesis that there is no privileged
method of interpretation.

6. De doctrina Christiana 3. 10. 14.

7. Thus Augustine (De Genesi ad Litteram 2. 1. 4, translated by J.

H. Taylor (New York: Newman, 1982)): ‘Let no one think that
because the Psalmist says “He established the earth above the
water”, we must use this testimony of Holy Scripture against
those people who engage in learned discussions about the
weights of the elements. They are not bound by the authority
of our Bible; and, ignorant of the sense of these words, they
will more readily scorn our sacred books than disavow the
knowledge they have acquired by unassailable arguments or
proved by the evidence of experience. The statement of the
Psalmist can with good reason be understood guratively.’
 8. Beryl Smalley, The Study of the Bible in the Middle Ages (Oxford:
Basil Blackwell, 1952), 14.

9. See Origen, De Principiis 4. 2. 4; and Augustine, De Utilitate

Credendi 5–8.

10. See e.g. Jerome’s Commentary on the Psalms 137:9—‘“Little

ones” means thoughts …the Rock is Christ.’

11. Origen interpreted the story of Joshua’s brutal execution of the

ve kings in Josh. 10: 15–27 allegorically. See the comparison
of his treatment of this story with that of others in R. P. C.
Hanson, Allegory and Event (London: SCM, 1959), ch. 5.

12. De Genesi ad Litteram 4. 22.

13. Ibid.1. 21.

14. Adversus Haereses 4. 32. 1.

15. Tertullian, De Praescriptione Haereticorum, 19.

16. Article 1. 9 of the Westminster Confession.

17. (New York: Palgrave, 2001).

18. John 1:1.

19. John 14: 28.

20. J. H. Newman, The Idea of the Development of Christian Doctrine;

expounded in Owen Chadwick, From Bossuet to Newman
(Cambridge: Cambridge University Press, 1957), ch. 7. See p.
157.
21. John Calvin, Institutes of the Christian Religion, trans. F. L.
Battles (Philadephia: Westminster, 1960), 1. 7. 2.

22. Second edn. (Oxford: Clarendon, 2004).

23. (Oxford: Oxford University Press, 2000). See pp. 268–80.

24. (Grand Rapids: Eerdmans, 2006).

25. For his general theory of warrant, see Alvin Plantinga, Warrant:
The Current Debate and Warrrant and Proper Function (Oxford:
Oxford University Press, 1993). In the text I summarize
Plantinga’s summary of his theory, ibid. 194. For Plantinga’s
application of his theory to Christian doctrines, see his
Warranted Christian Belief (Oxford: Oxford University Press,
2000).

26. See his collection of essays, Epistemic Justi cation (Ithaca:

Cornell University Press, 1989), and in it in particular ‘An
Internalist Externalism’. For the application to the results of
‘Christian mystical practice’ see his Perceiving God (Ithaca:
Cornell University Press, 1991). (Note his remarks on pp. 75–6
of this work about its relation to the previous essay.) Alston
has, however, now abandoned the view that there is a unique
true account (externalist or other) of epistemic ‘justi cation’,
and presents a pluralist account of the di erent epistemic
desiderata (externalist and internalist) that can be possessed
by a belief, in his Beyond ‘Justi cation’ (Ithaca: Cornell
University Press, 2005).
27. (Oxford: Clarendon, 1996).

28. (Aldershot: Ashgate, 2004). My quotations are from pp. 196

and 215.

29. Ibid. 209.

30. I discuss the kind of ‘faith’ necessary for salvation, and how
this di ers from belief, in my Faith and Reason, 2nd edn.
(Oxford: Clarendon, 2005).
 CHAPTER 2

REVELATION AND INSPIRATION

STEPHEN T. DAVIS

IN this chapter, we will consider the concepts of revelation and

inspiration.1 The two notions are distinct but closely connected in
Christian theology; they come together preeminently in discussions
of the Bible. There are three assumptions that form the basis of this
chapter but will not be argued for: (1) God exists; (2) there is divine
revelation; and (3) the Bible plays a crucial role in revelation. Let us
begin with revelation and then turn to inspiration.

One way to grasp the importance of revelation is to consider a

God who does not reveal, namely, the God of Deism. Deism was a
loosely de ned philosophical and religious movement that thrived
in Europe and America in the eighteenth century. The Deists
a rmed the existence of God and held that God created the
universe, with its immutable natural laws. They had con dence in
the power of human reason to reach proper conclusions about
religion, which for the Deists consisted of a few simple truths about
God, creation, and morality.
 The Deists importantly di ered from traditional Christian
thought in their rejection of all robust notions of divine revelation
(except ‘natural revelation’, i.e. conclusions about God and religion
reached by unaided human reason). Indeed, they rejected all claims
of divine intervention in the world—not only special revelation but
also miracles, epiphanies, and incarnations. God—so they implied—
is like someone who winds a clock and then lets it run on its own
without interference.

Oddly, Deism has made a recent comeback, although its

contemporary defenders do not use that term. Despite di erences,
there are striking similarities between the original Deists and those
recent and contemporary religious thinkers who a rm God but
deny divine intervention in the world.2 Some contemporary Deists
a rm a notion of revelation, although the concept ends up being
quite unlike traditional notions of God speaking or being
supernaturally manifested in the world. Their operative notion of
revelation seems to be more like: (1) insight into religious mysteries
gained by those blessed with the relevant sort of spiritual wisdom;
(2) enlightenment gained from a crucial event in personal or
communal history; or (3) new truth arrived at via interpersonal or
interreligious dialogue, truth that the participants would not have
discovered independently. So the idea is that God is ‘revealed’
through ordinary historical, cultural, intellectual, or scienti c
processes.
 In contrast, Christians believe in a God many of whose revelatory
acts constitute interventions in the ordinary course of history. But it
must be emphasized that human beings have no moral right to be
recipients of divine revelation, no claim on God that God must
reveal himself. So Christians should still—if they understand matters
rightly—feel a sense of wonder when they grasp the fact that God
has chosen to be revealed. There was no imperative or necessity that
God do so. God could have remained silent.

What would have been the consequences if God had remained

unrevealed? Suppose that the Deists were correct: suppose that God
did not much care about the a airs of the world and that God never
gave us a law or sent us prophets or a Son. What would follow?
Certainly human beings would still be interested in religious
questions. The Philippian jailer posed with elegant simplicity the
most important of them: ‘Sirs, what must I do to be saved?’ (Acts 16:
30). The obvious conclusion is that if God were silent, we would be
left to our own resources when we tried to answer questions like
this.

What types of answers to such religious questions might we

come up with? One possibility is some sort of legalism, i.e. a
religious system containing little more than a set of rules that
human beings must obey. This sort of religion would say: obey the
laws and you’ll be saved; disobey them and you’ll be lost. A second
possibility would be ritualism. This sort of religion would say: if you
follow the prescribed ceremonials, the gods will be placated and
you’ll be saved; if you fail to follow them, you’ll be lost. A third
possibility, much in the spirit of our age, is relativism. This sort of
religion would generously suggest that it doesn’t really matter what
you believe or how you behave; what matters is whether you are
sincere and try hard to behave according to your best lights. A
fourth possibility might be some version of nihilism: this sort of
theory would deny that there are any answers; even if God exists,
there is no way of knowing God or anything else for sure; there are
no ultimate values; there is no justice; death ends everything. The
most we can do is try as best we can to make life a bit more
endurable.

In short, if God were silent, i.e. if there were no revelation of any

sort, it is most probable that human beings would come up with
false religious ideas. When revelation is lacking or not recognized,
the world sets the intellectual and religious agenda. We see this
today: in many circles, religion is held to be relativistic, private,
syncretistic, deistic, or maybe all four. Such notions, in my view, are
mistaken. And they illustrate the point that apart from revelation,
there is no sure answer, based on human wisdom alone, to the
question of the Philippian jailer. If God were silent there would be
little hope of salvation; we would be like those who ‘go down to the
Pit’ (Psalm 28:1).

II
What is the purpose of divine revelation? It is to achieve God’s aims
in creation. Pre-eminently, God desires that human beings freely
love, worship, and obey God, and (as the answer to Question 1 of
the Westminster Shorter Catechism says), ‘enjoy him forever’.
Revelation, then, is not primarily for the purpose of imparting
information, issuing commands, or initiating ceremonies, although it
does all those things. Revelation exists for the essential purpose of
establishing a personal and loving relationship between God and
human beings.

To reveal is to unveil, show, or disclose something that was

hidden or unknown. Divine revelation is God disclosing things that
were hidden from human beings or unknown to them. Revelation is
a way for God to leap across the gap, so to speak, that separates God
from human beings. More precisely, there exist three gaps—an
ontological gap, an epistemological gap, and a moral gap. The rst is
due to the di erence between a necessary, eternal, all-powerful
creator and a dependent, temporal, feeble creature. The second is
due to the di erence between an omniscient being and an ignorant
being. The third is due to the di erence between a holy and morally
perfect being and a depraved and self-centered being.

Because of these gaps, human beings naturally know little of God

and God’s requirements, or at least not enough to accomplish God’s
redemptive purposes. And even things that they do know or can
learn on their own (perhaps that God exists, that murder is morally
wrong, etc.) can be denied, forgotten, or ignored. We are ignorant of
much that we need to know in order to be saved, and we are unable
to save ourselves. If we are to attain saving knowledge of God and
God’s purposes, that knowledge must come from God. It must be
revealed.

Scripture teaches that human beings were created ‘in God’s

image’ (Genesis 1: 26), which suggests that God and human beings
are in some ways signi cantly similar. Since ‘being similar to’ is a
symmetrical relationship (if A is similar to B, then B is similar to A),
then if we are (in some ways) similar to God, it follows that God is
(in some ways) similar to us. That signi cant similarity—whatever
exactly it is—is the ontological basis of revelation. Unless God and
human beings were importantly related in some such way (e.g. by
both being persons), revelation would be impossible. God has made
us to be receivers of revelation.3

Again, what God preeminently wants is for humans to come to

love, worship, and obey God. (For convenience we’ll lump all three
together and say that God wants us to ‘glorify God’.) Let us also
suppose that what God wants is for humans freely to opt to glorify
God. And the central obstacle that God faces in achieving these
desires is that human beings are both ignorant and sinful. They need
to be both taught and redeemed.

There are many areas in which humans are ignorant. We who

are asked to glorify God will need answers to at least the following
questions:
 • Does God exist?

• If God exists, what is God like?

• Exactly how do human beings ‘glorify God’, i.e. how do they go

about loving, worshipping, and obeying God? What exactly are
they required to do?

• What are the consequences of glorifying God, and what are the
consequences of not doing so?

Thus there are many things that human beings do not naturally
know or at least cannot easily come to know on their own and that
they must know, if God’s aims are to be realized (1 Cor. 15: 1–4;
Heb. 11: 6; 1 John 1: 1–3). Accordingly, one of the things that God
must do is nd a way to answer these and other questions.

The term ‘revelation’, then, preeminently refers to God’s actions

to answer for our bene t these sorts of questions. The claim that
God reveals himself is one of the most crucial claims of Christian
faith—as an act of sovereign grace, God reveals God to us. Thus the
prophet Amos says: ‘For lo, the one who forms the mountains,
creates the wind, reveals his thoughts to mortals, makes the morning
darkness, and treads on the heights of the earth—The Lord, the God
of hosts, is his name!’ (Amos 4: 13; italics added). Moreover, in
Christian theology, human ignorance and human sinfulness are
connected. Sin has the e ect of darkening our eyes to the truth,
especially to uncomfortable truth. So revelation has a redemptive
purpose. God does not just reveal interesting information to human
beings; revelation is God’s action aimed at creating a new
relationship between God and human beings and among human
beings.

Whatever revelatory methods God uses, the aim will surely be to

answer human questions in ways that are clear, lasting, and
convincing. A revelation is clear if it is unambiguous, if its meaning
is not easily mistakable. Of course any communication from one
person to another can be misinterpreted. Yet naturally God will
want the divine acts of revelation to be clear. A revelation is lasting
if its content is fairly easily passed on from one person to another
and from one generation to another. Now any revelatory act can be
forgotten or garbled through time, but God will clearly want the
divine acts of revelation to be lasting. A revelation is convincing if it
is powerful, illuminating, and tends to produce conviction. Of
course any revelatory act, no matter how convincing, can be
rejected, even if that rejection results from sheer stupidity or
stubbornness. Nevertheless, God will clearly want the divine acts of
revelation to be convincing.

III

There are many ways in which God might choose to reveal a given
message (let’s call it M) to some human being (let’s say Jones). Here
are a few: (1) God might create Jones in such a way as to be
naturally disposed to believe M. (2) God might telepathically cause
Jones to think or come to believe or at least cognitively entertain M.
(3) God might appoint someone as a spokesperson and cause that
person to say or write M to Jones. (4) God might cause Jones to
dream M. (5) God might miraculously bring it about that Jones sees
M written on a wall. (6) God might do some non-linguistic deed or
action whose proper interpretation or deep meaning is M.

Of possible modes of revelation, some are linguistic (i.e.

primarily involve words and sentences) and some are non-linguistic
(i.e. primarily involve deeds or actions). God’s revelatory action of
rescuing Israel at the Red Sea was primarily non-linguistic; it was an
action that spoke to Israel about the character of God. The
Decalogue that God gave to Israel via Moses was primarily
linguistic; it consists of words. This raises a theological question—
much discussed in the previous century—whether divine revelation
is by words or deeds.4 Some theologians denied that God reveals
propositions: such things—so they claimed—are timeless, static,
impersonal, and cold. God’s primary mode of revelation—so they
insisted—was through dynamic personal encounters; God is revealed
in deeds or persons or events rather than words.5

It does seem, a priori, that there are both values and problems in
either approach. Actions are sometimes more impressive, powerful,
gripping, and graphic than words. Just saying that you love your
spouse or child is typically less convincing than showing it. Yet the
problem with revelatory actions is that they seem more readily
susceptible to being misinterpreted, changed in the retelling of
them, and (unless they are written down) forgotten over time.
Perhaps words are often less powerful than deeds; but revelatory
words are valuable because they are not quite so easily
misinterpreted, are easier to preserve and pass on, and once
preserved are not quite so easily forgotten. These last two points are
important given God’s redemptive aims, i.e. given the assumption
that God would intend at least some revelations to be for the bene t
of other folk beside the original receivers or witnesses (see Psalm
22: 29–31).

It seems clear that God has used both modes of revelation.

Moreover, God’s great revelatory actions have typically been
accompanied by authoritative verbal interpretations.6 Indeed, it is
not easy to see how a bare event, action, or encounter could be
revelatory in any clear sense without interpretation or explanation.
Events do not interpret themselves. And as soon as an interpretation
is o ered, then we are talking about words and sentences. A bare
uninterpreted encounter with God (e.g. in the Exodus) will be far
more easily recognized as such if it is explained or at least
conceptualized as an encounter with God, as of course it is in
Exodus 11–15. And if a claim is made that there is truth involved in
a revelatory act or deed, it follows that the revelatory act be
expressible in propositions.

Concerning the debates about propositional versus non-

propositional revelation, it must be said that God is indeed revealed
in deeds, persons, and human experiences (see 2 Cor. 12: 1–4). But
it seems absurd to deny that God reveals words. To experience an
encounter with God logically presupposes the possession of at least
some knowledge about God. Thus Avery Dulles says: ‘If we had no
con dence in the propositional teaching of the Bible, we could
hardly put our trust in the persons or events of biblical history, or
even in the God to whom the Bible bears witness.’7

The enemies of ‘propositional revelation’ argued that Christian

faith is trust in the person of God or Christ rather than in any
proposition. As noted, their view was that propositions are too
abstract, impersonal, and static to constitute the proper objects of
religious faith; they elicit or call for no genuine response. This
criticism contains a kernel of truth, but only that. It is true that
Christian faith—or at least that crucial aspect of it that involves
trust ( ducia or faith in)—is in God rather than any proposition. But
surely the belief aspect of faith ( des or faith that) does essentially
involve acceptance of certain propositions as true. Moreover,
believing the truth of a proposition is inextricably tied to trusting a
person in those cases where the person trusted takes responsibility
for the proposition, e.g. by saying or writing it.8 My trust in ( ducia)
my mother, for example, makes no sense unless it involves
willingness to believe ( des) what she says to me.

It is misleading, then, to argue that propositions are too static

and invoke no genuine response. Christians would make no positive
response to God at all unless they accepted as true such propositions
as: God loves us; Christ died for our sins; and we should thank and
praise God. Furthermore, the Bible itself claims that God reveals
words. God is revealed in the words of the Decalogue, the oracles of
the prophets, the parables and other teachings of Jesus, and the
epistles of Paul. And many of those words were grammatically
con gured as propositions. They convey cognitive information; they
contain truths. It seems to me that those theologians who denied
that revelation was propositional believed (in fact, if not o cially)
in propositional revelation. They simply did not like the
propositions that more conservative interpreters of the Bible
claimed to nd there, and so (in e ect, if not by admission) they
looked for and—not surprisingly—found others.

IV

We have noted some ways in which God might have chosen to be

revealed to humans. But how did God actually choose to be
revealed? A distinction is often made between general and special
revelation. General revelation consists of those things about God,
human beings, morality, and religion that human beings can learn
on their own, i.e. without any supernatural (‘special’) assistance
from God. The Bible teaches that some such things can be learned in
this way (cf. Psalm 19: 1; Acts 17: 22–9; Romans 1: 18–23).
Theologians have suggested that we can see evidence for them, for
example, in considering the beauty and grandeur of creation, in
reasoning cogently about God, or maybe even in examining our own
consciences. But natural revelation, even at best, is incomplete,
hazy, and easily confused. It is in need of supplementation.
 Special revelation consists of those things about God, human
beings, morality, and religion that are relevant to our salvation and
that we can learn only as the result of, or are di cult to learn apart
from, some supernatural or special act of assistance by God.
Typically these would consist of things revealed by God to some
person or group through dreams, visions, epiphanies, prophecies,
miracles, or (supremely) through the life, death, and resurrection of
Jesus Christ. The Scriptures hold that there are such acts (cf.
Matthew 16:17; Luke 4:18–19; John 1:14; Galatians 1:11–12;
Hebrews 1: 1–4). Christians believe that reading Scripture can also,
through the illumination of the Spirit, constitute revelation.

Under the heading of special revelation, let me distinguish

among three sorts of revelatory acts on the part of God. I will call
them original revelation, recorded revelation, and appropriated
revelation.

Original revelation consists of divine revelatory actions in history,

some in the form of events and some in the form of words. This sort
of revelation would preeminently include: the Exodus, the giving of
the Decalogue, the oracles of the prophets, the teachings and
miracles of Jesus, and the cruci xion and resurrection of Jesus.
These are divinely ordained events or divine words that appear, by
God’s initiative, in the history of the people of God. Although God is
still at work in history, Christian tradition holds that original
revelation ceased with the death of the last apostle.9 Christians
celebrate the fact that God still speaks to people, moves them,
guides them, and works in various other ways in their lives. But
these divine actions do not count as original revelation and are not
theologically normative for all Christians.10

But original revelation will most likely be eeting and unstable

unless it is recorded in some way, such as being written down. Oral
traditions can certainly endure with amazing accuracy and for long
periods of time, and did so in the ancient world. Moreover, written
records can be corrupted or lost and forgotten. Still, if it is God’s
intention that as many human beings as possible know about God’s
acts of original revelation, then those acts must be recorded and
interpreted in some sort of authoritative written document. And this
is recorded revelation.

The text that Christians call the Bible is in part a record and
interpretation of original revelation. By reading the Bible it is
possible to learn some of what God has done in the past. Much of
what the biblical writers recorded need not have been
supernaturally revealed to them—doubtless they learned about
some of the events and words that they recorded by quite ordinary
means. The standard Christian claim is that recorded revelation, like
original revelation, ceased long ago—when the last book of the
Bible was completed. Since then presumably other Christian
writings—sermons, hymns, liturgies, creeds, and perhaps even
theological essays—were written under God’s in uence.
Pronouncements of some of the early ecumenical councils have, in
my view, a certain degree of normativity for Christians. But, as a
Protestant, I do not hold that they are theologically authoritative for
all Christians in the same sense that the Bible is.11

But the Bible is not merely (as is claimed in some circles12) a

record of previous divine revelation. The Bible itself is revelatory.
Christians hold that, through the illumination of the Holy Spirit,
God speaks to us in the Bible. When it is correctly interpreted, the
Bible in and of itself constitutes a unique and authoritative divine
revelation. As Paul Helm points out, probably some things that were
not necessarily revealed to the biblical writers, things they learned
quite naturally, do constitute revelation to us as we read them in the
Bible.13 When God does speak to us in Scripture, e.g. when a
biblical text, for example, inspires a reader to repent of a certain sin
or assist a certain needy person, this is appropriated revelation. It is
simply recorded revelation speaking to the reader, as we might say,
in God’s voice. Appropriated revelation adds nothing to original and
recorded revelation, nor did it cease with the death of the apostles,
nor are its deliverances theologically authoritative for all Christians.
Not all Christian denominations or traditions would accept this
strong statement; still, my own view is that no post-apostolic vision,
prophecy, dream, exegesis, creed, pronouncement, sense of
discernment, decision, or conviction is normative for all Christians.
It is authoritative only to the person by whom the revelation is
appropriated, to whom the Holy Spirit is speaking. Appropriated
revelation is an act of God; when I correctly sense that God is
speaking to me in Scripture, that is a supernatural act of the Spirit of
God.

Suppose God wanted to be revealed to as many people as

possible and to as many sorts of people as possible—young and old,
men and women, black and white, Eskimos and Hottentots, scholars
and peasants. If so, it would seem appropriate for God to decide that
at least part of divine revelation would consist of, or be crucially
located in, a book. Presumably God’s intention in doing so would be
to make the content of revelation available to as wide a population
as possible, and not just to those present at the appropriate time, to
an elite class of scholars, or to those initiated into mysteries. This
population would surely include people in generations and centuries
after the revelation was given. A book seems particularly well suited
to accomplish these ends.

It is inevitable that such a book would appear in a certain

historical, national, cultural, and linguistic context, and that those
who lived in that context would at least initially have the easiest
access to the book. We cannot explain why God would choose one
human group rather than another to be the original home of the
book we are envisioning.14 Still, if the revelatory book that rst
appeared in a given context were readily translatable into other
contexts, God’s aims could still be achieved.

If something like this picture were indeed God’s revelatory

choice, it would seem that God would then superintend the process
of the writing of the various documents of which the book would
ultimately consist, ensuring that they contained, at least in large
part, what God wanted them to contain. That is, enough of their
content would have to correspond to God’s revelatory intentions so
that people could read them and be reasonably sure of learning
what God wanted them to learn. It would also seem plausible to
suppose that God would ensure the church’s recognition of the book
as Scripture as well as superintend the process of copying,
translating, reproducing, and preserving the book. All this would
occur so that those who read the book could be reasonably sure that
what they read corresponded reliably, even if not perfectly, to what
was originally intended. There would also have to be ways for the
errors of copyists, translators, and printers to be detected and
corrected.15

According to Christians, this is precisely what did occur. God

actually did choose to be crucially and normatively revealed in a
book, and that book is the Bible. God did supervise the creation,
production, transmission, and preservation of the book in roughly
the indicated ways. The fact of God’s providential guidance of the
history of the Bible makes it unique among all the books that have
ever been written in the world’s history. It reveals God in a unique
and normative way.

However, the Bible, by itself, was not su cient to accomplish

God’s redemptive goals. A community was needed in order to
provide authoritative texts, translations, and interpretations of the
Bible. In this way, the original revelation could best be preserved
and communicated.16 The notion, sometimes heard in Protestant
circles, that the ‘Bible alone’ is su cient to accomplish God’s
revelatory and salvi c aims is quite mistaken.

This is not to deny the traditional Protestant notion or the clarity

or perspicuity of Scripture. The Bible’s basic message can be
understood by nearly anyone who is open and attentive to it, even
children. But it is obvious that not everything in Scripture is plain
and that a method and tradition of interpretation (e.g. what the
church Fathers called ‘the rule of faith’) is needed.17

Is revelation ‘progressive’? Are later divine revelations more

accurate and sophisticated than earlier ones? The concept of
progressive revelation has been used to solve problems caused by
the fact that some earlier revelations seem crude and even immoral
when compared with later ones. Some have felt, for example, that
this notion might be a way to reconcile (1) the insistence of the
books of Deuteronomy (see 2: 31–5; 3: 1–8; 7: 2) and Joshua (see 6:
15–21; 8: 25–6; 11: 12) that it was God’s will that the Israelites kill
every Canaanite man, woman, and child with (2) the law of love of
the New Testament. The argument would go something like this: the
earlier revelation was the best that God could do at the time, the
most that God could reveal, given the level of theological and moral
sophistication of the post-Exodus Israelites; they would have been
unable to understand the more complete revelation of God as found
in the New Testament. But by the time of the rst century, people
were ready to hear and receive that fuller revelation of the divine
nature and so were able to receive the teachings of Jesus.

There is both something right and something mistaken here. All

revelation, both earlier and later, involves accommodation or even
condescension on the part of God. To borrow Calvin’s metaphors,
God must ‘stoop’ and even ‘lisp’ (i.e. engage in a sort of baby-talk)
to speak to us.18 All revelation must meet us where we are. The
ontological, epistemological, and moral gaps between God and
human beings ensure that we cannot comprehend the fullness of
God’s nature and will. So all revelation is partial. As Paul says, we
now know ‘in part’, not ‘face to face’ (1 Corinthians 13: 12). What
God has revealed is trustworthy; we do indeed have knowledge of
God. But it is only provisional, incomplete, and inchoate. Even as
revealed, much about God remains mysterious to us.

What is mistaken in the concept of progressive revelation? It is

the suggestion that the ancient Israelites were so morally or
intellectually primitive that God could not get them to grasp the
fuller revelation that appears in the New Testament—a revelation
that the much more enlightened folk of the rst century (as well as,
of course, those of us in the twenty- rst) are clever and moral
enough to grasp. That is not a particularly convincing claim. There
is a sense in which revelation is progressive, but it has nothing to do
with any inability on God’s part to teach certain folk things that
they were too dumb to understand.
 Revelation is progressive in that: (1) as a sovereign choice, God
decided to be revealed over time rather than all at once, and in
earlier times did make concessions to then prevalent views, allowing
such things as slavery, polygamy, and divorce (Matthew 19: 3–9);
(2) as a sovereign choice, God decided to be revealed in such a way
that earlier revelations are interpreted by, incorporated within, and
sometimes superseded by, later ones (see Matthew 5: 21–48); and
(3) as a sovereign choice, God decided that normative revelation
occurs only for a time and then ceases. Thus the picture of God and
of God’s will that revelation paints for us is indeed more complete at
the end than it is at the beginning. All three points are contingent,
however; God could have chosen to do otherwise.

To accept a revelation from God is to be subject to something

like a paradigm shift. Revelation typically calls for a response, a
response of repentance, obedience, commitment, or trust. But it is
perfectly possible to receive a revelation from God, or to have
reliable evidence of one, and remain unmoved. Even the ‘signs and
wonders’ that accompany some revelatory acts—e.g. the Exodus or
the resurrection of Jesus—can be explained away by those who are
hard-hearted or spiritually blind. God structures all God’s revelatory
acts—even the Christ event—in such a way as to ensure our
cognitive and moral freedom to say no. God does not coerce the
proper response.

Again, the purpose of revelation is to bring it about that humans

come into a personal relationship with God, one that involves freely
chosen love (which cannot be coerced) as well as worship and
obedience. I argued above that God’s aim would be to make all acts
of divine revelation clear, lasting, and convincing. But they cannot
be so overwhelmingly convincing as to coerce consent—as, say, the
existence of my computer is to me now. The evidence for its
presence in front of me is so overwhelming that no sensible person
in my shoes could reject it. If God (as I believe) wants us to be free
rationally to say yes or no to God, then God’s revelatory acts cannot
be so convincing that no sensible person could reject them.19

Still, if revelation is accepted by someone who did not previously

know God, there must be change. A new notion of God is provided,
as well as a new notion of what we owe God. Indeed, there is
usually a radically new way of looking at all human experience and
reality. When this occurs in someone’s life, it is the work of the Holy
Spirit. That work marks the di erence between merely receiving a
new bit of information, a new command, or a new ritual, on the one
hand, and receiving a divine revelation, on the other.20

How then can our spiritual questions be answered? How can we

learn how to glorify God and thus satisfy God’s redemptive aims in
creation? The answer is: preeminently by accepting and attending to
the revelation of God in Scripture. All Christians a rm that Jesus
Christ is the supreme revelation of God. They should also a rm that
the Bible is the primary appointed place where Christians encounter
Jesus Christ and learn the divine will.
 VI

Let us turn to inspiration. In 1 Corinthians 7, Paul addressed himself

to questions such as marriage, divorce, celibacy, and sexual
relations. In doing so, he carefully distinguished between his own
teachings and those of Jesus. When Paul turned to the topic of the
freedom of widows to remarry (vv. 39–40), he noted that what he
said constituted his own teaching (‘in my judgment’); but he then
added, ‘I think that I too have the Spirit of God.’ Paul seems to have
been claiming, then, that he was teaching the Corinthians what the
Spirit of God would have him teach, that the Spirit stood behind or
validated his own teachings.

The much wider claim that the writers of the texts that we call
the Bible were so guided that what they wrote was what the Spirit
of God would have them write is called inspiration. Traditionally,
inspiration has to do with the in uence of the Holy Spirit on the
writers of the Bible so that what they wrote was, in some sense, the
word of God. Several of the biblical writers were convinced that
inspiration occurs. In 1 Corinthians 2: 9–15, Paul quoted an Old
Testament text, and then said: ‘These things God has revealed to us
through the Spirit.’ And 2 Peter 1: 20–1 speaks of scriptural
prophecy in this way: ‘No prophecy ever came by human will, but
men and women moved by the Holy Spirit spoke from God.’

But serious di culties face those who wish to write about this
topic. Most of them fall under this heading: the notion of inspiration
must t the kind of document that the Bible actually is.21 First, an
acceptable notion of inspiration must t the fact that while some
biblical writers (e.g. Paul, the Old Testament prophets) were at
times conscious of having a ‘word’ from God, others were not. Some
biblical writers seem to have been conscious simply of having
gathered information from various sources, and some of those
sources were incomplete or defective.22 Second, an acceptable
notion of inspiration must t the fact that in the Bible there are
distinct traditions about the same events and discrepancies between
them.23 There are not just numerical and historical discrepancies,
and not just apparently misidenti ed or misinterpreted quotations of
the Old Testament in the New Testament. There are also theological
discrepancies, or at least apparent ones. Third, an acceptable notion
of inspiration must t the fact that some biblical books had a long
history of composition and that the very notion of ‘the author’ is
irrelevant.

Traditional theories of inspiration have focused on the human

author or authors of the text and the text itself. They t what is
called the ‘prophetic model’ of inspiration, i.e. the model that
envisions a biblical author writing an inspired book or text under
the in uence of the Spirit. The central idea of this model is that God
inspired the writers of the Bible in a similar way as God inspired the
Old Testament prophets and gave them a ‘word’ from God to the
people. But under this model, there are several ways of formulating
a doctrine of inspiration. Let me mention several possible theories,
starting with the strongest and moving toward the less robust ones.
 1. Dictation. The strongest and simplest of the theories, dictation
holds that God alone is the author of Scripture; God dictated to the
human authors the very words that they wrote down, so that in the
(text-critically acceptable) words of Scripture we have the
unambiguous words of God. Now almost nobody who talks about
biblical inspiration admits to believing in dictation, although some
views of robust biblical inerrancy are not easy to disentangle from
dictation. The compelling reason that dictation is seldom defended
is that the many di erences in style, vocabulary, genre, and even
substantive content that exist in the various books and passages of
Scripture seem inconsistent with dictation. Unless God made sure to
dictate the words of Galatians to Paul in Paul’s own distinctive style,
and the words of Ecclesiastes to Qoheleth in Qoheleth’s own very
di erent style, the theory collapses. That is, it collapses as a general
theory covering the whole of Scripture. But dictation may well be
the most helpful theory of biblical inspiration for some texts
(assuming they capture accurately what was originally revealed),
e.g. the Ten Commandments, the teachings of Jesus, and perhaps
even some prophetic oracles.24

2. Inspired ideas. On this view, God inspired the biblical writers

by giving them the correct ideas or images, which those writers then
eshed out or expressed by means of their own words. God left the
choosing of the words pretty much up to them. The Bible’s good
points, so to speak, are a result of the divinely supplied ideas or
images; its bad points are a result of poorly chosen human words.
But this theory has received little support.25 Perhaps it has some
vague relevance to some parts of scripture, e.g. descriptions or
explanations of visions in Ezekiel or Revelation. But the basic
problem is the philosophical point that words are too closely
connected to the ideas they express for such a distinction as to their
origin to be consistently made.

3. Middle knowledge inspiration. William L. Craig has recently

proposed a theory of biblical inspiration that is both new and
serious.26 It is based on the notion of ‘middle knowledge’, a complex
and somewhat controversial notion that it is not possible to explain
here in any detail. Invented by the Jesuit philosopher Luis de
Molina (1535–1600), the idea is that at the moment of creation,
God knew three sorts of things. The rst is God’s ‘natural
knowledge’, which is knowledge of all necessities and possibilities,
i.e. of what could be (e.g. ‘Possibly the world will contain no ten-
foot tall human beings’) or must be (e.g. ‘No triangles are four-
sided’). The second is God’s ‘free knowledge’, which is knowledge of
all propositions that are or will be contingently true of the actual
world (e.g. ‘George W. Bush will be President of the United States’).
Located between these two—so Molina claimed—is God’s middle
knowledge (also called counterfactual knowledge); it is knowledge
of what would be the case if certain non-actual circumstances
obtained, e.g. ‘If Stephen Davis is o ered a peanut butter cookie to
vote for a certain candidate, he will refuse.’27
 Note that via his middle knowledge, God can providentially
control people without taking away their (libertarian) freedom and
thus can bring about the results that he desires. Craig then holds
that divine middle knowledge allows us to a rm that scriptural
inspiration was plenary (all of Scripture, not parts of it, were
inspired), verbal (the very words of Scripture, not just the ideas,
were inspired), and con uent (both God and the human authors are
the authors of Scripture). As middle knowledge applies to the Bible,
Craig a rms that ‘God knew, for example, that were he to create
the Apostle Paul in just the circumstances he was in around AD55
[given Paul’s background, personality, environment, and promptings
and gifts of the Holy Spirit], he would freely write to the Corinthian
church, saying just what he did in fact say.’28 So the letter of 1
Corinthians is truly the work of Paul (he freely wrote; the words are
Paul’s, etc.) and of God (since the letter says exactly what God
wanted said).

4. Degrees of inspiration. The idea here would be that there were

levels of inspiration of the writers of Scripture, i.e. some biblical
texts or writers were more directly or thoroughly inspired by the
Holy Spirit than were others. Thus James Orr said: ‘There were
lower grades of inspiration in the form of special charismata
(wisdom, artistic skill, physical powers), shading o till it becomes
di cult to distinguish them from heightened natural endowment.’29
As Paul Achtemeier points out, such a theory ts well to the claim—
frequently made in mainstream denominational and neo-Orthodox
circles—that the Bible is not in and of itself divine revelation, but is
rather a human and in places fallible record of past acts of divine
revelation.30

5. Social inspiration. Perhaps God inspired not an individual

writer but a group of people, speci cally the community that was
responsible for bringing the biblical text in question into existence.
But while there is no good reason to deny that God ever did such a
thing, the notion of social inspiration does not help us much in
understanding inspiration. The question remains in what precise
ways God inspired the community or the individuals who made up
the community.

6. Secular inspiration. Although this theory does not fall under the
prophetic model of inspiration, there is another possibility along the
route that we have been roughly traveling from stronger to less
robust theories. Possibly the Bible is an entirely human book and, so
far as origin is concerned, there is no intrinsic di erence between
the Bible and any other book. To the extent that biblical inspiration
occurred at all (and some liberal Protestants reject the notion
entirely), perhaps the writers of the Bible were ‘inspired’ in some
sense not unlike the way the writers of great but entirely secular
books were inspired. Perhaps God uses the book called the Bible for
religious purposes more than God uses other books; perhaps the
church accepts the Bible as more authoritative than other books; but
there was no special divine in uence on the writings of the Bible
per se. The reasons usually given for abandoning robust notions of
inspiration are (1) the discrepancies in the Bible, and (2) the fact
that the Bible seems to speak approvingly of such practices as
slavery, the oppression of women, mass murder, the torture of
prisoners, and polygamy.31

There are obviously other possible theories, and no one theory

has gained universal acceptance. Indeed, the problems of arriving at
an adequate theory of inspiration that is author-oriented or text-
oriented are so serious that some contemporary authors seek to
locate inspiration elsewhere. That is, some choose to abandon the
prophetic model. Not only do many biblical authors seem unaware
of having received any inspired ‘word’ from God, but it is sometimes
unclear precisely who would have been inspired. The point here is
not the truism that there are biblical books and passages whose
authors are unknown to us (e.g. Hebrews); the deeper point is that
the complicated textual history of some books (e.g. Isaiah, John)
makes the notion of one writer, blessed by God with the gift of
inspiration, apparently inapplicable. When so many people were
involved in creating the text, who was the inspired author?

VII

Accordingly, some recent and contemporary scholars attempt to

locate biblical inspiration elsewhere than with the author or the
text. Obviously, we cannot discuss all such attempts; we will
consider two such theories, one by a theologian and the other by a
philosopher.
 In his book, Evangelical Theories of Biblical Inspiration: A Review
and Proposal,32 Kern Robert Trembath argues that what we call
biblical inspiration is a fact not about the Bible, but about the
reader of the Bible, whether that be the church or the individual
Christian. He says: ‘I shall propose that “the inspiration of the Bible”
should be taken to refer not to empirical characteristics of the Bible
itself but rather to the fact that the church confesses the Bible as
God’s primary means of inspiring salvation within itself’ (p. 5).

The Bible is thus the means of biblical inspiration rather than its
locus or terminus. Biblical inspiration depends on the faith
perspective of the reader of the Bible; it is the way in which a
believer or Christian community recognizes the indispensable role of
the Bible in redemption. Trembath admits that inspiration is not
totally independent of the words of the Bible, ‘but neither is it a
property which applies exhaustively to those words’. The phrase
‘biblical inspiration’, he says, is an abbreviated reference to ‘the
experience of salvation by God through Christ as mediated through
the Bible’ (p. 111; see also p. 62).

What about scriptural authority? Trembath denies that the

authority of Scripture is located in its words, ‘as though they could
have an authoritative status apart from their reception and
appropriation by the believers’ (p. 52). Scriptural authority is
conditional upon the appropriation of its central message by
believers; it is the Bible’s ‘indispensability in witnessing to the
relationship which God establishes through the church’ (p. 53). The
words of the Bible can be said to be inspired only when they are
received as God’s word by the believing community (p. 116).

Trembath recognizes that the traditional Reformed doctrine of

the inward testimony of the Holy Spirit is, as he says, the functional
equivalent of his theory of biblical inspiration. But the question is
whether it constitutes an adequate theory of what it purports to be,
namely, biblical inspiration.33 If Trembath insists on using the term
‘biblical inspiration’ in his preferred way, one might ask what new
term we are to use to describe God’s guidance (if there was any such
guidance) of the process of the production of the book that we call
the Bible.

Few Christians will allow that there was no such process. For if
not, the Bible will have no special status among the world’s books
apart from the functional point about its having shaped the
Christian experience of salvation more than any other book. Yet
even if we grant that the Bible has this unique functional status for
the church, the question remains whether it has any special
ontological status. Is it or is it not intrinsically di erent from other
books? Is the Bible, in and of itself, unique? After all, some people
have been powerfully moved by The Iliad or The Apology or Anna
Karenina. And some people’s experience of Christian salvation has
been powerfully shaped by the Summa Theologica or The Institutes of
the Christian Religion or Mere Christianity.

Book-oriented theories of biblical inspiration34 usually involve

verbal inspiration, and this theory often leads to the notion that the
verbal claims made in the Bible are certainly true. Trembath
criticizes this notion on the grounds that certainty is not a property
of words (as the people who hold to the doctrine appear to hold),
but of the mind. He says: ‘Certainty is the conclusion of a judgment,
and judgments are acts of human minds rather than properties of
objects such as words’ (p. 91). But this claim is a frail reed on which
to hang a rejection of verbal inspiration. For some con gurations of
words—e.g. ‘Two plus two equals four’ and ‘Abraham Lincoln was
president of the United States’—have the property of being certainly
true. And this is surely what the friends of verbal inspiration want
to say about the words of the Bible. As regards the Bible, Trembath
says that ‘certainty results from the mental determination which
chooses the biblical accounts of redemption and salvation as those
which best summarize the actual experience of salvation undergone
by Christian believers’ (p. 91). Again, there is truth here, but the
problem is that there is much in the Bible that seems certainly true
to many believers and that has little or nothing to do with
describing their ‘actual experience of salvation’.

In the end it is puzzling that Trembath devotes his book to a

sustained discussion of the special status of the Bible in the Christian
experience of salvation—making points that could fall under the
rubric of the inward testimony of the Holy Spirit—without raising
the question of the special intrinsic character of the Bible. This is
curious because Trembath insists both that knowledge of God can
only come from God (pp. 5, 85, 90, 93) and that the Christian
community’s beliefs about God have been preeminently shaped by
the Bible (p. 111). Does this not at least suggest that the Bible is
‘from God’ in some interesting way? If so, in what precise way is the
Bible ‘from God’? This is the question that book-oriented discussions
of biblical inspiration were designed to answer. And the central
moral that I draw from Trembath’s work is that discussions of
biblical inspiration ought to continue to address the question of the
intrinsic character of the Bible.

VIII

In his ground-breaking book Divine Discourse,35 Nicholas

Wolterstor makes a suggestion that might be usable as part of a
non-book-oriented theory of biblical inspiration. But we must rst
note that Wolterstor ’s book is not primarily about inspiration or
even revelation, but about the concept of God speaking. Still, one of
the important points that he makes in his book is relevant to our
purposes.

Wolterstor points out that it occasionally occurs that a piece of

speech or writing is correctly attributed to a given person even
though it was composed by somebody else. Maybe a secretary who
knows the mind of the boss is asked by the boss to compose a letter;
the boss approves of the letter, signs it, and it then becomes the
boss’s letter. Ambassadors are entitled to ‘speak in the name’ of the
head of state whom they represent. Naturally, there can be di erent
modes and degrees of superintendence over the words and di erent
degrees in which the nal words of the text are authorized and
accepted as her own by the person who counts as the author.

So far as the Bible is concerned, we do nd in it the phenomenon

of prophets who are deputized or commissioned by God to speak a
message in God’s name. Wolterstor calls this ‘appropriated
discourse’; he suggests that the Christian claim that the Bible is
God’s book can naturally be understood, at least for much of the
Bible, as divinely appropriated human discourse (p. 53). So maybe
the Bible as a whole is God’s book in the sense that God
appropriates it. Accordingly, in the Bible God now speaks to us via
words that were composed by other people, namely the Bible’s
human authors. God is the ultimate author of Scripture (pp. 186,
283).

Wolterstor recognizes that the notion of appropriated divine

discourse must be supplemented by a doctrine of inspiration (p.
187); it can’t be that God had nothing whatever to do with the
composition of the books of the Bible, saw them (so to speak) lying
about, and simply decided to adopt the words as his own. Still, he
insists that the books of the Bible, human discourse as they are,
mediate to us God’s speaking. (Wolterstor never spells out the
notion of inspiration that he admits his theory needs as a
supplement.)

Wolterstor ’s notion of divine appropriated discourse seems

acceptable as a way of understanding how God speaks to us in the
Bible, at least in many of its books and texts. It is signi cant,
however, that Wolterstor sees clearly the need for this notion to be
supplemented by a theory of inspiration. That admission tends
strongly to con rm the suspicion that reader-oriented theories of
biblical inspiration, while valuable as far as they go, are insu cient
in themselves. They do not say enough about the intrinsic properties
of the Bible itself; they do not capture what most Christians want to
say about the Bible.

IX

Is it possible then to develop a defensible theory of inspiration that

concentrates on intrinsic properties of the Bible itself? Before trying,
let me mention three items that I reject. First, I do not hold that the
ideas of Scripture are inspired but the words are not. Second, I do
not hold that some parts of Scripture (e.g. the ethical or religious
parts, or the parts that pertain to its central salvi c purpose) are
inspired and the others are not. Third, I do not hold that some parts
of the Bible are more inspired than other parts (although certainly
some parts are more theologically crucial, religiously useful, and
hermeneutically central than other parts).

An acceptable theory of inspiration must recognize the fact that

the Bible is in some sense both the words of God and the words of
its human authors. There is both divine and human agency involved
in the production of the Bible. Accordingly, theories of inspiration at
both extremes are excluded: dictation because it denies human
agency and secular inspiration because it denies divine agency. The
Bible, then, is both the words of human beings and the word of God.

It is not hard to imagine scenarios in which God and human

beings work together to produce some result. The humans involved
may or may not be aware of God’s agency or in uence. Indeed, it
seems a clear presumption of biblical religion that this sort of
double agency occurs. Not only historical events (such as the
cruci xion of Jesus—Acts 2: 23) but also literary products
(according to 1 Peter 1: 20–1) can be the result of God and humans
working together: ‘No prophecy of scripture …ever came by human
will, but men and women moved by the Holy Spirit spoke from
God.’ That is, Scripture is the result of human beings speaking and
writing, but they spoke as moved by the Spirit.

What then is biblical inspiration? Let me de ne it as that

in uence of the Holy Spirit on the writing of the Bible that ensures that
the words of its various texts are appropriate both for the role that they
play in Scripture and for the overall salvi c purpose of Scripture itself. In
other words, working in conjunction with the human author, God
ensures that the words written are revelatory and t God’s purposes.
Because of inspiration, God speaks to us in the Bible. Scripture is not
just a record of revelation; it in itself is revelatory.36

Paul in 2 Timothy 3: 15–17 speaks of ‘the sacred writings that

are able to instruct you for salvation through faith in Jesus Christ’.
The sacred writings referred to here are surely the books of the Old
Testament; it then says: ‘All scripture is inspired by God and is
useful for teaching, for reproof, for correction, and for training in
righteousness.’ This text is one reason for holding, as argued above,
that the purpose of revelation is redemptive, i.e. its purpose is that
human beings come freely to glorify God. The word translated
‘inspired’ in 2 Timothy 3:16 (theopneustos) means ‘breathed out by
God’. Thus one venerable way of a rming that the Scriptures are
inspired is to claim that they are ‘breathed out by God’ for the
purpose of guiding people to life in Christ.

But it is also important to emphasize the point about God

working with the human author or authors. This means that
inspiration has nothing to do with authors being temporarily freed
from their human characteristics or limitations. It means that
despite those limitations, the words of Scripture t God’s purposes
for them. Nor does inspiration necessarily imply that the inspired
human authors experienced unusual psychological states such as
trances, visions, dreams, or the hearing of paranormal voices. In
some cases inspiration may have involved some such states, but in
other cases an observer would doubtless have noticed nothing
unusual about the behavior of the inspired author or authors.

I have argued against theories that locate biblical inspiration and

authority entirely in the reader. These notions must have something
to do with intrinsic properties of the Bible itself. Christians hold that
the primary reason that the Bible, in and of itself, is unique among
all the world’s books and theologically normative is because its
writers were inspired. That is why they hold that this book in some
sense originates from God, why they hold that God speaks to them
through this book.

Let us return to the point that the various books of the Bible
seem to have reached their nal form in di erent ways, and that
several such books had no one human author. In such cases, if
biblical inspiration is to make any sense, it must be that several
people were inspired, i.e. the Holy Spirit in uenced more than one
person. Perhaps inspiration a ected, say, the original prophetic
speaker (ensuring that the prophet spoke the words God intended to
reveal), the nal editor (assuming that this is a di erent person) of
the biblical book (ensuring that what it says is appropriate for God’s
purposes for it), and perhaps other people between them.

James Barr, a formidable critic of conservative views of biblical

authority, makes the following point:

If there is inspiration at all, then it must extend over the entire

process of production that has led to the nal text. Inspiration
therefore must attach not to a small number of exceptional persons
like St. Matthew or St. Paul: it must extend over a large number of
anonymous persons, so much so that it must be considered to belong
more to the community as a whole than to a group of exceptional
persons who through unique inspiration ‘gave’ the scriptures to the
community.37
In general, Barr’s point is well taken. We must insist that the process
can still include exceptional inspired persons who shape their
communities. But Barr is right that if inspiration is to make any
sense, it must mean that God superintends the whole process that
led to the existence of the book that we call the Bible.

Note the many ways in which God’s redemptive purposes tied up

in a book could be unful lled. The original revelatory event might
be misinterpreted. Over time, oral traditions about it might be lost
or distorted beyond recognition. The original piece of writing might
be inaccurate or misleading. Poor editing might make a ne original
document unintelligible, unhelpful, or otherwise salvi cally useless.
Writings not intended by God for inclusion in the book might be
included, and writings intended by God for inclusion in the book
might be excluded. The book could be copied or transmitted so
poorly that over time it became impossible to hear God’s voice in it.
The book, or crucial parts of it, might be subject to so much
systematic misinterpretation or mistranslation as to be useless in
achieving God’s purposes for it.

Small mistakes, perhaps at any point in the process, might be

consistent with the book’s achieving God’s purposes for it, especially
if they were either correctable or else so unimportant as to be
negligible. Still, the point stands: God must providentially
superintend the whole process of the production of the book that we
call the Bible. If the Bible is indeed an inspired book, then God has
done this very thing.
 Surely no one theory of biblical inspiration will cover the whole
of the Bible. Di erent theories seem relevant to di erent texts. As
noted above, dictation seems to be relevant to some passages. Social
inspiration, at least in some cases, must have occurred. Middle
knowledge seems promising as a way of understanding inspiration,
although the notion has yet to be thoroughly explored by scholars.
And Wolterstor ’s notion of divinely appropriated discourse is also
helpful.

There is a logical connection between the inspiration of the Bible, as

traditionally conceived, and the authority of the Bible. This
connection seems sensible: if a book has been so guided throughout
the history of its production that God speaks to us in it and it ful lls
God’s redemptive purposes for it, then that book will naturally have
a great deal of authority.

Authority has both a subjective and an objective sense. A given

civil law might have no authority over me in the sense of guiding
my behavior if I do not know about the law or do not submit to it.
The law, as we might say, has no subjective authority over me.
Some recent and contemporary scholars want to think of biblical
authority in this way: the Bible has authority over Christians
because they have chosen to receive and appropriate it as
normative. But some authority relationships (objective ones) exist
whether submitted to or not. Whether I know about or submit to the
law or not, it still has authority over me in the sense that my
behavior ought to be guided by it.

Similarly, Scripture has subjective authority over a person’s life

only if that person submits to Scripture. Sadly, many people today
fail to live their lives under Scripture’s authority. But because of
inspiration, Scripture is revelatory; God speaks to us in Scripture.
Accordingly, Scripture has objective authority over a person’s life
whether or not it is ever read, recognized, or submitted to. If in the
Bible God speaks to us, then we ought to submit to Scripture. Indeed,
as Swinburne points out, the Bible would have no authority for
Christians if the Christian community did not recognize its
inspiration.38 Of course all authority is from God, and belongs to
Jesus Christ (Matthew 28: 18); accordingly, the Bible’s authority
over Christian faith and practice is derived rather than original or
absolute. Nevertheless, the Bible is and ought to be the central way
in which God exercises authority over the Christian community.

NOTES

I would like to thank Professors Tom Flint, Susan Peppers-

Bates, Michael Rea, and Richard Swinburne for their helpful
comments on an earlier draft of this chapter.

1. Some of the material in the rst half of this chapter is a

slightly revised version of my Christian Philosophical Theology
(Oxford: Oxford University Press, 2006), 44–57.
2. It is possible to posit a robust notion of miracles that do not
amount to interventions. For example, God might foreordain
that certain events occur at certain points in time that would
not otherwise have occurred. But I will ignore that possibility
here.

3. This point is rightly insisted on by Kern R. Trembath in his

Divine Revelation: Our Moral Relation With God (New York:
Oxford University Press, 1991), pp. 4, 114–15, 136, 168–9.

4. The distinction that I am making between revelation by words

and revelation by deeds is close to the distinction that George
Mavrodes makes between the communication model of
revelation and the manifestation model of revelation. See his
helpful book Revelation in Religious Belief (Philadelphia:
Temple University Press, 1988), 35.

5. See e.g. John Baillie, The Idea of Revelation in Recent Thought

(New York: Columbia University Press, 1956), 49: ‘what is
fundamentally revealed is God Himself, not propositions about
God’.

6. This point was emphasized even in the neo-Orthodox period by

James D. Smart. See his The Interpretation of Scripture
(Philadelphia: Westminster, 1961), 172–3.

7. Avery Dulles, SJ, Models of Revelation (Garden City: Doubleday,

1983), 205. See also William Abraham, Divine Revelation and
the Limits of Historical Criticism (New York: Oxford University
 Press, 1982), 21, and Ronald H. Nash, The Word of God and the
Word of Man (Grand Rapids: Zondervan, 1982), 43–54.

8. See Paul Helm, The Divine Revelation (Westchester: Crossway

Books, 1982), 26.

9. See the Westminster Confession, I. 1, in John H. Leith (ed.),

Creeds of the Churches (Garden City: Doubleday, 1963), 193.

10. See Richard Swinburne, Revelation: From Metaphor to Analogy

(Oxford: Oxford University Press, 1972), 102. I should note
that much of what I say in this section of the chapter
presupposes the work of Swinburne.

11. My own more complete views on Scripture and tradition can be

found in my Christian Philosophical Theology, 265–82.

12. See Swinburne, Revelation, 103: ‘Holy Scripture must be

regarded by Protestants as it is by Catholics, as no more than a
true record of a revelation which existed before it.’

13. Helm, Divine Revelation, 70.

14. Ancient Israel was similarly puzzled over its election as God’s
people. In Deuteronomy 7: 7–8a, the best explanation that
they could nd was a kind of tautology: God loved Israel
because God loved Israel.

15. Many of these points are rightly emphasized by Swinburne. See

Revelation, 75–84.

16. I entirely agree with the emphasis placed on this point by

Swinburne (see ibid. 81–4, 113, 119, 178, 183–4.) But I
 believe that he is led to his emphasis on the church for the
wrong reason. He concedes too much to historical-critical
biblical scholarship, and concludes that on historical grounds
alone (that is, apart from authoritative church teaching) there
is little that we can know, for example, of the life of Jesus.
Note this telling remark: ‘The task of discovering some vague
outlines of what Jesus said and did and what happened to him
is not, I suggest, an impossible one’ (p. 105). On the other
hand, in the second edition of the book, Swinburne will take
the slightly stronger stand that we have a reasonable amount
of historical evidence for believing quite a lot of what is said
by and about Jesus in the Gospels.

17. See James Callahan, The Clarity of Scripture (Downers Grove:

InterVarsity Press, 2001).

18. See John Calvin, Corpus Reformatorum (Calvin’s works, ed. W.

Baum, E. Cunitz, and E. Reuss [Braunschweig, 1865]), xxvi.
35, 312, 387.

19. This point is well argued in W. S. Anglin, Free Will and Christian
Faith (Oxford: Oxford University Press, 1990), 187–92.

20. See Mavrodes, Revelation in Religious Belief, 150.

21. See Raymond Brown, The Critical Meaning of the Bible (New
York: Paulist Press, 1981), 7–14.

22. See James Orr, Revelation and Inspiration (Grand Rapids: Baker
Book House, 1969; original edn. London: Duckworth & Co.,
 1910), 164–5, 179–81.

23. See Paul J. Achtemeier, The Inspiration of Scripture

(Philadelphia: Westminster, 1980), 58–74.

24. A variation on this theory would be dictation at the stage of

editing rather than the stage of composing the text. That is,
perhaps the Bible’s human authors supplied both the ideas and
the words but God occasionally intervened to correct mistaken
or misleading expressions.

25. For a discussion of it, see Robert Gnuse, The Authority of the
Bible (New York: Paulist Press, 1985), 42–6.

26. Craig’s Theory is found in his ‘“Men Moved by the Holy Spirit
Spoke from God” (2 Peter 1: 21): A Middle Knowledge
Perspective on Biblical Inspiration’, Philosophia Christi, ser 2,
1/1(1999), 45–82.

27. Counterfactuals can be true or false. According to the standard

Lewis/Stalnaker theory, we test a counterfactual such as, ‘If
Bryant had not been injured, the Lakers would have won the
game’ by asking whether the Lakers do win the game in the
closest possible world to the actual world in which Bryant is
not injured. This analysis is somewhat controversial among
logicians, and things quickly get more complex than this, but
the point is that there do exist procedures for verifying or
falsifying counterfactual claims.

28. Craig, ‘Men Moved by the Holy Spirit’, 72.

29. Orr, Revelation and Inspiration, 177.

30. Achtemeier, The Inspiration of Scripture, 43.

31. Those who a rm robust inspiration deny that the Bible,

properly interpreted, approves of such practices. But this
hermeneutical issue is outside the scope of this chapter.

32. (New York: Oxford University Press, 1987). References to this

book will be placed in parentheses in the text.

33. John Reid’s comment of some half-century ago is appropriate:

‘Now it turns out that the Bible is being recommended, not
because it is an inspired book, but because it is an inspiring
book …This may indeed be all that some people require of
biblical inspiration; but it is an impoverished conception.’ The
Authority of Scripture (New York: Harper & Row, 1957), 168.

34. One of Trembath’s criticisms of book-oriented biblical

inspiration gets nowhere. He says the idea that inspiration is
an intrinsic property of the Bible is wrongheaded because
properties of things normally hold independently of anybody’s
beliefs about them, but only the Christian community believes
in inspiration; ergo, inspiration is not an intrinsic property of
the Bible (p. 116). But it surely cannot be true that a property
can be a belief-independent property of some object only if
everybody accepts it as a property of that object.

35. Nicholas Wolterstor , Divine Discourse: Philosophical Re ections

on the Claim that God Speaks (Cambridge: Cambridge
 University Press, 1995). References to this work will be placed
in parentheses in the text.

36. It should be pointed out that the de nition of biblical

inspiration just suggested might t with various theories of
precisely how the Bible or the biblical writers were inspired.
The fact that the de nition does not entail a commitment to
any such theory and is thus somewhat exible is, I believe, an
asset.

37. James Barr, Holy Scripture: Canon, Authority, Criticism

(Philadelphia: Westminster, 1983), 27.

38. Swinburne, Revelation, 192–3.

 CHAPTER 3

SCIENCE AND RELIGION

DEL RATZSCH

INTRODUCTION

THE natural sciences have profoundly shaped modern life and have
notoriously generated challenges for religious belief—even being
credited by some with having destroyed religion’s rational
defensibility. Most people, however, see both science and religion as
having important truths to tell us, and try to t both into a coherent
world-view. Among that wider group, some see science and religion
as occupying separate, isolated territories, with any alleged con icts
resulting from failure to respect proper boundaries, while others see
varying relationships and legitimate (or illegitimate) interactions
between the two. The competing views arise from a history, of
course—a history widely misconstrued.

Brief History

There is a widespread perception that whatever their present

relation (if any), historically science and religion have continually
been at each other’s throats. That picture is far from accurate—
indeed it is widely acknowledged that Christian theology and
practice played a substantial, positive role in the birth and rise of
modern science.1 The overwhelming majority of sixteenth- and
seventeenth-century scientists were committed Christian believers
who deliberately mined their theology for concepts, principles,
boundary conditions, models, and rational justi cations for their
edgling science.2 But as a science initially unsure of its place and
powers gained con dence and competence, a divide began to
develop between science (increasingly seen as the autonomous,
authoritative voice on nature) and religion/theology (increasingly
seen as the intellectually disconnected, nature-irrelevant authority
only on spiritual and moral matters—if that). That divide, beginning
in the eighteenth century, became increasingly prominent, peaking
in the Positivist movement of the early to mid-twentieth century.

Although the conceptions of reality, of science, of religion, of

philosophy, and of humans undergirding Positivist claims were
subsequently recognized as inadequate and impoverished, a broad
science/religion disjunction still dominates most academic and
much popular discussion. In recent decades, however, a renewed
science/religion discussion has become lively, widespread, and
substantive, with numerous standpoints vociferously advocated.
Resolution is not in the o ng, but visible progress—and the
unmasking of various loud errors—is taking place.
The Relata

Assessing science/religion proposals requires understanding the

relata. One would thus think that the initial step would be to de ne
science and religion. Unfortunately, that’s not going to happen,
because there is no settled, consensus de nition for either, and the
standard candidates all have known problems. For present purposes
I take religion to involve belief in a transcendent supernatural
being(s), plus (typically) closely associated moral codes, ritual
practices, personal/group commitments, convictions concerning
meaning, purpose, value, and post-death conscious existence, all
integrated into an encompassing world-view.3

Science I take to be a deeply empirical project aimed most

fundamentally at understanding and explaining the natural realm,
typically in natural terms. Conceptions of science have varied
widely over the centuries, changing as shortcomings of earlier
conceptions became progressively evident.4 Incorporating lessons of
some of those previous failures, contemporary pictures of science
recognize that theories can be neither generated, rigorously proven,
nor rigorously falsi ed by any quantity of empirical data. It is
further recognized that science does not have a rigid logical
structure and that theories are creatively constructed and not
inferred. Philosophers now almost universally acknowledge that
observation is not passive, and that data to some (often
exaggerated) degree carry undertones of the theories involved in
their generation. Quite a few contemporary philosophers of science
have become convinced that echoes of human gender, political,
class, psychological, and societal structures are sometimes imported
into scienti c concepts as well as into scienti cally essential
metaphors and theories. And few would now deny that theory
evaluation and selection inescapably involve a wide range of
scienti cally essential extra-empirical conceptual resources—e.g.
presuppositions (uniformity of nature, appropriateness of human
cognitive resources, etc.), epistemic values (e.g. fruitfulness,
consistency, explanatory power, empirical adequacy, wider t,
simplicity), concepts (e.g. law, explanation, con rmation, truth),
and so forth. Over all, the trajectory of recent decades is the
growing awareness of the centrality of humanness to science—with
the consequence that the standard perceived dichotomy between
science and religion is increasingly blurred.

Although disagreeing on speci cs, most mainline commentators

argue that, despite the complexities and despite the ineradicable
dependence of science upon human-shaped resources beyond just
empirical data and reason, scienti c results can nonetheless often
rightly claim powerful rational justi cation and epistemic
legitimacy. Rigor, objectivity, and warrant may be less than absolute
and less than often wished and claimed, and science may have
human ngerprints and human DNA traces all over it, but science
can still give us rational reason to accept various scienti c theories
as true or probably true. A tempered or ‘critical’ realism (the
position that science attempts and sometimes succeeds in
uncovering genuine theoretical truth, and that critically vetted
scienti c results are rationally warranted) is still defensible.

PROVISIONAL TYPOLOGIES

Science and religion each have complex hierarchical internal

structures involving multiple, multilevel components, with complex
and varied internal connections between them. Indeed, science and
religion each exhibit so many variations and such complex and
dynamic internal structures that prospects of there being any simple,
unitary account of the relationship(s) between science and religion
seem dim.

Still, it seems worthwhile to try to identify some sort of gross

order. There have been and continue to be numerous proposals
concerning if and how science and religion relate to each other.5
Most fall somewhere into one of three very broad categories:

A. Independence (or separation): Science and religion are both

legitimate, justi able, valuable human pursuits. Each has its own
proper authority and domain (e.g., issues, methods, results), but the
respective domains have no substantive overlap, connections or
interactions. A person can be a competent practitioner of both with
integrity.

B. Con ict (or warfare): Science and religion operate in the same
(or overlapping) domains, make con icting claims and/or demands.
At most one of the two is rationally legitimate, their basic con icts
being ultimately irreconcilable. A person cannot, with competence
and integrity, be a practitioner of both.

C. Dialogue (or interaction, integration): Science and religion are

both legitimate, justi able, valuable human pursuits, and while each
has its own (possibly even overlapping) authority and domain,
science and religion each can and do substantively interact with,
contribute to, and even properly self-correct in the light of the
other, especially in such areas as foundations and interpretative
stances, with apparent con icts ultimately reconcilable.

As indicated, terminology is not uniform among science/religion

scholars, and a number of in uential proposals constitute subtypes
of one (sometimes more) of the above. Brief discussion of major
views follows.

Independence (Separation)

The popular cultural perception is that religion has not fared well in
its contacts with science, and it is commonly suspected that
religion’s problem has resulted from attempts to speak in areas
where it has no rightful authority or competence. But (on
Independence views) were science and religion to stay in their
appropriate domains, there would be no problem. The only issue
would involve locating the relevant boundaries and identifying what
lay within each.
 There are indeed religious issues that appear to have no overlap
with any scienti c issues and vice versa. But Independence does not
seem adequate as a general position. For one thing, as indicated
earlier, not only were science and religion not independent
historically, but had they been, the early history of science might
have been quite di erent—and not in a positive way. For another,
no one knows how to delineate hard boundaries to science (or
religion)—demarcation issues being as intractable as ever.

Furthermore, the history of science contains numerous cases of

fruitful cross-fertilization. Newton, for instance, was emboldened by
the doctrine of divine omnipresence to countenance something very
like the otherwise (at that point) proscribed concept of action at a
distance—in his case, gravitational. And in his Mysterium
Cosmographicum, Kepler said:

[T]here were three things above all for which I sought the cause as
to why it was this way and not another—the number, the
dimensions, and the motions of the orbs. I have dared to carry out
this search because of the beautiful correspondence of the immobile
Sun, the xed stars, and the intermediate space with God the Father,
the Son, and the Holy Spirit.6

On a more general level, science de facto takes the cosmos to be

in relevant respects like a creation—uniform, orderly, intelligible,
predictable, and even beautiful. And science de facto takes scientists
to be in relevant respects like creatures made in God’s image—
rational, creative, knowing, capable of understanding.

And intriguingly, there is growing neurophysiological evidence

that human cognition is not neatly compartmentalized—and that
things typically taken as de nitive of science (reason, perception,
and so forth) and things popularly taken as de nitive of religion
(emotion, faith, and so forth) are inseparable cognitive functions.
(More later.) That is at least suggestive, along the same lines as
Einstein’s remark that ‘The compartments into which human
thought is divided are not so watertight that fundamental progress
in one is a matter of indi erence to the rest.’7

Given these sorts of cross-talk, attempts sharply to separate

science and religion almost inevitably do violence to one or the
other—e.g. reducing religion to subjective value and morality, or
reducing science close to a sterile positivism.

Methodological Naturalism

The currently most in uential supporting thesis for Independence is

methodological naturalism. According to this doctrine, science takes
no position on philosophical naturalism,8 but by its very nature (or
alleged de nition) restricts its methods, presuppositions, conceptual
resources, and results wholly to the realm of nature and must
proceed as if philosophical naturalism were true. As a strategic
recommendation, this type of de facto prohibition might well be a
harmless and indeed practically useful restriction. But that benignity
fades when advocates forget that it is a stipulated prohibition and
take the pronouncements of a science operating under that human
prohibition as the unvarnished truth about nature. There is no
rational reason why the supernatural might not exist and have a
hand in the structure and running of nature at some points. Were
that the case, any science con ned to and driven to purely
‘naturalistic’ theories would go irretrievably wrong on such points.
Thus, to merely assume that a methodological naturalistically
restricted science generates truth would derail science at exactly all
the places that a mistaken assumption of philosophical naturalism
would. The following:

(C1) T is the best theory consistent with philosophical naturalism

and

(C2) T is the best theory

are not interchangeable and require di erent justi cations—unless,

for example, we take philosophical naturalism as de ning the
relevant (or normative) context for science. To forget the
quali cation thus smuggles a substantive and far from innocent
philosophical doctrine into science. And that restriction is not
scienti cally essential. For example, science can be done—and
historically sometimes was done—employing a non-naturalistic
conception of law as regularities in God’s immediate governance of
the cosmos. Such views may even o er the only viable explanation
of the unique logical characteristics of natural laws—i.e. their being
located between material generalizations and necessities, their
support of counter-factuals, etc.9

With all Independence views there could still be ghts over

exactly whose turf speci c matters fell within—even if each side
scrupulously avoided meddling in what it saw as the sovereign
territory of the other. And exactly who is going to adjudicate those
disputes and on what basis?

Complementarity

Complementarity—a subtype of Independence—is the most popular

of the usual science/religion positions. In con ning science and
religion each to its own realm, Complementarity dictates that
scientists and theologians tend their own level and not meddle in
others, thereby precluding nasty science/religion turf wars.
Complementarity also insists on the coordinateness of the truths
arising from both pursuits, seeking a more cohering, uni ed picture
of reality. Advocates also typically accept a completeness principle
for both science and religion agreeing with van Huyssteen that

[A]t its own level science is capable of providing adequate and

complete explanations.10

and that

[A]ll it [the universe] contains is explicable by the natural

sciences.11 Large stretches of science and religion do seem
orthogonal and autonomous. But how—or if—the claims work
generally is problematic.

Boundaries

First, there would have to be a clear demarcation between the

explanatory/causal economies of the two realms in question, as well
as a matching disjunction between the results (and perhaps the
methods, etc.) of the corresponding disciplines. I know of no non-
question-begging case for either. Second, some complementarists
hold that a theistic world-view provides the undergirding
conceptual matrix required by science but that science and religion
diverge past that point. That presents technical di culties
maintaining the required separations, and has consequences for the
possibility of miracles, providence, and other things important to
some religions.

Connections vs. Completeness

Given the precision, order, complexity, and generally staggering

beauty of the cosmos, it seems wildly implausible that the sole
purpose of the cosmos is to generate (as it inarguably does) charred
Camel cigarette butts. The di erent levels of reality thus clearly
have some inter-relevance—or consonance—bearing upon science
and religion. Complementarists thus routinely claim that religion
must ‘take science seriously’.12 But if religion is complete in the
appropriate sense it can presumably in principle arrive at a
complete religious picture purely on its own. Thus, if religion should
pay attention to science, it is evidently for practical reasons—
although if the realms are disjoint and complete, that practicality is
puzzling.

But the typical requirement for science/religion consonance

seems right, although involving di erent levels (as it must) it
constitutes an awkwardness for completeness claims. If consonance
really is a substantive requirement, then there could very well be
speci c religious claims that were not in the requisite way
consonant with speci c (purported) characteristics of the physical
realm. Thus those speci c religious claims could not be held without
violating the consonance requirement. But given the stipulated
completeness, religion is supposed to be complete and autonomous in
its own realm and should thus have the right to make exactly the
‘non-consonant’ claim in question. There would thus be con ict,
contrary to Complementarity. But if there can properly be no such
con ict, consonance is a singularly tenuous demand.

Advocates frequently see such tenuous connections as ‘science

pointing beyond itself’—i.e. raising questions it cannot answer. (For
instance, science can neither explain the ultimate origin of the
natural laws or lawlike structures that it both investigates and uses
in its investigations, nor explain the ultimate boundary conditions of
the cosmos, nor explain the ultimate source of nature’s
intelligibility.) But if there are questions in its own domain that
science cannot answer, then it is not complete. Thus those questions
must fall outside the domain of science. But if science generates
questions in other domains, then it must have some grip on
resources and issues of relevance to that other domain, violating
separation.

Complementarists want separation as a remedy for meddling

hostilities, want completeness within each realm, and also want
both consonance and science’s pointing beyond itself so that science
and religion can be laced into a whole picture whose components
constitute a harmonious unity—not a compacted conceptual
conglomerate or the fragmented compartmentalization of simple
separation. It is not obvious that those are mutually realizable
desiderata. Genuine uni cation will require smudging the boundary
lines and softening the claims of completeness and autonomy. That,
however, would involve moving into the area of various alternative
modes of contact. To those we now turn.

Con ict (Warfare)

Once the barriers are lowered and science and religion are
construed as potential competitors in the same (explanatory) tasks
in the same arena, interaction and con ict become very real—some
would say inevitable—possibilities. There is naturally disagreement
over who gets eviscerated in such con ict, but some cautions are in
order. First, con ict of content requires realist conceptions of both
science and religion. (Realist views in this context are, roughly,
those according to which truths and entities of the relevant sorts do
exist, that discovering truths involving such entities constitutes an
important aim, and that such truths or something near to them are
sometimes actually secured.) Thus, neither non-realist views of
science nor non-realist views of religion will serve the con ict
thesis.13 Second, neither outdated history and philosophy of science
nor uninformed views of religion will do. Much of the perception of
historical con ict is sheer invention. The loci classici for that
perception (for instance Draper’s and White’s books14) are polemics
with ‘history’ manipulated or manufactured to match
preconceptions. And some purveyors of more recent popular screeds
on science and religion have exhibited equal ‘respect’ for historical
fact when it stands in the way of preconception.15 Of course, the
mere fact of science/religion con ict would not reveal which side
was mistaken. Sorting that out would take further work.

Creationist/ID ‘Con ict’ Views

Creationists, Intelligent Design (ID) advocates, the church of

Galileo’s day, etc., are habitually (and polemically) cited as holding
a Con ict thesis, but that is largely misrepresentation. On many
views (e.g. Dialogue) there can be prima facie science/religion
disputes—indeed, even legitimate interaction of nearly any sort
creates the possibility of disagreement. The key di erence is that on
other views such disputes are ultimately legitimately resolvable
whereas Con ict says that they are not—that con icts partially
de ne the relationship.
 But creationists (and many ID advocates) insistently and atly
deny that legitimate science and legitimate religion are in any ultimate
opposition whatever. In contrast to most others, some creationists
do hold that religious/theological matters can not only play
legitimate roles in science but take precedence in resolving apparent
science/religion con icts. But that is a very di erent issue from
endorsing Con ict. Creationists and ID advocates typically claim
that the only real con ict is between religion and an agenda-
twisted, polemical, atheism-fueled faux ‘science’—not real science.

Unfortunately, some, especially among popular creationists, have

endorsed Baconian inductivism, according to which science involves
only strict inductive inferences from pure observational data, have
taken that as the criterion of scienti c legitimacy, and have argued
(correctly) that most contemporary theoretical science fails to meet
Baconian conditions. But given the well-known and insurmountable
di culties with Baconian and allied conceptions of science, even if
parts of contemporary science fail Baconian standards, nothing
much follows.

Naturalist/Atheist Con ict Views

Contentions in this category come in several types, brief discussion

of the more prominent of which follows.16

Foundations
Various commentators claim that science rests upon a
presupposition of philosophical naturalism, note that that contradicts
the belief of most religions in supernaturalism, and conclude that
the very project of science is ‘diametrically opposed’ to any religious
belief worth its salt.17

But that allegation is suspect.18 For one thing, if science operates

in a cosmos structured like a creation, if the historical roots of
science lie in theology, and if the historical justi cation for essential
scienti c presuppositions or stances was theological, then
contradiction between the foundations of science and religion seems
implausible. Nor is it plausible that Newton, Boyle, Faraday and
other major historical gures weren’t quite perceptive enough to
notice their own incoherence (unless, of course, in their religious
‘commitments’ they were just liars, cowards and/or hypocrites—also
wildly implausible).

Deconstructing Religious Rationality

One common ‘con ict’ line attempts to defeat (speci cally:

undercut) the rational propriety of religious belief rather than
directly attacking its truth. This challenge involves:

a. citing purported causes/sources of religious belief;

b. claiming that those causes/sources are aimed at something

other than capturing truth;

then
 c. arguing that such causes/sources, so aimed, cannot provide
rational justi cation for the beliefs generated.

As an historical example, Freud argued that religious belief

represented wish ful llment aimed at comfort and reassurance in a
world where life was threatening and death terrifying. As the
product of processes not reliably aimed at truth (wish ful llment),
religious beliefs would have no rational warrant.

But making a case here requires showing that the alleged source
is the actual source, is not aimed at truth, is thus unreliable, etc.
None is trivial, and Freud failed miserably on the rst at least. More
recently, there have been a number of attempts to defeat the
rational legitimacy of religious belief by citing evolutionary theory
as fully explaining the emergence, and/or the preservation, of
religious beliefs. Explanatory proposals vary. For instance, some
argue that religious belief is/was just an evolutionary spandrel,19
with no special evolutionary upside.20 Others, e.g. David Sloan
Wilson, argue that religion was preserved for its group-selective
advantages,21 while Daniel Dennett argues that religion in the form
of early animism arose out of a Hyperactive Agent Detection Device
(HADD), and served tness by making the welter of experiences of
nature not only cognitively more manageable but even
predictable.22

It is worth noting that all the evolutionary scenarios about

religious belief currently on o er are highly speculative ‘imaginative
illustrations’ (to borrow a term from Darwin). But suppose that one
of them was prehistorically absolutely right. Exactly how is that
supposed to undermine either the truth or the rational legitimacy of
religious belief? That a propensity toward religious beliefs (and
practices) should have arisen by evolutionary means (whether
having direct tness value or not) does not in the slightest imply
that basic religious beliefs are false or irrational, any more than the
fact that a propensity to believe our senses arose by evolutionary
means implies that we should be deeply suspicious of that
propensity.

One can perfectly sensibly, and consistently with everything

science knows, maintain that evolution is the means by which
humans were equipped to achieve recognition of some crucial
religious realities. And if evolution has preserved those propensities,
that strongly suggests that they have some (at least evolutionary)
value. Given evolution, that religious propensities were both selected
for and preserved as near-universals is nicely consistent with what
one might expect were some basic religious outlook true. And the
implied value is atly contrary to what most foes of religion have
stridently proclaimed for, literally, centuries.

Furthermore, if Darwin is right, then natural selection produced

the faculties and mental structures with which we form all other
beliefs, and by which we even pursue science. The governing aim of
natural selection is reproductive success (not theoretical truth) as
contemporary Darwinians acknowledge.23 But if a belief-production
mechanism having some governing aim and purpose other than truth
means that the resultant beliefs are not rational legitimate, then
exactly the same principle poses problems for scienti c beliefs and
for anti-religious arguments produced, as they must be if Darwin is
correct, by the cognitive faculties developed via Darwinian processes
ultimately pointed toward enhancing reproductive tness—not
truth. In fact, Darwin himself worried about his theory undercutting
the mind’s reliability.24 Creation-based religion—with humans,
however produced, re ecting God’s image—may be in less trouble
here than the ‘blind, pitiless indi erence’ of Dawkins-style
naturalism.25

Erosion of Religion—Closing Gaps

One challenge cites gradual historical erosion rather than speci c

decisive episodes. Religion, the story goes, supplied (perhaps then-
reasonable) pre-scienti c explanations, but science has progressively
conquered those areas, evicting religion in the process. Thus, Alfred
North Whitehead:

[F]or over two centuries religion has been on the defensive …The
period has been one of unprecedented intellectual progress…Each
such [advance] has found the religious thinkers unprepared.
Something which has been proclaimed to be [theologically] vital,
has nally, after struggle, distress, and anathema, been modi ed
and otherwise interpreted…The…continuous repetition of this
undigni ed retreat, during many generations, has at last almost
entirely destroyed the intellectual authority of religious thinkers.26

On this view, the bits and pieces that have crumbled away from the
religious conceptual scheme might not individually have been
essential to religious belief, but over the longer haul they constitute
a dismal track record of serial failure.

The most common way of packaging this challenge involves the

‘God-of-the-gaps’ picture. According to this picture, since religion
arose as and remains nothing beyond a primitive attempt at
explaining natural phenomena, believers inevitably attribute any
phenomenon not understood at a given moment to divine activity,
then interpret that phenomenon as empirical evidence for theism. In
short, believers bridge any gaps in our understanding of nature—or
any actual causal gaps in nature—by reference to divine activity.
But, it is claimed, there is a recurring historical pattern of science
hitting upon natural explanations of the relevant phenomena, and as
religious ‘explanations’ are sequentially displaced by science, the
available elds for religious explanations (gaps) continually shrink.
Religion is thus reduced to ghting for its life in rearguard actions
from within whatever gaps happen to be (as yet) unclosed. But even
those temporary hideouts are well within the gunsights of scienti c
inevitability, spelling the inevitable doom of rational religious belief
as its major substantive eld of operation—explanation of natural
phenomena—is wholly taken over by science.
 A variant version is that while science has perhaps not rendered
religious belief completely irrational, it has made explanatory
appeal to any supernatural agency unnecessary—even pointless.27
God is super uous because his primary function is explanatory, and
anything one might claim to explain theistically will ultimately be
covered by alternative scienti c explanations. Perhaps one can still
accept religion if one wants, but religious belief does not provide
resources for any essential explanatory work not equally (indeed,
better) doable by science.

Many religious believers also reject God-of-the-gaps views, but

for di erent reasons. They reject the whole underlying picture of
religion as an explanatory scheme for natural phenomena. They
accept the historical perception that religion has been serially
displaced from earlier gaps, then argue (a) that when science does
produce a natural explanation and religion is forced to retreat,
religion is made to look foolish; (b) that the religious faith of anyone
who rested their belief upon that now-discredited ‘evidence’ is thus
weakened; and (c) that such gap theories involve a derogatory
picture of God not getting things right initially and resorting to
emergency tinkerings to keep things on track. Religion is thus
indeed harmed and it would be vastly better for religion to avoid
such vulnerability altogether by not being construed as in the
business of providing competing explanations to begin with. It
should rather be construed as having a wholly di erent, exclusively
religious and moral content, structure, and purpose.
 Despite their constituting (largely unargued) ritual incantations,
all those objections strike me as seriously weak. First, it should be
recognized that the logic of gap views is impeccable—if there is no
other viable explanation, then divine activity must be the answer.
Furthermore, science has a long track record of overturning its own
positions, but that doesn’t make those who held the earlier,
overturned scienti c views look foolish, lose their trust of science,
or anything of the sort. Thus, the mere fact of exhibiting a history of
having views overturned by science is not alone necessarily
problematic for some doxastic region—religion as well as science.
And perhaps God likes running the cosmos, and deliberately left
dials he gets to turn from time to time. I know of no reason—aside
from speculative supernatural psychology—for thinking that if only
he could have, God would have produced a cosmos that precluded
his special activity in running it.

The ‘history’ isn’t exactly bulletproof either. Religion hasn’t

always been driven into gaps, and has in fact sometimes made
substantive contributions. As mentioned earlier, Newton embraced a
theory of gravitation involving the then-proscribed idea of action at
a distance (short-range repulsion in this case) on the basis of his
belief in God’s omnipresence. Impetus theory, a crucial forerunner
to inertial theories, was defended by Francis of Marchea on the basis
of doctrines connected with the sacraments. Faraday’s eld theory
was connected to his theology of God as creator and sustainer.
Maxwell’s eld equations modeled his views concerning
relationships within the Trinity. Further, religion-friendly aspects of
contemporary cosmic ne-tuning cases cannot be dismissed as easily
as some would like.

More generally, there still are gaps in our scienti c pictures, and
it takes a-historical hubris (or massively undersupported
philosophically avored induction) to claim that science has, does,
can, or ever will explain everything having to do with the cosmos.
Recurrent claims that we nally have in hand all necessary
materials for completing the scienti c picture have just as
recurrently failed. Moreover, Kuhn argued that revolutionary
advances sometimes reopen scienti c issues previously thought to be
settled. Closed gaps may thus be an unstable launching platform for
critiques.

And in any case, an essential presupposition underlying the rst

two cases—that religion is merely explanatory—is seriously
historically and theologically speculative, and controversial to boot.

Comparative track records—rampant success of science vs.

visible stagnation of theology—also require caution. Most scientists
take methodological naturalism as a norm for the acceptability of
scienti c theories. Thus any inadequate theory will by
methodological policy be replaced by some alternative theory that
also meets methodological naturalistic conditions—non-naturalistic
theories being ruled out by at. Thus, only naturalistic theories can
even be candidates for ‘success’. The claim, then, that naturalism has
a monopoly on scienti c success bears some logical resemblance to
that of the ruling party in a one-party country citing its history of
electoral success—where only party members can run—as evidence
of the voters’ high regard.

Speci c Scienti c Results

The simplest challenge science could pose would be inarguable

results of science contradicting fundamental religious beliefs. But
while accepted scienti c results may contradict speci c doctrines of
speci c groups (e.g. young earth views), one cannot seriously hold
that plate tectonics, biological or stellar evolution, relativity, or
quantum mechanics entail the non-existence of God. Obviously, one
could coherently claim that God used evolution, or that the world is
quantum mechanical because that is the way God wanted it. Since
those are coherent claims, then those theories do not contradict the
existence of God.

a. Inconsistency. Mere inconsistency does not show religion

mistaken unless science is given precedence in such con icts.
Showing that precedence is not trivial. And inconsistency is not
automatically fatal even with in science itself. It has been known for
decades that general relativity and quantum mechanics are in
tension with each other, but—rightly—no one seriously proposes
abandoning either. Furthermore, the clash of ideas is often how
truth emerges.

b. Randomness. It has also been claimed that the key role played
by randomness in such central theories as evolution and quantum
mechanics undermines the religiously central doctrine that the
cosmos, life, and humans are products of deliberate purpose. That
claim does not stand up to scrutiny,28 and in fact (as will emerge
later) the very randomness alleged to be religiously problematic can
serve some theologically important functions.

Styles of Rationality

Many challenges to the rationality of religious belief assume that

rational justi cation must conform to a ‘scienti c’ model. Though
that assumption is far from evident, some believe that religious
belief does (or can) meet that standard.29 Others have argued that
there is no single canon of rationality—that di erent domains have
their own characteristic rationality. If either contention is right,
these objections fail, as they would were the proposed ‘scienti c’
model mistaken.

Very recently, some scholars have argued that virtually none of

our fundamental, common sense, life-governing beliefs are
generated by argumentation, nor are they mere provisional
explanatory hypotheses. A belief that one’s spouse loves one is
neither a hypothesis to explain puzzling behaviors, an induction,
something to be empirically tested, a candidate for stringent
falsi cation e orts, nor an empty irrelevancy. We simply do not
acquire or justify beliefs that those around us have minds or that
there has been a past via testing, con rmation, and argumentation.
Nor could we—as the long history of failed philosophical attempts
attests.

Yet such beliefs are surely sensible if any of our non-trivial

beliefs are, and non-accidentally track the truth if any do. Rational
justi cation here must have some di erent, deeper character, and
arti cially limiting acceptable sources to explanatory hypotheses,
scienti c argumentation, and the like, does serious violence to
human rationality. Some have suggested analogously that arti cially
limiting rational justi cation for core religious beliefs to procedures
(e.g. foundationalist procedures) that are demonstrably inadequate
even for ordinary experience is equally misguided.30

Independence and Con ict can each point to historical and

current episodes that illustrate their respective patterns. But despite
these apparent successes, each is inadequate as a general approach,
and formally inadequate for broader conceptual reasons. We must,
then, look further.

Dialogue (Interaction, Integration)

Various Dialogue advocates argue that science and religion share

normative patterns of inference, theory evaluation, evidence
assessment, explanatory styles, and the like. On some views, those
patterns are specialized, on others simply honed instances of
ordinary, everyday patterns. Nancey Murphy argues that
‘philosophy of science …provides the de nitive account of the
canons of probabilistic reasoning’ but that ‘(potentially at least)
theology is methodologically indistinguishable from the sciences’.31
Others (e.g. Mikael Stenmark) argue that scienti c rationality is
parasitic on deeper structures of rationality. Thus science does not
de ne canons of reason.32

Many Dialogue advocates argue that religion supports science by

providing conceptual, presuppositional, meaning-assigning, and
practical milieus within which science thrives and ts better than in
any non-theistic foundation. Some see science reciprocating,
o ering discoveries (e.g. ne tuning) best explained by theism and
thus providing indirect evidence for theism.

Some argue that science provides models for various religious

matters. For instance, Peacocke and Clayton argue that while
science’s picture of nature as a ‘seamless web’ has undercut
theology’s traditional interventionist (or ‘interfering’) view of God’s
activity in the world, science’s picture of dynamic emergence within
nature (e.g. consciousness emerging from neural complexity) o ers
a new model of God’s immanent activity33—a model with less need
for a separate supernatural realm (and hence for an ‘outside
interventionist’ model of God’s action in the world) and which in
turn ts best with panentheistic theology.34 (This view involves an
interesting partial reprise of some earlier views in that it sees ‘even
inner-worldly causality as (in at least some sense) a manifestation of
divine agency…there [being] no qualitative or ontological di erence
between the regularity of natural law and the intentionality of
special divine actions’.35) Similarly, evolutionary theory provided a
primary model (and perhaps much of the impetus) for process
theology. Some (e.g. Stump and Kretzmann) have argued that
relativistic conceptions of simultaneity provide a model for what
God’s being outside time would look like.36 And yet others (e.g.
Torrance, Pannenberg) have tried to connect relativity and eld
theories to the speci cally Christian doctrine of the Incarnation.

Although most Dialogue advocates decry scientism, the de facto

position of most is that science can have decisive consequences for
theological positions but that attempts to draw from religion or
theology similar consequences for properly constituted scienti c
positions are ‘interfering’ with science.37 Despite the science-heavy
tilt of most advocates, the rejection of scientism is genuine—science
(as noted earlier) being seen as incomplete in inevitably raising
questions which it cannot answer, and which only religion has the
resources to address.38

There are important exceptions to that tilt tendency. Murphy

argues that there is ‘no reason in principle why theories originating
in theology cannot be included as auxiliary hypotheses in scienti c
research programs and vice versa’.39 More signi cantly, she argues
that ‘We sometimes have to correct our theology as science
advances…. But sometimes theology must correct science as well.’40
On a parallel tack, Wolfhart Pannenberg believes that theology
establishes that ‘If God…is the creator of the universe, then it is not
possible to understand fully…the process of nature without
reference to God.’41 Of course, science is now attempting that
‘impossibility’ and thus for its own sake needs redirection.

Dialogue views see science and religion as each maintaining its

own recognizable identity, but with a permeable (or semi-
permeable) interface or shared stratum on some level(s) rather than
a sealed boundary between them. Such views, it seems to me, point
in the right direction, but there are deeper interfusions between
science and religion, which Dialogue proposals do not adequately
acknowledge. (More shortly.)

One other Dialogue view tries to trace connections to a ner-

grained level than most others. On this view, the fundamental
indeterminacy of quantum mechanics means that the future is
genuinely open, with multiple divergent, live, unknown futures. But
God deliberately created the world that way. Thus, just as the
clockwork Newtonian universe was taken as indicating a creator
favoring mathematical precision, sharp distinctions, overriding
micro-order, autonomous and static identities, and in nite
predictability, some Dialogue advocates argue that the quantum
world suggests a creator favoring openness, freedom, fuzzy edges,
dynamic uidity, interconnectedness, novelty, innovation,
emergence, surprises. This openness, it is argued, allows for free will
and also suggests new ways of conceptualizing God’s
immanentness.42

A number of authors (e.g. Polkinghorne, Murphy) have sought to

locate some divine activity within the open causal interstices of
quantum phenomena—seeing God as in uencing otherwise
indeterminate quantum events in directions leading toward desired
ends.43 While this idea would allow direct, special divine action in
the world, such action need not violate probabilistic natural law and
need not even be scienti cally determinable. These are seen as
virtues, because they avoid the objectionable (to many) idea of God
‘interfering’ in nature (a dislike often related to God-of-the-gaps
worries noted earlier). And even many who don’t object to
intervention as such, nonetheless dislike the idea of God’s action in
nature being so ‘heavy-handed’ that even something so lacking in
religious sensitivities as science can detect it.44

Those objections do not seem particularly fatal, meaning that the

‘interstitial’ view of this sort of divine action remains a live
possibility.

Deeper Integration

Some commentators have argued that the connections between

science and religion run even more profoundly deep—arguing that
the two constitute an indissoluble, intermingled unity. Such
philosophical or religious positions as coherentism, pantheism,
process theology, and panentheism could quite organically
accommodate such a view, but prominent advocates of those
positions generally adopt a less extreme version of the Dialogue
view. (Note that panentheists Clayton and Peacocke gured into the
previous section.)
Some Candidates

Among the scholars who suggest this deeper Integration picture are
theologian Thomas Torrance and philosopher Roy Clouser. I shall
brie y describe Clouser’s view in order to provide a glimpse of what
such a theory might look like.

According to Clouser, any belief in something as divine is a

religious belief, and to believe that something is divine just is to
hold that thing to have unconditional, non-dependent existence. But
all theories, according to Clouser, substantively embody
presuppositions about the nature of reality (e.g. what kinds of things
are real and upon which explanations must be grounded) and thus
can’t avoid presupposing something or other to have that
unconditional, non-dependent status.

That means, says Clouser, that every theory (scienti c,

mathematical, or otherwise) carries the freight of some religious
belief concerning a divinity. Further, every theory is regulated by
some presupposition concerning what is divine, and every scienti c
theory and concept has its content bounded and colored by such
divinity presuppositions. Those presuppositions play a ‘pivotal’ role
in the construction, assessment, acceptance, or rejection of theories.
The e ects are pervasive, having implications for positions
concerning ‘human nature, destiny, values, happiness and so on’.45
This does not necessarily make scienti c theories themselves
components of religion proper, but it means that every scienti c
explanatory theory is unavoidably and substantively a ected by
some religious (divinity) belief.46

Clouser further contends that attributing unconditional non-

dependent existence to anything other than God leads inevitably to
‘reductionistic’ theories (in this case, theories restricted to
conceptual resources too impoverished to encompass the relevant
truth seven in principle), and that any such will be partially (often
subtly) false, since its internal religious shape will embody
something other than the non-reductionistic truth that God alone is
divine. Thus, the science of anyone not committed to God’s being
the only unconditional, non-dependent existent will be, strictly
speaking, false. In fact, Clouser argues that similar conclusions hold
for all truth, knowledge, and even concepts.47 Thus religious
believers and non-believers cannot share even quite the same
scienti c concepts.

Although we must recognize that intuitive conceptual boundaries

are permeable, things get pushed a bit far here. I am suspicious of
de nitions that entail that belief that numbers and propositions
exist necessarily is a religious belief. Nor does it seem to me that
materialists do not truly and fully understand that the earth orbits
the sun because their concept of the earth has a deeply buried,
underlying intuition that the earth is composed of material
substance which does not ultimately depend on God for its
existence.48
Further Dialogue/Integration Exploration

Most philosophers of science believe that science cannot function

without some underlying de facto metaphysical principles that are
not consequences of science itself. Given science’s contingency,
there are a variety of broader matrices into which a science can be
embedded. But no complete metaphysical system is neutral with
respect to philosophical naturalism and theism, and it is not clear
how that non-neutrality is to be prevented from seeping into
science. The metaphysical foundations contribute something to
science—otherwise they would not be essential—and that
contribution is not merely in providing a generic, content-irrelevant
platform for science to sit on. The consequences of science’s
fundamental presuppositions (or predispositions) for boundaries and
initial conditions could provide a signi cant connection between the
content of religious belief and the content of science. And
theological developments that a ected that content could in
principle generate ripples into science.

Moreover, aspects of deeper levels can seep into other levels of

the scienti c hierarchy, and just as observational data are (widely
taken to be) theory-laden, theories in this circumstance may be
metaphysics- or even theology-laden.49 And, as noted, science can
function only in a world like a creation—coherent, uniform,
intelligible, and so on. If, as is sometimes argued, structure is in
subtle part content, then a science that requires a creation-like
structure may have subtle creation-like content.
 There are other possible factors as well. Human natural science
is con ned ultimately to humanly-accessible concepts and
procedures. In creating and constructing theories, the concepts
employed are nearly always imported or assembled from descriptive
and theoretical contexts from other areas of cognition and
experience. Typically, such imports drag more with them than the
phenomenon speci cally warrants. And where our human concepts
don’t quite map exactly onto reality (think quantum phenomena),
we must do the best we can with the concepts we can manage,
applied partially metaphorically.50 In fact, Mary Hesse has argued
(correctly, I think) that theoretical explanations just are ‘metaphoric
redescriptions’ of relevant phenomena.51

Historically, theology has provided some of those concepts

employed metaphorically. It has been argued more recently that
conceptual imports into science include gender, political, class,
psychological, and societal structures. If importation and
employment of metaphoric content are cognitive and practical
necessities, then scienti c theories and concepts will inevitably be
infused with humanness, and the standard prohibitions—meant
originally to sieve out religion—aren’t going to work.

There are additional facets of science that soften its (earlier-

perceived) contours, making it less inconceivable that religion plays
some role in science proper. Perception, recall, is an active process,
and there are no obvious boundaries concerning what sorts of
factors can a ect (however subtly) the shape of perceptual content.
Values and value judgments play an ineliminable role in the
sciences. Science pursues intelligibility and explanation, but our
recognition of both is in part via a cognitive phenomenology, as a
particular feel, a particular seeming. Even our most rigorous reason
—mathematics, logic—rests ultimately upon our human intuitions
concerning deductiveness and entailment, on involuntarily
embraced axioms, etc.

More surprisingly, the neurophysiological underpinnings of our

cognition simply do not respect the dichotomies we try to impose.
Recent research indicates that some of the very cognitive processes
by which we judge beauty are also among those by which we judge
truth.52 And on a number of levels, reason and emotion function as a
unit. As one example, various structures in the brain (e.g. tonically
active neuron sites in the caudate) directly integrate reason inputs
(from the prefrontal cortex) with emotion inputs (from the
amygdala).53 In some of those structures, integration occurs prior to
our conscious access to the process. One upshot of all this is that
certain sorts of rationality are compromised in the absence of
emotion.

Antonio Damasio and others suspect that that pattern applies to

scienti c reasoning and mathematics.54 Paul Thagard has argued
that one important factor in scienti c theory justi cation is whether
acceptance of the theory maximizes coherence—including emotional
coherence.55 And Christopher Hookway has argued that felt
emotional responses are one way that unarticulated epistemic
evaluative criteria present themselves to us cognitively, and that
‘e ective epistemic evaluation could turn out to be impossible
without …appropriate emotional responses’ (my emphasis).56

Thus, some things often seen as closely linked to religion operate

deep within—indeed, constitute operative components of—cognitive
processes often seen as closely tied to science. Still deeper, our very
selfhood may be involved. Physicist Sir Denys Haig Wilkinson asks,
‘How do we then choose between alternative scienti c hypotheses
when we have used up all our scienti c criteria? We are, of course,
left face to face with ourselves. The only remaining criterion is what
seems right to us …in the deepest seat of human feeling …’57

So values, intuitions, seemings, a sense of beauty, emotions, the

deepest seat of human feeling—all are built into science. And into
reason itself. Again none of them—much less Einstein’s ‘rapturous
amazement at the harmony of the laws of nature’—are strangers to
the religious domain of the cognitive landscape. And if the reality
science studies is an integrated creation, the infusion nestled within
any wholly successful science may well be theistic.

WIDER PICTURES

Each proposal above typically works in some cases but contorts

other cases when pushed as universal. Con ict views are often
agenda-driven and fail historical and philosophical scrutiny.
Independence views often deny the real connectedness at various
points, miss some basic human epistemic facts, and often involve
impoverished views of religion and/or outdated conceptions of
science. The Complementarity subtype represents an improvement,
but does not give adequate weight to the depth of some connections,
and often involves untenable pictures of both science and religion as
having independent, substantive, and complete takes on everything
in the natural realm. Dialogue gets a lot right, but often tilts toward
science, limits the scope and level of interpenetration, overestimates
science’s scope and autonomy, and (like Complementarity) is often
accompanied by nearly pathological fear of God-of-the-gaps. The
Deep Integration subtype of Dialogue alone recognizes the depth of
some crucial connections, but often overshoots the mark, and
sometimes ignores the fact that religion really is irrelevant to some
scienti c contents and vice versa.

What procedures, principles, and criteria can we trust in trying

to put science and religion together? We need a perspective broader
than any of the usual candidates and ultimately broader than just
science/religion interactions. Most humans ultimately seek a world-
view that is whole, uni ed, and satisfying—one that includes science,
religion, and an organic relationship between them. We look for
inclusive unity because few could take rejection of science seriously,
yet most believe that science alone cannot provide wholeness and
satisfyingness—either intellectual or deeper.58 And human
judgments of satisfactoriness are bedrock even for science,
mathematics, and reason itself. There being no escaping that, we
had best learn to appropriate and employ it epistemically
properly.59

With scienti c theory, satisfactoriness has to do ultimately with

certain sorts of theory-generated, science-relevant successes—making
intelligible, unifying, explaining, predicting, and so on, relevant
judgments and the standards themselves being embedded within a
human-tinged conceptual, philosophical, and praxical matrix. A
theory might be propped up arti cially in the short term, but any
theory or discipline that fails such success tests is not stable in the
long term. With religion, the same thing holds. Any religion that
fails to a ord religiously-relevant success—making intelligible,
unifying, explaining, generating social coherence, meaning, inner
peace, hope, charity, love, gratitude, awe, and human ourishing—
may be propped up arti cially in the short term, but is not viable in
the long term.

Viable world-views must achieve world-view-relevant successes.

What those are is not wholly clear, but they must include something
at least like the science-relevant and religion-relevant successes
noted above. And any world-view that claims that its science-
relevant successes demonstrate the impossibility of realizing
important religion-relevant aims has failed. Any world-view that
claims that its religion-relevant successes demonstrate the
impossibility or impropriety of realizing important science-relevant
aims has likewise failed. And any world-view that insists that the
two are disconnected or irreconcilable has failed, having turned its
back on some of the deeper drives of both science and religion—
coherence, uni cation, system, wholeness.60 Peacocke’s remark that
in our search for intelligibility and meaning science and religion are
now ‘inextricably interlocked…in the common human enterprise of
seeking both’61 is spot on.

The relevant successes are not just propositional. Satisfyingness

is not merely, for instance, an axiomatic structure, even with respect
to standard evaluative criteria for science—e.g. explanatory power.
And some contemporary thinkers cite even broader than usual non-
epistemic criteria for evaluating theory legitimacy—considerations
having to do with human wellbeing, global human solidarity,
democracy, social justice, equality, ecological concerns, and other
‘progressive interests’. Marxists, many postmodernists, some
feminists, and even some Darwinians fall into this category.

Although some of that involves (to use Darwin’s wonderful

phrase) ‘mental rioting’, broad, not very well-behaved criteria
beyond propositionally linked considerations play substantive and
legitimate roles in both religion and science. Criteria for
science/religion interaction models and criteria will likely be at
least as broad as that. And still other, farther- ung constraints
involving an array of facets of human life will likely sift down from
any world-view matrix in which that relationship is embedded.

CONCLUSION
All this inter-infusion conjoined with the lack of absolute boundaries
between science and non-science means that science/religion issues
will not be the clean, simple, no-strings-dangling phenomenon many
would like them to be. There is unlikely to be any single elegant, all-
encompassing solution—at least any available to us humans. Just as
there may be aspects of nature forever beyond our cognitive
capacities, just as there may be religious facts forever beyond our
understanding, some facets of the actual relationship between
science and religion may be forever beyond us as well. There may
always be tensions and loose ends. But that is true in nearly every
human endeavor—whether involving science, religion, theology,
philosophy, or world-views. Indeed, given human history,
propensities, and nitude, the slick tidiness of any proposed
science/religion resolution (or dissolution) is very probably
legitimate cause for epistemological alarm.

NOTES

I thank colleagues in the Calvin College Philosophy

Department, and David van Baak of the Calvin Physics
Department, as well as Michael Rea and Thomas Flint for their
often disquieting and annoyingly correct criticisms.

1. A wide variety of prominent historians—R. Hooykaas, Eugene

Klaarrens, and John Hedley Brooke—have developed this
theme, which goes back at least to Collingwood.
2. The literature is huge, involving Newton, Kepler, Boyle,
Descartes, Maxwell, Faraday, Mersenne, Von Helmont, and
others.

3. That is, naturally, controversial. Some (e.g. the Wittgensteinian

D. Z. Phillips) have challenged the essentiality of belief to
religious faith, advancing a type of anti-realism according to
which religious ‘belief’ involves minimal or no (theistic)
metaphysical commitments or claims, but is rather a ‘form of
life’. Others argue that e.g. Buddhism is a religion although
some forms of it involve no belief in the supernatural. While
acknowledging such positions, I am restricting the present
discussion to theistic religions.

4. Nearly any good introduction to philosophy of science will

contain discussions of various past failures.

5. Many are descendants of Ian Barbour’s typology developed in

e.g. Religion in an Age of Science (San Francisco:
HarperSanFrancisco, 1990).

6. For further historical discussion see e.g. John Hedley Brooke,

Science and Religion: Some Historical Perspectives (New York:
Cambridge University Press, 1991).

7. Max Jammer, Einstein and Religion (Princeton: Princeton

University Press, 1999), 156.

8. Philosophical naturalism is the view that the natural realm is all

that exists—that there is no supernatural realm.
 9. See my ‘Nomo (theo)logical Necessity’, Faith and Philosophy 4
(1987), 383–402. Reprinted in Michael Beaty (ed.), Christian
Theism and the Problems of Philosophy (Notre Dame: University
of Notre Dame Press, 1990), 184–207.

10. J. Wentzel van Huyssteen employs this phrase more than once
in Duet or Duel: Theology and Science in a Postmodern World
(Harrisburg: Trinity), e.g. 56, 77. Many others, e.g. Alan G.
Padgett, Science and the Study of God (Grand Rapids:
Eerdmans, 2003), 81, concur.

11. Ibid. e.g. 55, 75.

12. e.g. van Huyssteen, Duet or Duel, 57, 78.

13. e.g. instrumentalist and some social constructivist views.

14. J. W. Draper, History of the Con ict Between Religion and Science
(1875), and A. D. White, History of the Warfare of Science with
Theology in Christendom (1896).

15. The Galileo, Scopes, Bruno, and other frequently cited cases
contain in their currently popular versions vast quantities of
invention.

16. Much of this section derives from my ‘Demise of Religion:

Greatly Exaggerated Reports from the Science/Religion “Wars”
in Michael Peterson and Raymond Van Arragon (eds.),
Contemporary Debates in the Philosophy of Religion (Oxford:
Blackwell, 2004), 72–87.
17. Norman F. and Lucia K. B. Hall, ‘Is the War between Science
and Religion Over?’, The Humanist (May/June 1986), 26–8.

18. Although I will not pursue the issue here, some argue that
there are—or can be—non-supernaturalistic religions.
Buddhism is sometimes cited as an example, and some
philosophers argue for the possibility of an ‘anthropic theism’
involving a ‘deity’ who while powerful, knowing, etc., is
wholly natural (see e.g. Peter Forrest, God without the
Supernatural: A Defense of Scienti c Theism (Ithaca, NY: Cornell,
1996). However, I am restricting discussion to
supernaturalistic systems.

19. In the evolutionary context, a spandrel is a mere by-product of

the evolutionary production of some characteristic actually
selected for (which may itself later be coopted for some
signi cant biological function).

20. Scott Atran, In Gods We Trust (New York: Oxford University

Press, 2002), is often read this way.

21. Darwin’s Cathedral (Chicago: University of Chicago Press,

2002).

22. Breaking the Spell (New York: Viking, 2006), 117.

23. Patricia Churchland, ‘Epistemology in the Age of Neuroscience’,

Journal of Philosophy 84 (1987), 544–53 at 548–9.

24. See his 3 July 1881 letter to William Graham, in The Life and
Letters of Charles Darwin, ed. Francis Darwin (New York: n.p.,
 1889).

25. Alvin Plantinga has precipitated quite a debate on this issue.

See James Beilby (ed.), Naturalism Defeated? (Ithaca: Cornell
University Press, 2002).

26. ‘Religion and Science’, in Science and the Modern World

(Cambridge: Cambridge University Press, 1925).

27. Steven Weinberg, ‘A Designer Universe?’, New York Review of

Books (21 October 1999), 48.

28. See my ‘Design, Chance, and Theistic Evolution’, in W.

Dembski (ed.), Mere Creation (Downers Grove: InterVarsity
Press, 1998), 289–312, and my ‘Saturation, World Ensembles
and Design’, Faith and Philosophy 22 (2005), Special Issue;
‘Proceedings of the Russian-Anglo-American Conference on
Cosmology and Theology, Notre Dame’, 667–86.

29. See e.g. Nancey Murphy, Theology in the Age of Scienti c

Reasoning (Ithaca: Cornell University Press, 1990), and
Michael Banner, The Justi cation of Science and the Rationality
of Religious Belief (Oxford: Clarendon, 1990).

30. Among these are a number of ‘Reformed Epistemologists’,

including Alvin Plantinga, Nicholas Wolterstor , and Kelly
Clark, among others.

31. Murphy, Theology, 99, 198. Many others argue for a

commonality of inference types, canons of reasonableness, etc.
32. See e.g. Mikael Stenmark How to Relate Science and Religion
(Grand Rapids: Eerdmans, 2004), ch. 5.

33. See Philip Clayton and Arthur Peacocke (eds.), In Whom We

Live and Move and Have Our Being (Grand Rapids: Eerdmans,
2004), e.g. pp. xix-xx (Peacocke) and 87 (Clayton).

34. Panentheists disagree about what the view is. The basic
description is: ‘the world is in God, but does not exhaust God’.
But the ‘in’ is metaphorical, with no agreement on what the
metaphor means. In discussing God’s relation to and action in
the world, panentheists often use the analogy of the world
being ‘God’s body’—God’s action in the world being analogous
to the human mind’s relation to the human body.

35. P. Clayton, ‘The Case for Christian Panentheism’, Dialog 37

(Summer 1998), 201–8 at 206.

36. See e.g. Eleonore Stump and Norman Kretzmann, ‘Eternity’,

Journal of Philosophy, 78 (1991), 429–58.

37. John Haught, Science and Religion, e.g. 4, 18. Others, including
e.g. Clayton, Polking-horne, van Huyssteen, and Peacocke,
often sound like this.

38. Nancey Murphy, Reconciling Theology and Science (Kitchener:

Pandora, 1997), e.g. ch. 1 pp. 46, 67, 79, et passim. Many
others (e.g. Peacocke, 21; J. Polkinghorne, Reason and Reality
(Philadelphia: Trinity, 1991), 75; Serious Talk (Philadelphia:
Trinity, 1996), 7, T. F. Torrance, The Christian Frame of Mind
 (Edinburgh: Handsel, 1985), 77–8, van Huyssteen, Duet or
Duel, 77) make similar claims.

39. See Murphy, Reconciling Theology and Science.

40. Ibid. 4.

41. See e.g. W. Pannenberg, Toward a Theology of Nature: Essays on

Science and Faith (Philadelphia: Westminster/Knox, 1999), 17
., 30–1, et passim.

42. There are obvious connections to ‘open theism’. Kelly James

Clark describes open theism thus: According to open theists,
God changes, cannot do certain things, is dependent on
creatures for emotional states and will, and genuinely su ers
along with his creation. More particularly, God does not know
future contingents (especially those involving libertarian free
will). All of this is taken to mean that the future is open both
to God and to human beings, who work in partnership to carve
out an unforeseen but hoped for future.

43. e.g. Serious Talk, 53–4;id., Quarks, Chaos, and Christianity (New
York: Crossroads, 1994), 71. Polkinghorne in some places
seems a bit ambivalent about this view.

44. Christian critics of ID routinely assert that nature is in fact

designed, but that its designedness is not a scienti cally
determinable matter.

45. Knowing With the Heart (Downers Grove: InterVarsity, 1999),

41.
46. See e.g. ibid. 29.

47. R. Clouser, ‘Prospects for Theistic Science’, in Perspectives on

Science and Christian Faith 58 (2006), 2–15 at 15 n. 25.

48. Note that on e.g. coherentist views, Independence advocates

are mistaken in thinking that there is no con ict if both sides
respect proper boundaries. Given the infusion of religious (or
anti-religious) foundations into everything, con ict between
believing and non- (or un-)believing conceptual systems would
be global and pervasive.

49. Integration via the same foundational metaphysics in uencing

both science and religion is the basic idea underpinning
Padgett’s mutuality model (Science and the Study of God).

50. Mary Midgley has done particularly interesting work in this

area. See e.g. her Science as Salvation (New York: Routledge,
1994).

51. See her Revolutions and Reconstructions in the Philosophy of

Science (Bloomington: Indiana University Press, 1980).

52. Norbert Schwarz, ‘On Judgments of Truth and Beauty’,

Daedalus 135 (2006), 136.

53. See e.g. J. M. Schwartz and S. Begley, The Mind and the Brain
(New York: Regan, 2002), 67 ., and Pinker, 371 ., for
discussion.

54. Damasio has written a number of interesting books dealing

with this and related topics.
55. See e.g. his ‘The Passionate Scientist: Emotion in Scienti c
Cognition’, in Peter Carruthers, Stephen Stich, and Michael
Siegal (eds.), The Cognitive Basis of Science (Cambridge:
Cambridge University Press, 2002), 235–50. See also Gerald
Clore and Karen Gasper, ‘Feeling is Believing: Some A ective
In uences on Belief in Nico Frijda, Antony S. R. Manstead, and
Sacha Bem, Emotions and Belief: How Feeling In uences Thought
(Cambridge: Cambridge, 2000), esp. 10–44, 24 ., and Eddie
Herman-Jones, ‘A Cognitive Dissonance Theory Perspective on
the Role of Emotion in the Maintenance and Change of Belief
and Attitude’, in Frijda, Manstead and Bem, ibid., 185–211.

56. ‘Emotions and Epistemic Evaluations’, in Carruthers, Stich, and

Siegal (eds.), Cognitive Basis, 251–62 at 257. A related position
is taken in Clore and Gasper, ‘Feeling is Believing’.

57. Sir Denys Haig Wilkinson, ‘The Quarks and Captain Ahab’,
Schi Memorial Lecture, Stanford, 1977.

58. There are exceptions on both ends. For the former see e.g.
Andrew Pickering, in Constructing Quarks (Chicago: University
of Chicago Press, 1984), and for the latter see e.g. Richard
Dawkins, Peter Atkins, et al.

59. That requires presumptive acceptance that our cognitive

faculties are basically reliable, but any epistemological hope
requires that, and the only workable way of securing that may
be a doctrine of creation. Stenmark’s theory of presumptionism
is useful here (see Stenmark, How to Relate Science and
 Religion, p. 90 et passim), as are Alvin Plantinga’s works on
Reformed Epistemology.

60. Imre Lakatos, Larry Laudan, et al. have led important parts of
the charge here.

61. Theology for a Scienti c Age, 5

 CHAPTER 4

THEOLOGY AND MYSTERY

WILLIAM J. WAINWRIGHT

CHRISTIANITY’S critics have often accused it of mystery mongering.

Hume, for example, says that

all popular theology, especially the scholastic, has a kind of appetite

for absurdity and contradiction. If that theology went not beyond
reason and common sense, her doctrines would appear too easy and
familiar. Amazement must of necessity be raised: Mystery a ected:
Darkness and obscurity sought after: And a foundation of merit
a orded to the devout votaries, who desire an opportunity of
subduing their rebellious reason, by the belief of the most
unintelligible sophisms.1

And John Toland asserts that Christian theologians and priests have
gone even further than the hierophants of the ancient mystery cults.
The latter swore their initiates to secrecy but their mysteries were
intelligible in themselves. Only Christians dared maintain that their
doctrines were mysterious in a more radical sense, ‘that is,
inconceivable in themselves, however clearly revealed’.2
 Hume’s and Toland’s explanations of this phenomenon di er.3
Whatever one thinks of their explanations, however, there is little
doubt that the appeal to, and adoration of, mystery is a
characteristic feature of much Christian thought and practice. The
Pseudo-Dionysius, for example, begins his Mystical Theology by
asking the Trinity to guide him to the ‘most exalted’ and hidden
secrets of Scripture ‘which exceedeth light and more than exceedeth
knowledge, where…the mysteries of heavenly truth lie hidden in the
dazzling obscurity of the secret silence, outshining all brilliance with
the intensity of their darkness’.4

Nor are themes like these peculiar to Christian mystics and

mystical theologians. They are commonplace in the Church Fathers
and in a number of later Christian theologians. Two examples will
su ce.

Consider rst John Chrysostom. St Paul said,

‘The Lord…dwells in unapproachable light.’ And pay heed to the

accuracy with which Paul speaks…He does not say: ‘Who dwells in
incomprehensible light,’ but: ‘in unapproachable light,’ and this is
much stronger than ‘incomprehensible.’ A thing is said to be
‘incomprehensible’ when those who seek after it fail to comprehend
it, even after they have searched and sought to understand it. But it
does not elude all inquiry and questioning. A thing is
unapproachable which, from the start, cannot be investigated nor
can anyone come near to it.5
Yet,

suppose…we forget Paul and the prophets for the moment, [and]
mount up to the heavens…Do you think that the angels in heaven
talk over and ask each other questions about the divine essence? By
no means! What are the angels doing? They give glory to God, they
adore him, they chant without ceasing their triumphal and mystical
hymns with a deep feeling of religious awe. Some sing: ‘Glory to
God in the highest;’ the seraphim chant: ‘Holy, holy, holy,’ and they
turn away their eyes because they cannot endure God’s presence as
he comes down to adapt himself to them in
condescension. (Chrysostom 65–6)

My second example is from Jonathan Edwards, who concludes a

philosophically sophisticated explication of the Trinity by saying: ‘I
don’t pretend fully to explain how these things are, and I am
sensible a hundred other objections may be made, and puzzling
doubts and questions raised, that I can’t solve. I am far from
pretending to explain the Trinity so as to render it no longer a
mystery’, or ‘asserting that [my account of the Trinity is] any
explication of this mystery that unfolds and removes the
mysteriousness and incomprehensibleness of it: for I am sensible
that however, by what has been said, some di culties are lessened,
others that are new appear; and the number of those things that
appear mysterious, wonderful, and incomprehensible are increased
by it’.6
 As these examples illustrate, both Christians themselves and
their critics have historically thought that the concept of mystery is
central to Christian re ection and Christian worship. It is initially
surprising, then, to nd that the indices of four important recent
reference works contain few if any references to mystery. None are
found in the index to Philip Quinn and Charles Taliaferro’s A
Companion to the Philosophy of Religion, for example, and the indices
to my The Oxford Handbook of Philosophy of Religion and to William
Mann’s The Blackwell Guide to the Philosophy of Religion contain only
one each. What is perhaps most surprising is that the only reference
to mystery in Peter Byrne and Leslie Houlden’s massive Companion
Encyclopedia of Theology is to a discussion of Rudolf Otto’s numinous
experience in my article on ‘Religious Experience and Language’ in
that volume. What explains this?

Partly, I think, a not unreasonable fear of obfuscation—a

suspicion that appeals to mystery can be an excuse for avoiding
hard thought and a justi cation for obscurantism and superstition.

Part of the explanation may also be a suspicion that the

apophatic tradition that fuels much classical discussion of mystery
isn’t really Christian or is not Christian enough.7 But this, I believe,
is a mistake. Even ‘Dionysius is not without a sense of personal
devotion to the God-man, as when he prays that his discourse may
be guided “by Christ, by my Christ,” at the beginning of [the]
Celestial Hierarchy.’8 Christ also plays an important role in
Dionysius’s mystical vision. For God reveals and communicates
himself through the celestial and ecclesiastical hierarchies, and
Christ is the head and inner power of both. And the centrality of
Christ is even clearer in Dionysius’s rst major western disciple. For
John the Scot, the end of creation as a whole is the Word of God.
‘The beginning and the end of the world are in the Word of God,
indeed, to speak more plainly, they are the Word itself, for It is
manifold end without end and beginning without beginning…save
for the Father.’9 Moreover, only the Word made esh makes the
return of all things to God possible. ‘God’s Word cried out in the
most remote solitude of divine Goodness…He cried out invisibly
before the world came to be in order to have it come to be; he cried
out visibly when he came into the world in order to save it.’10 Or
consider Pierre de Berulle, who ‘never stopped drinking at the well
of Dionysius’,11 yet more or less seamlessly wedded his mystical
apophatism to a very rich Christ mysticism.

The most important reason for the neglect of mystery, however,

may be this. William Alston begins his recent ‘Two Cheers for
Mystery’ by observing that ‘contemporary Anglo-American analytic
philosophy of religion’ exhibits ‘a considerable degree of con dence
in’ its ability ‘to determine what God is like; how to construe his
basic attributes; and what his purposes, plans, standards, values and
so on are’. No one ‘thinks we can attain a comprehensive knowledge
of God’s nature and doings. But on many crucial points, there seems
to be a widespread con dence in our ability to determine exactly
how things are with God.’12 And, of course, the more con dent one
is, the less one will see any need for according the concept of
mystery a central place in one’s re ections on God. But what if
failing to do so distorts these re ections? The burden of this chapter
is that it does.

Philosophical and theological discussions of the Christian mysteries

are frequently confused by a failure to carefully distinguish
distinct13 religiously relevant uses of the term. Four seem to me
particularly important.

First, a sense of mystery can refer to wonder, surprise, or

astonishment at something the human mind did not expect, and
could not have anticipated. For example, ‘things beyond our seeing,
things beyond our hearing, things beyond our imagining, prepared
by God for those who love him’ (1 Corinthians 9). Mysteries in this
sense, however, need not (though they may) be things we can’t
know or understand once they have occurred or been revealed to us
(the Resurrection, for example, or the new life in Christ).

Second, ‘mysteries’ may refer to doctrines that are either

incongruent or formally inconsistent with ‘common notions’.
Jonathan Edwards uses ‘mystery’ in this sense when he says that we
should expect a revelation of ‘spiritual’ or ‘invisible’ things to be
attended with much mystery and di culty since they are remote
from ‘the nature of things that language is chie y formed to express
—viz. things pertaining to the common a airs and vulgar business
of life, things obvious to sense and men’s direct view…and of an
exceeding di erent nature from the things of this world…and not
agreeable to such notions, imaginations, and ways of thinking that
grow up with us and are connatural to us’.14 Thus Edwards argues
that without a love of holiness and its consequent horror of sin, the
doctrine of hell will seem absurd (unfair, unjust, and, in that sense,
irrational) since one fails to appreciate what makes in nite
punishment appropriate.

Again, a common pagan reaction to the ‘scandal’ of the cross was

that it was not ‘God-be tting’. It is often assumed that what
troubled re ective or philosophically minded pagans was the
doctrine of a cruci ed God’s seeming inconsistency with God’s
impassibility. Paul Gavrilyuk has convincingly shown, however,
that, for the most part, pagan theologians did not think that the
divine’s impassibility precluded emotions (love or benevolence, for
example), or an involvement in human a airs. What troubled them,
rather, was the attribution of su ering to God, and the kind of
involvement implied by the doctrine of the cruci xion. ‘A slave’s
death on the cross was unanimously regarded as shameful and
degrading.’ Docetic forms of Christian Gnosticism were at least
partly motivated by a desire to defuse this scandal. Thus, ‘a
Basilidean account of the cruci xion…puts the following confession
into the mouth of Christ… “I did not die in reality but in
appearance, lest I be put to shame by them…”’ Similarly, ‘in the
docetic segment of the apocryphal Acts of John’, Christ reveals
himself to John on the Mount of Olives at the very moment at which
the cruci xion is taking place at Golgotha as the cosmic being he
really is, but says that others (namely orthodox Christians) ‘“will
call me something else, which is vile and not worthy of me.”’15

Third, a doctrine or truth may seem absurd, unbelievable, or, at

the very least, mysterious because we are deprived of relevant
information. Suppose, for example, that it is true that God exists and
that every evil is necessary for a greater good, and consider Rowe’s
example of a fawn who dies slowly and painfully in a raging forest
re. Even if we believe that because God exists, the fawn’s death is
necessary for a good that outweighs it, how it does so may well seem
mysterious. But, the reason it seems mysterious is, arguably, that we
are ignorant of certain relevant facts. Finite intellects or, in any
case, nite intellects in via, cannot grasp all the relevant logical
connections between goods and evils, and are unavoidably ignorant
of some of the relevant goods and evils (either because they don’t
know what these goods and evils are or, if they do, fail to fully
appreciate their goodness or badness). The truths in question are not
intrinsically mysterious, however, since the mystery would be
dispelled if we were to come into possession of the relevant
information—through a special revelation, perhaps, or at the
eschaton or divinely ordained denouement of the world’s history.
Thus, Edwards thinks that some central doctrines are ‘very
mysterious’, and ‘have di culties in them, inexplicable by us’. The
doctrine of predestination, for example, is ‘very di cult to reconcile
with the justice of God’. Nevertheless, ‘the time is coming when
these mysteries will all be unfolded, and the perplexing di culties
that have attended them will all be perfectly vanished away, as the
shades of night before the sun in a serene hemisphere’.16

In other cases, however, the mystery may be uneliminable by

any addition of information or strengthening of our intellectual
powers. In cases of this sort the mystery is irreducible because its
object is intrinsically mysterious. These are mysteries in Gabriel
Marcel’s sense. Marcel distinguishes ‘problems’ from ‘mysteries’.
Problems17—a chess problem or mathematical problem, for
instance, or the puzzles brought to Sherlock Holmes—have
solutions, although these solutions may never be discovered. If and
when the solutions are discovered, however, the problems
disappear. A mystery, on the other hand, has no solution. No matter
how much we may come to learn about it, it remains as mysterious
as it was before.18 A mystery in Marcel’s sense is more or less the
same as what Rudolf Otto called the ‘mysterium…that which is
“mysterious”’, ‘in the religious sense’, is ‘the “wholly other,”…that
which is quite beyond the sphere of the usual, the intelligible, and
the familiar, which therefore falls quite outside the limits of the
“canny” [of what is ‘within our ken’], and is contrasted with it,
lling the mind with blank wonder and astonishment’.19 The
concept of mystery, like the concept of transcendence, is formally
negative. ‘On the side of the feeling-content’, however, ‘it is
otherwise; that is in very truth positive in the highest degree,
though…it cannot be rendered explicit in conceptual terms.’20

Mysteries in this fourth sense are, arguably, the subject of the

passages from the Pseudo-Dionysius, John Chrysostom, and
Jonathan Edwards that I quoted at the beginning of this chapter.
They are also my primary concern in what follows.

II

Alston o ers four reasons for what he calls the ‘Divine Mystery
Thesis’, namely, that ‘God is inevitably so mysterious to us, to our
rational capacities…that nothing we can think, believe, or say about
him is strictly true of God as he is in himself.’ These reasons include
the experiences of the great Christian mystics and the doctrine of
simplicity to which I shall turn in a moment. The others are, rst,
‘the puzzles, paradoxes, and insoluble problems which theological
thought [about the Trinity, for example] seems so frequently to
lead’; and, second, our limited capacities: ‘If we think about the
relation of human cognitive powers to the absolutely in nite source
of all that is other than itself, it seems reasonable to suppose that
the former would not be in a position to get an account of the latter
that is exactly correct, even in certain abstract respects’21 (Alston
100–1). Or as Edwards argues: ‘A very great superiority, even in
beings of the same nature as ourselves’, makes many of their
actions, intentions, and assertions ‘incomprehensible and attended
with inexplicable intricacies’. Witness the relation that ‘little
children’ bear to ‘adult persons’, for example, or the ‘vulgar’ to
‘learned men, [or] great philosophers and mathematicians’. God ‘is
in nitely diverse from and above all in his nature’, however. So if
God vouchsafes a revelation of himself (of his triune nature, say)
‘which is entirely diverse [not only] from anything we do now
experience in our present state, but from anything that we can be
conscious or immediately sensible of in any state whatsoever that
our nature can be in, then especially may mysteries be expected in
such a revelation’.22

Or consider Aquinas. In Part 4 of the Summa Contra Gentiles,

Aquinas argues that because ‘the human intellect’ must ‘derive its
knowledge from sensible things, [it] is not able through itself to
reach the vision of the divine substance in itself, which is above all
sensible things and, indeed, improportionately above all other
things’. Yet ‘man’s perfect good is that he somehow know God’.
Therefore, because ‘it was [but] a feeble knowledge of God that man
could reach’ by unassisted reason, ‘God revealed certain things
about himself that transcend the human intellect…These things [the
doctrine of the Trinity, for example] are revealed to man as…not to
be understood but only to be believed as heard, for the human
intellect in this state…is connected with things sensible [and]
cannot be elevated entirely to gaze upon things which exceed every
proportion of sense’, although ‘when it shall have been freed from
the connection with sensibles, then it will be elevated to gaze upon
the things which are revealed’.23 The arguments I have discussed so
far focus on the relation between our necessarily limited intellects,
on the one hand, and the divine, on the other. Other arguments
focus more directly on God’s own being.

According to John the Scot, for example, God knows himself but
does not know what he is. John’s reasons for this are essentially
these: to know what God is, one would have to grasp God’s essence;
God ‘is not essence’, though, ‘but More than Essence and the in nite
Cause of all essences, and not only in nite but the In nity of all
in nite essence, and More than in nity’. God is beyond essence
because essences are the sorts of thing that can be captured or
expressed in de nitions, and de nitions proceed by marking out a
thing’s boundaries. God has no boundaries, however, since he is
in nite or limitless. Not even God, then, can know what God is. But
this does not imply that God is ignorant (because ‘He does not
understand of Himself what He is’) nor does it imply that God is
impotent (because ‘He is unable to de ne His Substance’). For God
has no what.24 Nor does it imply that ‘God does not know himself’,
or know himself as in nite, and as indeed ‘above every nite thing
and every in nite thing and beyond nitude and in nity’.25 (John
the Scot, Bk. 2. 585A-590D).

John’s case for the claim that God has no essence or ‘what’
depends on the following subargument. If a de nition is to succeed
in distinguishing its de niendum from other things, it must exclude
the item being de ned from other things. An ‘absolutely in nite’
God would exclude nothing, however, since if it did exclude
something, it would have limits and thus not be ‘absolutely in nite
(or limitless).

This subargument has a certain (specious) plausibility if

‘absolutely in nite’ is construed in arithmetical or geometrical
terms. The series of natural numbers excludes no natural numbers,
for example, and an in nite line excludes none of its segments. Yet
why should we construe ‘absolutely in nite’ in this fashion? Why
not, instead, construe it as ‘absolutely perfect’? If we were to do so,
however, the subargument wouldn’t go through because absolute
perfection does exclude things, for it excludes imperfections such as
ignorance and injustice and mixed perfections26 such as
repentance.27 Let us turn, then, to another attempt to show that
God’s nature is such that he transcends what can be thought of him.

In Proslogion 15, Anselm exclaims: ‘Lord, not only are you that
than which a greater cannot be thought, but you are also something
greater than can be thought. For since it is possible to think that
there is such a one, if you were not this same being something
greater than you could be thought—which cannot be.’28

Commenting on this passage, M. J. Charlesworth observes that

Anselm is reminding us that ‘even if we understand God to be “that
than which nothing greater can be thought,” we do not thereby
have a positive or determinate knowledge of God’;29 and refers us to
the reply to Gaunilo where Anselm says that just as one can think or
understand ‘the ine able’ though one can’t ‘specify [or describe]
what is said to be ine able; and just as one can think of [or
understand] the inconceivable—although one cannot think of what
“inconceivable” applies to—so also “that than which a greater
cannot be thought”…can be thought of and understood even if the
thing itself cannot be thought and understood.’30 But granting this,
one may still wonder why Anselm thinks that a being greater than
can be thought is greater or more perfect than one lacking this
property.

Being such that one cannot think it is not itself a perfection if for
no other reason than because our inability to think something
(adequately capture it in concepts) may be a function of its
imperfection. Plato’s Receptacle, or Aristotle’s or Plotinus’s hyle, are
examples.31 Again, a thing might be too complicated or too hidden
for our intellects to comprehend it. The true nature of the physical
universe might be an example. It doesn’t follow that its
impenetrability to nite intellects is a good-making feature of it.

Being too perfect for us to conceive, on the other hand, might be

a second-order perfection, that is, a perfection parasitic on a thing’s
other perfections. And perhaps God has this second-order property.
Yet why think he does?

Well, perhaps because while we know that God’s joy, for

instance, or his knowledge, are perfect we don’t have a good
conceptual grasp of either of them. In the case of a degreed property
that lacks an upper limit, such as joy, this might be because the bliss
that God enjoys is incommensurable with any nite analogue of it.32
Or in the case of a degreed property such as knowledge which has
an intrinsic maximum (namely, maximal knowledge or
omniscience), nite intellects may know that God knows all truths,
and yet not know how he knows them, or just what his knowledge of
them is like. Moreover, if God is simple, as much of the tradition
maintains, then no positive characterization of God is strictly true of
him since’ all our propositional thought and speech is necessarily
carried on by making distinctions’ (Alston 101). Or perhaps Karl
Rahner is right, and mystery in Marcel’s or Otto’s sense is an
intrinsic positive feature of the Godhead. (More on this later.)

But suppose we agree that being greater than can be thought is a

consequence of (some of?) God’s rst-order perfections, or of the
relation between them (that is, of his simplicity),33 or of the fact
that mystery is an intrinsic feature of the divine essence.34 Does it
straightforwardly follow that being greater than can be thought is
not only a divine property but also a divine perfection? Only if we
assume that any property entailed by a perfection is itself a
perfection. And this assumption is false. Repentance entails a prior
sin, and while the former is a perfection or good-making property,
the latter is not. Again, intelligence trivially entails two plus two
equaling four but two and two summing to four is not a perfection.
Even so, since being greater than can be thought is arguably
entailed by God’s rst-order perfections, if God were not greater
than can be thought, he would necessarily lack one or more of those
perfections and so would not be God, a being greater than which
none can be thought.
 There is reason to think, then, that God necessarily transcends
what can be said and thought of him. Just how mysterious is he,
though? We will turn to this question in the next section.

III

At one extreme are the views of deists such as Charles Blount who
assert that ‘that rule which is necessary for our future happiness
ought to be generally made known to all men…Therefore, no
revealed religion’, with its attendant mysteries, ‘is necessary for
human happiness.’35 John Toland and other deists think that,
because God’s perfection entails his ‘justice and reasonableness’,
‘nothing in true Christianity…is either contrary to or above reason
…Nothing in true Christianity’, therefore, ‘can be a mystery—a
proposition or notion impenetrable to ordinary, human intellectual
capacities.’36

At the other extreme lies Pierre Bayle who professes faith in the
Christian mysteries but insists on their contra-rationality. The
doctrine of the Trinity, for example, contradicts the self-evident
principle that if x = y and y = z then x = y = z, for any x, y, and
z. And more generally, ‘there is a clear incompatibility between
accepting the Cartesian standard of clarity and distinctness’, which
Bayle believes to be appropriate in philosophy, ‘and accepting the
Christian doctrine [s]…Mysteries…are necessarily non-evident’.37
Hence, while faith ‘produces a perfect certitude…its object will
never be evident. Knowledge, on the other hand, produces together
both complete evidence of the object and full certitude of
conviction.’ So ‘if a Christian…undertakes to maintain the mystery
of the Trinity’, for instance, ‘against a philosopher, he would oppose
a non-evident object to evident objections’.38

Bayle claims that skeptical arguments drive us to faith. ‘It is

through a lively awareness’ of ‘the di culties that surround the
doctrines of the Christian religion…that one learns of the excellence
of faith and the blessing of heaven. In the same way, one also learns
of the necessity of mistrusting reason and having recourse to grace.’
‘It has pleased the Father, and the Son, and the Holy Ghost,
Christians ought to say, to lead us by the path of faith, and not by
the path of knowledge and disputation…We cannot lose our way
with such guides. And reason itself commands us to prefer them to its
direction (Bayle 435, 423, my emphasis).

Yet can one ‘assert doctrines while simultaneously holding that

they contradict self-evident principles’?39 Bayle thinks that one can.
For in doing so, one is not inconsistently ‘believing and not
believing the same [proposition] at the same time’. Rather, one
consistently believes that ‘(1) The light of reason teaches me that
[the doctrine] is false; [yet] (2) I believe it nonetheless because I am
convinced that this light is not infallible and because I prefer to
submit…to the Word of God, than to a metaphysical demonstration’
(Bayle 298). Is this coherent? Penelhum thinks that it isn’t. For in
spite of Bayle’s explicit disclaimer, he does believe two inconsistent
things. Through faith Bayle believes the doctrine of the Trinity, for
example, and yet at the same time believes that the doctrine’s falsity
is entailed by self-evident principles. But because one can’t believe
that a proposition, p, is entailed by self-evident principles without
believing p, Bayle must also believe that the doctrine of the Trinity
is false.40

I do not nd this objection compelling. Granted, one can’t both

nd a set of principles self-evident and nd that they self-evidently
entail p, and not believe p. To rest on this, though, seems to miss
Bayle’s point, namely, that his radical skepticism has led him to
doubt so-called self-evident principles. Since he no longer nds them
self-evident, he no longer nds the propositions they entail evident
either, and so does not inconsistently believe the doctrine of the
Trinity, for instance, while simultaneously believing its denial.

Whether Bayle can escape all charges of incoherence is more

doubtful. For if all principles of reason are called into question, it is
di cult to see how Bayle can be justi ed in asserting that reason
itself directs us to set reason aside.41 As far as I can tell, Bayle has
only two recourses. The rst is to restrict his skepticism to only some
of reason’s self-evident principles. This seems to me a non-starter,
however, for if his doubt extends to self-evident logical and
mathematical principles, as his animadversions on the doctrine of
the Trinity imply, it is di cult to see how he could retain his
con dence in any principle of reason. The second is less obviously
incoherent, and this is to regard the apparent bankruptcy of reason
as a cause of, rather than a reason for, the ight to faith. Bayle’s
procedure would then bear some resemblance to that of the
Buddhist Madhyamikas who employ reason to show the incoherency
of reason, thereby, in their view, clearing a space for the occurrence
of the non-dual, non-conceptual intuition of the Suchness of things.

There may also be another problem. John Toland characterizes

views such as Bayle’s thus: while doctrines such as the Trinity
‘cannot in themselves be contradictory to the principles’ of reason
yet, because ‘of our corrupt and limited understandings’, they may
seem to us to contradict them. On ‘the authority of divine
revelation’, however, ‘we are bound to believe…in them’, and ‘to
adore what we cannot comprehend’ (Toland 24).

Toland has two responses to this position, the second of which is

more telling,42 namely, that we cannot understand what is or seems
to be contradictory,43 and cannot believe what we don’t understand,
although ‘a man may give his verbal assent to he knows not what,
out of fear, superstition…interest, and the like…motives’. ‘For what
I don’t conceive, can no more give me right notions of God, or
in uence my actions, than a prayer delivered in an unknown tongue
can excite my devotion’ (Toland 35, 28, my emphasis). Of course,
strictly speaking, Toland is mistaken because we can and do
understand contradictions. Indeed, it is just because we understand
them, that we know they are necessarily false. In a larger sense,
though, Toland is right. Because contradictions imply everything, in
believing a contradiction I am implicitly committed to believing that
God both is and is not evil, that we both should and shouldn’t
worship him, and the like, and it is di cult to see how a belief of
this sort could give me right notions of God or guide my behavior.
Moreover, this is true even if the doctrines proposed for our belief,
while not contradictory, are wholly impenetrable by our
understandings. The divine mysteries, then, cannot be
contradictions and they cannot be totally opaque to nite intellects.
So precisely what about them can’t we understand?

Leibniz understands Bayle to say that while God can comprehend

the mysteries of his own being (and so see that they involve no
inconsistencies), they are contrary to human and, indeed, any nite
understanding. He nds this incoherent, however. For, in his view,
human reason is part of divine reason, and what is contrary to part
of reason must be contrary to the whole of reason since the latter
includes the former.44 Why does Leibniz think this?

Some of the divine mysteries, at least, are necessary truths. The

doctrine of the Trinity is an example.45 Now Leibniz believes that a
comprehension of necessary truths depends upon a proof or analysis
that demonstrates or explicates their necessity. In the case of the
divine mysteries, however, the required analysis or proof involves
an in nite number of steps, and so God alone can perform it. Hence,
while God can comprehend the mysteries of his own being, we
cannot. We can, nonetheless, meaningfully and justi ably believe
and assert them, and this for three reasons. First, we possess or have
been given ‘analogies’. The soul’s union with its body, for instance,
is an analog of the Word’s union with a human nature. Second, even
though we can’t prove a mystery and, in that sense, comprehend it,
we can refute objections to it. A ‘defense of the mysteries’ is similar
in this respect ‘to a defense of the consistency and completeness of
arithmetic’. Kurt Gödel (1906–78) has shown that arithmetic’s
consistency and completeness can’t be proved ‘but this does not
prevent one from [successfully] defending it against speci c
accusations of inconsistency’.46 Finally, one has the standard
‘motives of credibility’ for believing the mysteries (miracles, ful lled
prophecy, and so on). Leibniz has thus explained, to his own
satisfaction, how divine mysteries can be both incomprehensible to
nite intellects and justi ably believed and asserted by them.

Leibniz’s refutation of Bayle depends on his claim that human

reason is not a mere analog of God’s reason but literally a part of it.
Why does he say this? ‘When functioning properly, humankind’s
reasoning is simply “the linking together of truths and objections in
due form”…[and] as such cannot be mistaken’. Divine reason does
not di er from human reason in kind but only in degree. Although
the demonstration or analysis of the divine mysteries ‘require[s]
in nite cognitive power’,47 the kind of demonstration or analysis
required is essentially the same. But this strikes me as doubtful at
best and, in any case, not what the tradition as a whole has had in
mind: the di erence between God’s nature and capacities and ours
is not merely quantitative but qualitative, a di erence of kind and
not just of degree. Let us therefore try again.
 Thomas Aquinas claims that some truths about God exceed ‘the
power of human reason’. In a paper published in 1988, George
Mavrodes suggested that we gloss ‘exceeds the power of human
reason’ thus: ‘A truth exceeds the ability of all human reason if and
only if it is not possible to prove that truth demonstratively.’ Why
think that, in this sense, the doctrine of the Trinity, say, exceeds the
power of human reason? Perhaps because, as Aquinas suggests,
there are only two ways of demonstrating a truth about something—
by deducing it from our ‘understanding of the substance [or
essence] of the thing which is the subject of that knowledge’ or from
the thing’s causal e ects.48 Truths about such things as the Trinity
can’t be demonstrated in this way, however. For we can’t (prior to
the beati c vision at least) grasp God’s essence. And Aquinas
assumes that ‘the creative power of God is common to the whole
Trinity, and hence…belongs to the unity of the essence, and not to
the distinction of the persons’. So while ‘by natural reason we can
know what belongs to the unity of the essence’ from God’s casual
e ects, we cannot know ‘what belongs to the distinction of the
persons’ from them.49

Mavrodes’ principal objection to this line of thought is that there

are other no less reasonable interpretations of ‘demonstration’. For
example, following Plantinga and others, we might identify a
demonstration with a sound, non-circular deductive or inductive
argument from universally (or nearly universally) accepted
premises. Or, following Mavrodes and Penelhum, we might identify
a demonstration with a sound, non-circular argument from premises
that its proponent or recipient knows to be true or has strong
reasons to believe are true. Although it is unlikely that one can
provide demonstrations in the rst sense of the doctrine of the
Trinity (or of any other interesting philosophical or theological truth
for that matter), the prospects of doing so in the second may be
brighter. And, in fact, Richard of St Victor, Jonathan Edwards, and
Richard Swinburne claim to have provided demonstrations of the
doctrine of the Trinity in the sense in question.50

The tradition does more or less unanimously attest that God’s

essence is unknown or unknowable by nite creatures.51 The fact
that a thing’s substance or essence is unknown or unknowable does
not entail that we can’t rationally establish many truths about it,
however. Locke thought that the substances or essences of physical
objects were currently unknown but did not think that there were
no well-grounded truths about them. Again, many truths about
water were known before it was discovered that water is H2O. Nor
would matters obviously have been di erent if we had not
discovered that water is H2O or even if, for some reason, humans
never could have discovered this.

Yet if God’s essence or inner being is unknown or unknowable,

what can we know about God? One of the most common answers
hinges on a distinction between knowing that a proposition is true
and knowing or understanding how it can be true. As Edwards
observes in ‘The Mind’, 71, ‘it is not impossible to believe or know
the truth of mysteries, or propositions that we cannot comprehend,
or see the manner how the several ideas that belong to the
proposition are united…we may perceive that they are united and
know that they belong one to another, though we do not know the
manner how they are tied together’.52 In view of remarks Edwards
makes elsewhere, it is clear that he is referring not only to
propositions that we know to be true because their truth is attested
by reliable authority but also to propositions that we can prove or
demonstrate without fully comprehending them. In a late entry, for
example, Edwards asserts that ‘’Tis not necessary that persons
should have clear ideas of things that are the subject of a
proposition, in order to being rationally convinced of the truth of
the proposition,’ and cites ‘many truths of which mathematicians
are convinced by strict demonstration…concerning which they have
no clear ideas’, such as propositions about ‘surd quantities and
uxions’.53 It is in this sense, presumably, that after having given his
rational account of just why God must be triune, Edwards exclaimed
‘I don’t pretend fully to explain how these things are… I am far from
pretending to explaining the Trinity so as to render it no longer a
mystery.’54

In what sense, though, can propositions which we know to be

true elude our comprehension? Well, in some cases we may observe,
and hence know, that an event has occurred without understanding
how it could have occurred, or be cognizant of a phenomenon such
as lightning or combustion yet be utterly ba ed by it. In cases like
these, what is lacking is a grasp of the occurrence’s causes or of the
mechanisms or internal structures underlying the phenomenon that
puzzles us. Something like this may be involved in our
understanding of God’s providence, for example. As James
Kellenberger observes in discussing Job’s faith, one can believe or
even know that God exists, is good and merciful, and the like
without understanding how God is good and merciful, since ‘his
ways of goodness, mercy’, and so on, ‘are beyond our conceiving…
This source of mystery—the inconceivability of God’s ways—would
remain just as it is if it were known that there is a God, that he is
good and that he is merciful.’55 What Job lacks, in short, is an
understanding of the mechanisms of divine providence. The extent to
which divine providence is a mystery in this sense is a matter of
debate, of course. Job confesses that he has ‘obscured [God’s]
designs with [his] empty-headed words …holding forth on matters
[he] cannot understand, on marvels beyond [him] and [his]
knowledge’ (Job 42: 3). Edwards, on the other hand, thought that
he did have a grasp of the mechanisms of providence, devoting an
entire sermon series to an explication of the wondrous ‘machine’,
composed of wheels within wheels, that underlies and constitutes
the history of redemption.

Edwards’s Discourse on the Trinity suggests a somewhat di erent

sense of ‘incomprehensibility’. He compares the Christian enquirer
to a student of nature. When the latter ‘looks on a plant’, or an
animal, ‘or any other works of nature, at a great distance, [he] may
see something in it wonderful and beyond his comprehension’, and
so desire to view it more closely. And if he does, he ‘indeed
understands more about them…and yet the number of things that
are wonderful and mysterious in them that appear to him are much
more than before. And if he views them with a microscope, the
number of the wonders that he sees will be much increased still. But
yet the microscope gives him more of a true knowledge concerning
them.’56 What this comparison suggests is that while Christian
divines’ investigations of the Christian mysteries are progressive in
the sense that more and more facets of these mysteries are revealed,
more questions answered, and more puzzles removed, the
knowledge that is gained only leads to new questions and new
puzzles: ‘However…some di culties are lessened, others that are
new appear; and the number of those things that appear mysterious,
wonderful and incomprehensible are increased by it.’57 Edwards’s
point in this passage, I think, is not that some questions about the
Trinity, for instance, are unanswerable (although they may be) but,
rather, that any answers we may discover simply lead to more
questions. The problem the passage isolates is thus quantitative
rather than qualitative. But unlike some of the more optimistic
estimates of the possibility of progress in science, Edwards seems to
think that our mounting success in answering questions about the
Christian mysteries takes us no closer to the goal of answering all
questions about them.
 Yet another sense of ‘incomprehensibility’ is suggested by
‘Miscellany 839’: divine mysteries ‘are not only so above human
comprehension that men can’t easily apprehend all that is to be
understood concerning them, but [they] are di cult to the
understanding in that sense, that they are di cult to be received by
the judgement or belief’.58 Why is this the case? At least partly
because the Christian mysteries are attended with ‘paradoxes’ and
‘seeming contradictions’. Di culties of this sort are not peculiar to
divinity. ‘The reasonings and conclusions of the best metaphysicians
and mathematicians concerning in nities’, for instance, are also
‘attended with paradoxes and seeming inconsistencies’.59 (Hilbert’s
hotel is an example.60)61 The problem is partly due to the
limitations of language. As Edwards observes in ‘Miscellany 83’,
‘The things of Christianity are so spiritual, so re ned, so high and
abstracted, and so much above the things we ordinarily converse
with and our common a airs, to which we adapt our words,’ that
‘we are forced to use words…analogically…and therefore [does]
religion [abound] with so many paradoxes and seeming
contradictions.’62 If I understand Edwards, however, the problem is
not just with our language but with our imagination or sense of
grasp.63 Our analogies and metaphors are ultimately inadequate. We
lack an adequate model or, perhaps more accurately, an adequate
uni ed model of the deep things of God. Edwards sometimes
employs an Augustinian psychological model of the Trinity, for
example, while at other times employing patristic social models. He
makes no attempt to unify them, however—presumably because he
sees no way of doing so.64

IV

But suppose that we grant that the veil between God and ourselves
cannot be removed in this life. Will we, indeed, can we, behold God
unveiled in the next? Some texts suggest that we will. Paul, for
example, asserts that while ‘we now see only puzzling re ections in
a mirror, we shall then see face to face’. Our ‘knowledge now is
partial’ but ‘then it will be whole, like God’s knowledge of me’ (1
Corinthians 13: 12). Does the beati c vision, then, include an
unclouded vision of God’s essence? Some Christian theologians, at
least, have thought not.65

Thus, Aquinas says that ‘since our mind is not proportionate to

the divine substance, that which is the substance of God remains
beyond our intellect and so is unknown to us. Hence the supreme
knowledge which man has of God is to know that he does not know
God, in so far as he knows that what God is surpasses all that we
can understand of him.’66 Rahner seems to me to be correct in
arguing that, because ‘the reason for saying’ that knowing God
involves knowing that one does not know God67 ‘holds good for the
beati c vision’ as well as for the veiled glimpses of God we have in
this mortal life, ‘there is no reason for not applying [it] to the
knowledge of God in the beati c vision’68 (Rahner 59).
 Some of the things Edwards says have similar implications.
Edwards believes that while many of the di culties and perplexities
surrounding Christian doctrines will be cleared up in the ‘future
appointed time of joy and glory to the church [on earth]…the
perfect and full explication of these mysteries is part of the last and
eternal state of the church [in heaven], to heighten the joy and
praises of the wedding day of Christ and his church’.69 It doesn’t
follow that all mystery will be dispelled, however, and at least some
things Edwards says suggest that it won’t. Thus, as we have seen, in
‘Miscellany 1340’ Edwards asserts that because God ‘is in nitely
diverse from and above all others in his nature’, any revelation he
chooses to vouchsafe of his intrinsic nature (of his triunity, say) will
be so ‘entirely diverse [not only] from anything that we do now
experience in our present state, but from anything that we can be
conscious or immediately sensible of in any state whatsoever that our
nature can be in’, that many ‘mysteries may be expected in such a
revelation’.70 The clear implication is that even the light of heaven
won’t, and indeed cannot, dispel all mystery.

One of the most powerful statements of this view is given by

John Chrysostom, who exclaims: ‘let us call upon him, then, as the
ine able God who is beyond our intelligence, invisible,
incomprehensible…Let us call on him as the God who is inscrutable
to the angels, unseen by the Seraphim, inconceivable to the
Cherubim, invisible to the principalities, to the powers, and to the
virtues, in fact to all creatures without quali cation, because he is
known only by the Son and the Spirit’ (Chrysostom 97, my
emphases). Why do the Seraphim ‘stretch forth their wings and
cover their faces? For what other reason than that they cannot
endure the sparkling ashes nor the lightning which shines from the
throne. Yet they did not see the pure light itself nor the pure essence
itself. What they saw was a condescension accommodated to their
nature’71 (Chrysostom 101). So unless the beati ed see God more
clearly than the angels do, they do not grasp God’s essence. The
mystery of God is thus ineluctable.

And, indeed, it is possible that it is ineluctable in an even

stronger sense. For there may be a sense in which God’s own
knowledge of himself does not dispel the mystery. Karl Rahner
points out that in post-medieval scholastic theology (and, I would
add, in much if not most contemporary analytic philosophy of
religion), mystery is a property statements have when they exceed
our reason or cannot be fully understood. This conception of
mystery has three noteworthy features. First, while it is admitted
that the mystery of doctrinal statements is rooted in features of their
object, the focus of scholastic theology’s discussion of mystery is on
the statements, rather than on what those statements are about.
Second, mystery is regarded as a function of the relation of the
propositions in question to human reason. Third, reason is construed
in its modern sense as ratiocination or ‘calculation’, and thus sharply
distinguished from the will and a ections. Rahner thinks that each
of these features re ects a mistake. In the rst place, mystery is
primarily a characteristic of the deep things of God—not of the
doctrinal statements that express them. In the second, mystery is not
(in the rst instance at least) a function of the relation between
certain propositions about God and nite intellects but, instead, an
intrinsic property of God himself. Finally, we can cognize mystery
by faith alone, and faith is an expression of our will and a ections
as well as our intellect. To recognize mystery is not just to
acknowledge that certain propositions exceed our grasp; it is to
prostrate ourselves in loving wonder before something which can’t
be comprehended by any sort of propositional cognition.

In short, as Rahner says, if God’s incomprehensibility ‘is the very

substance of our vision and the very object of our blissful love’, then
‘vision must mean grasping and being grasped by the mystery, and
the supreme act of knowledge is not the abolition or diminution of
the mystery, but its nal assertion’ (Rahner 41). The beati c ‘vision
of God face to face’ does remove many mysteries, ‘but this only
means that what [the mysteries] express is manifested in its own
being and substance, is experienced therefore in itself and must no
longer rely for its manifestation on the [authoritative] word that
does duty for it72…Nonetheless, these mysteries [that is, what one
now directly beholds] remain mysterious and incomprehensible’
(Rahner 56). The Greek Fathers are thus right when they speak of
‘the highest stage of life and knowledge’ as entering ‘into the
darkness in which God is’ (Pseudo-Dionysius), or a not-knowing
which is ‘the supra-rational knowledge’ (Maximus the Confessor), or
tell us that ‘to enter the holy of holies is to be encompassed by the
divine darkness’ (Gregory of Nyssa) (Rahner 58).

But all this, if correct, has a potentially startling consequence.

For if mystery is not, in the rst instance, a function of the relation
between God and nite intellects but, rather, an intrinsic property of
God’s own nature, then God’s complete and perfect knowledge of
himself must include a recognition of it. ‘The absolutely clear self-
awareness of’ God may thus include ‘something positive which does
not appertain to the [propositional] intellect73 but to the mystery in
contradistinction from such an intellect.’ If this is so, then mystery
‘appertains to God’s knowledge [of himself], essentially, in a
preeminent and analogous sense’ (Rahner 48–9). There may thus be
a sense in which God himself can’t comprehend his own essence74
but must enter into the ‘divine darkness’, knowing himself, or
aspects of himself, through a ‘not-knowing’ that is at one and the
same time a supreme ‘supra-rational knowledge’ of the deep things
of his own being.

To see why this suggestion isn’t outrageous, we need to say more

about the kind of knowledge involved in apprehending a mystery in
Marcel’s or Rahner’s sense.

An awareness of mystery in the sense in question is perhaps best

construed as a species of appreciation or knowledge by
acquaintance.75 Other examples of this sort of knowledge are my
knowledge of what strawberries taste like or silk feels like, my
awareness of a thunderstorm’s sublimity or the beauty of a Bach
fugue, or my knowing what it is like to su er or to love. Notice that
these forms of knowledge by acquaintance are related analogically,
and vary with their objects. Our appreciation of beauty, for
instance, is importantly di erent from our acquaintance with sense
modalities such as the taste of strawberries or the feel of silk, and
both di er signi cantly from a rst-hand knowledge of the horror of
war or what Kierkegaard calls ‘ rst love’.

Instances of knowledge by acquaintance aren’t just di erentiated

by their objects, however. Consider the wonder over her own beauty
that Semele expresses in an aria in Handel’s opera of the same
name, or the Greeks’ wonder at the world (the fact of its being),76 or
the Christian’s or Muslim’s or Sri Vaisnava’s wonder at the glory of
God. These instances of wonder di er not only in object but in
phenomenological quality or feel. Semele’s wonder, for example, is
qualitatively di erent from the Greeks’ wonder at the world, and
both di er qualitatively from the theists’ wonder at the glory of
God.77

The object of a sense of mystery in Marcel’s or Rahner’s sense is

God’s own being or nature, and the best description of its
phenomenological character is probably Otto’s—’blank wonder, an
astonishment that strikes us dumb, amazement absolute’, occasioned
by coming ‘upon something inherently “wholly other,” whose kind
and character are incommensurable with our own, and before which
we therefore recoil in a wonder that strikes us chill and numb’, but
whose ‘feeling-content’ is ‘positive in the highest degree’.78
 Does a sense of mystery in this sense entail lack of understanding
or grasp? The answer is ‘Yes’ if understanding or grasp is de ned in
terms of conceptual grasp or propositional knowledge (knowledge
that). The answer is ‘No’ if an a rmative answer is understood to
imply the existence of a gap in understanding that could in principle
be lled in, or that the sort of knowledge involved in this and other
instances of knowledge by acquaintance isn’t adequate to its object.

Are there two quite di erent senses of mystery,79 then, which we

might call epistemological mysteries and ontological mysteries,
respectively? (Epistemological because the mysteries in question are
a function of the relation between God’s nature or being, on the one
hand, and the limitations of created intellects, on the other.
Ontological because the mystery this expression gestures at is an
intrinsic aspect of God’s own being rather than a feature of human
or divine knowledge of it.) These senses of mystery are at home in
very di erent places. Discussions of epistemological mysteries are at
home in philosophical theology, for example, and are the principal
subject of this chapter’s rst three sections. The ontological sense of
mystery, on the other hand, is perhaps most at home in liturgical
worship and the prayer of adoration.80 But the two senses of
‘mystery’ are not entirely equivocal. For both senses are partly
de ned by a lack of conceptual mastery. Epistemological mysteries
and ontological mysteries elude conceptualization in very di erent
ways, however, and for very di erent reasons. Epistemological
mysteries elude it because while adequate concepts are in principle
available (if only to God), they are not available to us. Ontological
mysteries elude conceptualization because no concepts can fully
express them. They are best (albeit imperfectly) expressed by
symbols, images, songs, poetry, and, perhaps ultimately, by the
silence of mystical prayer.

The question raised by Rahner’s claim that ‘mystery appertains

to God’s knowledge’ of himself is thus roughly equivalent to this.
Can God adopt the attitudes described by Otto toward himself? On
the whole,81 I don’t see why not. Even if we understood everything
about ourselves that can be conceptually grasped we might still
wonder at, or be amazed or astonished by, our own being. (‘We are
fearfully and wonderfully made’ (Psalm 139: 14)) Why, then, can’t
God wonder at, or be amazed by, his being? Moreover, that God can
adopt these attitudes toward himself seems even clearer in a
Trinitarian context. A standard eucharistic prayer concludes, ‘All
this we ask through your Son Jesus Christ. By him, and with him, and
in him, in the unity of the Holy Spirit, all honor and glory are yours,
Almighty Father, now and forever’ (my emphasis). Just as we adore
or glorify God, so the Father and Son and Holy Spirit adore and
glorify each other. Nor do I see why the Father’s and the Son’s and
the Holy Spirit’s mutual adoration can’t be tinged with awe,
wonder, and (even) astonishment.

So in precisely what sense does mystery ‘appertain to God’s

knowledge’ of himself? God, I suggest, knows everything that can be
propositionally known about himself82 through an analog of
propositional cognition. But other aspects of God’s being (those
which are mysteries in Marcel’s or Rahner’s sense) can’t be grasped
in this way even in principle, and God knows these by an analog of
appreciation or knowledge by acquaintance. Nothing about God is
thus unknown to God. He is neither ‘ba ed’ nor ‘puzzled’ nor
‘mysti ed’ by his own being.83 Nevertheless, God’s nature is for him
an object of an amazement, wonder, and awe, which are the felt
aspects, as it were, of a perfect experiential acquaintance with
depths of his own being that necessarily elude even his own
complete conceptual comprehension.84

NOTES

1. David Hume, The Natural History of Religion (London: Adam &

Charles Black, 1956), 54 (henceforth Hume).

2. John Toland, Christianity Not Mysterious (London, 1696; repr.

New York: Garland, 1978), 73 (henceforth Toland). I have
modernized Toland’s spelling and capitalization.

3. Toland ultimately explains it by the activity of ‘cunning priests’

who exploited people’s gullibility for ‘their own advantage’
(Toland 70). Hume traces its origin to people’s fear of God and
of his ill favor; and who, ‘supposing him to be pleased, like
themselves with praise and attery’, spare ‘no eulogy or
exaggeration…in their addresses to him…Thus they proceed;
till at last they arrive at in nity itself…And it is well, if, in
striving to get farther…they run not into inexplicable mystery,
 and destroy the intelligent nature of their deity, on which
alone any rational worship or adoration can be founded’
(Hume 43).

4. Dionysius the Areopagite on The Divine Names and The Mystical

Theology, trans. with introduction by C. E. Rolt (London:
Macmillan, 1957), 191 (henceforth Dionysius).

5. John Chrysostom, On the Incomprehensible Nature of God, trans.

Paul W. Harkins, Fathers of the Church (Washington: The
Catholic University of America Press, 1984), 100 (henceforth
Chrysostom).

6. Jonathan Edwards, Discourse on the Trinity, in The Works of

Jonathan Edwards (New Haven: Yale University Press, 1957–),
xxi. 134, 139.

7. The apophatic tradition (or tradition of negative theology)

insists that God is most accurately described by saying what
God is not.

8. Bernard McGinn, The Foundations of Mysticism: Origins to the

Fifth Century (New York: Crossroad, 2003), 180.

9. Johannes Scotus Erigena, Periphyseon: The Division of Nature,

trans. I.P. Sheldon-Williams, rev. John J. O’Meara (Montreal:
Bellarmin, 1987), Book 5. 893A (henceforth John the Scot).

10. John the Scot, Commentary on John. Quoted in Bernard

McGinn, The Growth of Mysticism: Gregory the Great through the
12th Century (New York: Crossroad, 2004), 108.
11. William L. Thompson (ed.), Berulle and the French School:
Selected Writings, intro. (New York: Paulist Press, 1989), 14.

12. William P. Alston, ‘Two Cheers for Mystery’, in Andrew Dole

and Andrew Chignell (eds.), God and the Ethics of Belief: New
Essays in Philosophy of Religion (New York: Cambridge
University Press), 99 (henceforth Alston).

13. Distinct, though not necessarily mutually exclusive.

14. Jonathan Edwards, ‘Miscellany 1340’, in Works, xxiii. 368.

15. Paul Govrilyuk, The Su ering of the Impassible God (Oxford:

Oxford University Press, 2004), 81–2. ‘Many apologists
considered an open attack upon the prevailing [sentiments
and] social conventions to be the most successful defense
strategy.’ They admitted ‘that the divine birth, su ering, and
cruci xion were unseemly, scandalous, and o ensive in the
eyes of the world’, but insisted that what the world deemed
shameful and o ensive is in fact a most God-be tting way of
securing the redemption of a ruined creation since only by
assuming esh and su ering and death could God redeem
them (ibid. 87). I would only add that scandals in this sense
don’t entirely disappear upon conversion. For old sentiments
and habits of evaluation linger on even though they no longer
dominate one’s life and outlook. In so far as they do, the sense
of o ense and scandal isn’t entirely eradicated.

16. Jonathan Edwards, ‘Miscellany 654’, Works, xviii. 195–6.

17. At least when well formed.

18. A mystery, in this sense, should be distinguished from the

‘unknowable’, for the latter is a purely negative notion. ‘The
recognition of mystery’, on the other hand, ‘is an essentially
positive act of the mind’ (Gabriel Marcel, Being and Having
(New York: Harper & Row, 1965), 118).

19. Rudolf Otto, The Idea of the Holy, trans. John W. Harvey
(NewYork: Oxford University Press, 1958), 26.

20. Ibid.30. Although it can be expressed in images and symbols.

Otto is explicitly speaking in this passage of the concepts of
transcendence and the supernatural. The context makes it
clear, however, that what Otto says of the concepts of
transcendence and the supernatural also applies to the concept
of mystery.

21. Alston asks us to consider, in this connection, how di cult it is

to even craft ‘a picture of the physical world…the complete
correctness of which we can be assured’ (Alston 100).

22. Jonathan Edwards, ‘Miscellany 1340’, Works, xviii. 370–1, my

emphasis.

23. Thomas Aquinas, On the Truth of the Catholic Faith, Book Four:
Salvation, trans. Charles J. O’Neil (Garden City: Doubleday,
1957), 35–7, my emphasis. While the general thrust of this
passage is clear enough, it does raise an important question to
which we will return later. For if God ‘improportionately’
 transcends all nite things, as Aquinas says, why think that
freedom from connection with sensibles will be su cient to
remove the mystery or darkness that surrounds God ‘s nature?

24. So there is nothing God has that God doesn’t know, and no task
that someone could perform (such as de ning God’s essence)
that God cannot do.

25. The idea here is presumably this. Finitude and in nity contrast
with, and hence circumscribe, each other. An in nite line or
quantity isn’t a nite line or quantity, and vice versa, and is, in
that sense, limited. In nite lines or quantities are thus only
limitless in certain respects. God, by contrast, is limitless in all
respects.

26. Mixed perfections are good-making properties that entail an

imperfection.

27. John’s claim that God has no essence also rests upon the
traditional notion that a good de nition states a thing’s genus
and di erentia and/or proceeds by locating the de niendum
within the Aristotelian categories. If God is simple and/or
transcends the Aristotelian categories, then neither of these
requirements can be met. What is unclear, however, is just
why de nitions must meet either of these requirements. Does
‘God = df. a being greater than which none can be thought’
do so, for example? I doubt that it does, but a traditionalist
could reply that it doesn’t capture God’s essence either and, for
that reason, isn’t a so-called ‘real de nition’.
28. Anselm, St. Anselm’s Proslogion, with a Reply on Behalf of the
Fool by Gaunilo and the Author’s Reply to Gaunilo, trans. M. J.
Charlesworth (Oxford: Clarendon, 1965), 137.

29. Ibid. 81, my emphases.

30. Ibid.189.

31. Matter for Plotinus, for example, is ‘not an independently

existing principle, but the point at which the out ow of reality
from the One fades away into utter darkness’; or, alternatively,
the point at which the process of fragmentation reaches its
logical limit, a plurality without any unity at all—and hence
not even a real plurality, since a real plurality is ‘always a
plurality of [determinate] things, each of which is one’. Matter
is thus sheer formless indeterminacy. Matter is below being
(although not non-existent) just as the One is above it. Since
only what has being and form can be conceptualized, both
matter and the One elude our intellect (R. T. Wallis,
Neoplatonism (New York: Charles Scribner’s Sons, 1972), 48–
50).

32. Not only is God’s joy or happiness greater than any nite joy or
happiness, no nite joy or happiness is half as great, or two-
thirds as great, or almost as great as God’s.

33. If simplicity supervenes on the relations between God’s rst-

order properties, being greater than can be thought would be a
third-order divine property: being greater than can be thought
 supervenes on God’s simplicity, which supervenes in turn on
the relations between his rst-order properties. If, however,
God’s simplicity is simply identical with the relations between
his rst-order properties, then being greater than can be
thought would presumably be a second-order divine property.

34. Strictly speaking, the latter two cases are also cases in which
being greater than can be thought is a consequence of God ‘s
rst-order properties. For (as we saw in n. 33), since God’s
simplicity depends on his rst-order properties, so too does the
property of being greater than can be thought which is a
consequence of it; and mystery, in Rahner’s view, is
presumably a rst-order property of the divine essence (?).

35. Charles Blount, Oracles of Reason, 198–9. Quoted in Peter

Byrne, Natural Religion and the History of Religion: The Legacy of
Deism (London: Routledge, 1989), 53–4.

36. Byrne, ibid. 54, 71. Toland allowed that revelation might be a
‘source’ of true propositions but insisted that ‘once such truths
come to my notice by revelation’, they must pass the bar of
human reason, i.e. I must be able to make out their truth for
myself on the basis of publicly available evidence. More
radical questioners of revealed religion refused to grant ‘even
this limited role to’ tradition since doing so would preclude
some people from salvation (those who lived before the
decisive moment or kairos, for example) (ibid. 72–4).
 But even if God’s justice and reasonableness were to
preclude his making human salvation depend on assent to
mystery, why should it preclude either the possibility that
certain truths about him are irreducibly mysterious or his
communicating those truths to us? Perhaps it wouldn’t. But
the deists’ con dence in the transparency of a just and
reasonable religion doesn’t comport well with either the
existence or communication of irreducibly mysterious truths.
In any case, if knowledge of them isn’t necessary for salvation,
then (from the deists’ point of view) their communication is
super uous at best and, at worst, a distraction from what is
truly important.

37. Terence Penelhum, God and Skepticism’ (Dordrecht: D. Reidel,

1983), 27.

38. Pierre Bayle, Historical and Critical Dictionary: Selections, trans.

Richard H. Popkin (Indianapolis: Bobbs-Merrill, 1965), 414
(henceforth Bayle).

39. Penelhum, God and Skepticism, 28.

40. Ibid.56–9.

41. For a related objection see Toland: ‘The very supposition, that
reason might authorize one thing, and the Spirit of God
[Scripture] another, throws us into inevitable skepticism. For
if contradictions can be true, the authority of reason is called
into question. And if the authority of reason is called into
 question, then so too is the authority of Scripture since the
latter rests on the former. ‘We believe the Scripture to be
divine, not upon its own bare assertion [or upon an inner
impulse or intuition, or the so-called testimony of the Holy
Spirit], but from a real testimony consisting in the evidence of
the thing contained therein; from undoubted e ects, and not
from words and letters’ (Toland 30, 32). We believe Scripture,
in other words, because what it tells us—that God exists, that
we are immortal, that the best o ering we can give God is a
moral life—is intrinsically evident to reason. However
inadequate this account of the authority of Scripture may be,
Toland has put his thumb on a real problem. For calling
reason’s authority into question undercuts any attempt to o er
rational arguments for relying on Scripture, including not only
standard appeals to miracles, ful lled prophecy, and so on, but
also existential or pragmatic arguments such as Bayle’s.

42. Toland’s rst response is this: to the claim that while doctrines
such as the Trinity are indeed ‘not contrary to sound reason…
no man’s reason is sound’, Toland objects that even though the
reason of most people is indeed unsound, the de facto defects
of human reason can be remedied without divine assistance.
For we can learn to ‘compare ideas, distinguish clear from
obscure conceptions, suspend our judgments about
uncertainties, and yield only to evidence’ (Toland 57, 60). One
may reasonably doubt whether these measures are su cient to
 restore an impaired or fallen intellect, however. They clearly
are not if, as I have argued elsewhere, a rightly disposed heart
is needed to reason rightly about religion and other value-
laden matters. See my Reason and the Heart: A Prolegomena to a
Critique of Passional Reason (Ithaca, NY: Cornell University
Press, 1995).

43. And ‘a seeming contradiction is to us as good as a real one’

(Toland 34, my emphasis).

44. Leibniz’s idea, I think, is that the set of propositions certi ed by

human reason (that is, that seem clearly true to a properly
functioning human reason) is a proper subset of the set of
propositions that are certi ed by reason as such. The set of
propositions that are certi ed by reason as such, however, is
identical with the set of propositions that are certi ed by
divine reason. If so, then if it were to seem clearly true to a
properly functioning human reason that (e.g.) God can’t be
one and three, then ‘God can’t be one and three’ would be a
member of the set of propositions that are certi ed by divine
reason. Since, by hypothesis, the divine reason also certi es
‘God is three and one,’ the set of propositions certi ed by
divine reason would be inconsistent.

45. This is not absurd since presumably, if God is triune, he is

necessarily triune. It is more di cult to see how the doctrine
of the Incarnation is a necessary truth, however, except on the
peculiarly Leibnizian view that all truths are analytic.
46. Adrian Bardon, ‘Leibniz on the Epistemic Status of the
Mysteries’, Philosophy and Theology13 (2001), 153.

47. Ibid. The internal quote is from Leibniz’s Theodicy, trans.E.M.

Huggard (La Salle: Open Court, 1985), 108.

48. George Mavrodes, ‘“It is Beyond the Power of Human Reason”’,

Philosophical Topics 16 (1988), 77.

49. Thomas Aquinas, Summa Theologica Iq. 32 a. 1, trans. the

Fathers of the English Dominican Province (New York:
Benziger Brothers, 1947), i (my emphases).

50. Indeed, if we follow Locke, Swinburne, and others who believe

that a rationally compelling case can be made for the authority
of the Christian revelation, and that that revelation includes
the doctrine of the Trinity, and if we broaden the concept of
proof or demonstration to include proofs from testimony, then
reason can prove or demonstrate that God is triune. As
Jonathan Edwards says, ‘divine testimony’ cannot be opposed
to reason, evidence, and argument because it is a rule of
reason, a kind of evidence, and a type of argument like the
‘human testimony of credible eye-witnesses’, ‘credible history’,
‘memory’, ‘present experience’, or ‘arithmetical calculation’
(Jonathan Edwards, ‘Miscellaneous Observations’, The Works
of President Edwards (New York: B. Franklin, 1968); repr. of
the Leeds edn. reissued with a 2-vol. supplement in Edinburgh
(1847), 228. While Edwards doesn’t clearly distinguish
between evidence, argument, and rule of reason, the
distinction is presumably this: apparent memories, for
example, are a type of evidence, justifying claims by appeal to
memory is a type of argument, and the appropriate rule is
‘one’s memories are normally reliable’. Similarly, the contents
of Scripture are a type of evidence, justifying claims by
appealing to Scripture is a type of argument, and the
appropriate rule is ‘Scripture is trustworthy.’) ‘Scripture is
reliable’ resembles such rules as ‘The testimony of our senses
may be depended on,’ ‘The agreed testimony of all we see and
converse with continually is to be credited,’ and the like
(Edwards, ‘Miscellany 1340’, Works, xviii. 361). Principles
such as these can be established, or at least certi ed, by reason
and then used to establish other truths that cannot be
established without their help. However, that reason can
appropriately be used to establish the credentials of a rule of
reason, and that that rule can in turn be used to establish
other truths that can’t be demonstrated without it, does not
imply that opinions formed by a reason that does not employ
the rule can be used to determine the truth or falsity of
opinions established only by its means. The naked eye, for
example, ‘determines the goodness and su ciency’ of the optic
glass, yet it would be absurd for someone to ‘credit no
representation made by the glass, wherein the glass di ers
from his eye’, and to refuse to believe ‘that the blood consists
partly of red particles and partly of limpid liquor because it all
appears red to the naked eye’ (‘Miscellaneous Observations’,
 Works of President Edwards, 227). Similarly, the fact (if it is a
fact) that the credibility of Scripture can be established by a
sound non-circular deductive or inductive argument does not
show that every truth the Scripture contains can be
demonstrated in the same fashion.

51. As we shall see, e.g. John Chrysostom clearly thinks that God’s
essence can’t be known by even the most exalted of creatures.
The cherubim and seraphim themselves know God only by
‘ gures’.

52. Jonathan Edwards, ‘The Mind’, Works, vi. 385.

53. Jonathan Edwards, ‘Miscellany 1100’ Works, xx. 485.

Propositions about in nities and, perhaps, quantum mechanics
provide other examples.

54. Jonathan Edwards, Discourse on the Trinity, Works, xxi. 134.

55. James Kellenberger, ‘God and Mystery’, American Philosophical

Quarterly 11 (1974), 99.

56. Jonathan Edwards, Discourse, Works, xxi. 140.

57. Jonathan Edwards, ibid 139.

58. Jonathan Edwards, ‘Miscellany 839’ Works, xx. 54–5.

59. Jonathan Edwards, ‘Miscellany 1340’, ibid. xviii. 371.

60. David Hilbert described a hotel with an in nite number of

rooms, each of which is occupied. Yet even though the hotel is
full, its proprietor can accommodate an in nite number of
 new guests. If Jones requests a room, for example, the
proprietor can simply move the guest in room number 1 to
room number 2, the guest in room number 2 to room number
3, the guest in room number 3 to room number 4, etc., and
assign Jones room number 1. And the process can be repeated
for each new arrival.

61. Or again, in discussing some of the implications of idealism

(which he thinks is not only coherent but true), Edwards
asserts, ‘But we have got so far beyond those things for which
language was chie y contrived, that unless we use extreme
caution, we cannot speak, except we speak exceedingly
unintelligibly, without literally contradicting ourselves’, and
adds the following corollary: ‘No wonder, therefore, that the
high and abstract mysteries of the Deity, the prime and most
abstract of all beings, imply so many seeming contradictions’
(‘The Mind’, sect. 35, Works, vi. 355).

62. Jonathan Edwards, ‘Miscellany 83’, ibid. xiii. 249.

63. The two are not unrelated of course.

64. See e.g. William J. Danaher, Jr., The Trinitarian Ethics of

Jonathan Edwards (Louisville: Westminster John Knox, 2004).
Cf. Amy Plantinga Pauw, The Supreme Harmony of All: The
Trinitarian Theology of Jonathan Edwards (Grand Rapids:
William B. Eerdmans, 2002).

65. Or, more cautiously, have said things that imply that it doesn’t.
66. Thomas Aquinas, de Potentia, q. 7 a. 5. Quoted in Karl Rahner,
‘The Concept of Mystery in Catholic Theology’, Theological
Investigations (New York: Seabury, 1974), iv. 58–9 (henceforth
Rahner).

67. Namely, that our minds are ‘not proportionate to the divine
substance’.

68. Aquinas, of course, thinks that the redeemed in heaven have a

direct vision of God’s essence. Yet how is this possible if ‘no
proportion exists between the created intellect and God’
because of the ‘in nite distance between them’? His answer is
that while there is no quantitative proportion between God and
created intellects (God isn’t twice as great, or four times as
great, or a thousand times as great, or…as created intellects),
‘every relation of one thing to another is called proportion’,
and created intellects are related to God as e ects to their
cause (Summa Theologica, I q. 12, a. 1, my emphasis). Whether
this response is adequate seems to me doubtful. It may be
su cient to show that created intellects are capable of
knowing that there is a rst cause and that that cause must
have those properties entailed by its being rst cause. It isn’t
clearly su cient to show that created intellects are capable of
directly beholding God’s essence. But even if they are, Aquinas
is clear that the lack of ‘quantitative’ proportion between
created intellects and God implies that God’s essence can’t be
known by created intellects as God himself knows it, namely, ‘in
 an in nite degree’. The beati ed see God ‘wholly’, and there is
nothing ‘of him [which] is not seen’. But he is not seen
‘perfectly’. Aquinas illustrates the distinction between knowing
a whole in its entirety yet not knowing it perfectly by
contrasting two ways of knowing that a triangle’s angles are
equal to two right angles, the rst resting on a ‘probable
reason’ and the second on a ‘scienti c demonstration’. A
person possessing the latter ‘comprehends’ the truth while
someone who only possesses the former does not. Yet it ‘does
not follow that any part of [the truth] is unknown [to the
former], either the subject, or the predicate, or the
composition; but [only] that it is not as perfectly known as it
is capable of being known’ (ibid. I q. 12 a. 7) Aquinas’s point, I
think, is not that certain truths about God’s essence necessarily
elude the saints in heaven but, rather, that their apprehension
of them is unavoidably imperfect or obscure to one degree or
another. The object of the saints’ vision (the divine essence) is
apprehended directly (without a medium) and in its entirety,
but the manner in which they behold it is necessarily
imperfect. None of this, as far as I can see, is inconsistent with
the thesis of this section, namely, that there are aspects of God
that can’t be fully captured by the propositional or conceptual
intellect.

69. Jonathan Edwards, ‘Miscellany 654’ Works, xviii. 196.

70. Jonathan Edwards, ‘Miscellany 1340’ Works, xxiii. 370–1, my
emphasis.

71. ‘God condescends whenever he is not seen as he is, but in the

way one incapable of beholding him is able to look upon him’,
that is, by images, visions (including, presumably, so-called
intellectual visions), and the like (Chrysostom 101).

72. Namely, the Bible, creeds, and the like.

73. That is, to concepts or their divine analogs.

74. That is, doesn’t know himself by any analog of propositional

knowledge.

75. I owe this suggestion to Alvin Plantinga.

76. See e.g. Michael B. Foster, Mystery and Philosophy (London:

SCM, 1957), ch. 2.

77. Note that appreciation should not be construed as a purely

subjective reaction to something that can, in principle, be fully
captured in concepts. The taste of strawberries, for example, or
the nature of rst love elude adequate conceptualization
although both can be expressed in poetry or song.

78. Otto, Idea of the Holy, 26, 28, and 30.

79. As Plantinga also suggested to me.

80. See e.g. the eucharistic hymn adapted from the Liturgy of St
James that begins ‘Let all mortal esh keep silence and with
fear and trembling stand.’
81. On the whole, because God presumably isn’t ‘numbed’ or
‘chilled’ by the sight of his own being.

82. That he is omnipotent, say, or good.

83. Reactions, which notice, are only appropriate to puzzles in

Marcel’s sense, i.e. to problems that would be resolved by the
acquisition of the right propositional knowledge.

84. Thus, to put the point in terms of the distinction introduced

earlier, God is not an epistemological mystery to himself but
does regard his own being as an ontological mystery.
 PART II

DIVINE ATTRIBUTES
 CHAPTER 5

SIMPLICITY AND ASEITY

JEFFREY E. BRQWER

THERE is a traditional theistic doctrine, known as the doctrine of

divine simplicity, according to which God is an absolutely simple
being, completely devoid of any metaphysical complexity. On the
standard understanding of this doctrine—as epitomized in the work
of philosophers such as Augustine, Anselm, and Aquinas—there are
no distinctions to be drawn between God and his nature, goodness,
power, or wisdom. On the contrary, God is identical with each of
these things, along with anything else that can be predicated of him
intrinsically.1

Although divine simplicity was once regarded as an essential

part of philosophical theology, having been upheld for over a
millennium by a veritable army of philosophical theologians—not
only Christian, but also Jewish and Islamic—the doctrine has, in
more recent history, fallen on hard times. Philosophers and
theologians now seldom speak of divine simplicity, and when they
do, their remarks are almost always critical. Indeed, contemporary
analytic theists often take themselves to have conclusive reasons for
rejecting it. ‘The trouble with the idea’, C. B. Martin once remarked,
‘is just that it is hogwash.’2 Many others would agree, perhaps even
be willing to go so far as to claim, with Quentin Smith, that divine
simplicity is not only ‘plainly self-contradictory’ but actually
‘testi es to the predominance of faith over intellectual coherence’.3

In this chapter, I take the rst steps necessary for restoring the
doctrine of divine simplicity to its former glory, arguing that its
widespread rejection in contemporary philosophy and theology is
certainly premature, perhaps ultimately unwarranted. There can be
no question that this doctrine comes with substantial and
controversial commitments in metaphysics. But in each case, I shall
argue, these commitments are perfectly respectable, having been
ably defended and taken very seriously on independent grounds in
the contemporary literature. If my argument is successful, it will be
clear that this doctrine—together with the conception of divine
aseity that traditionally motivates it—deserves more attention than
it has yet received at the hands of contemporary philosophers and
theologians.

My discussion is divided into three main parts. In sect. 1, I

provide a brief introduction to divine simplicity, describing its chief
motivation historically and explaining why, on the standard
contemporary interpretation, the doctrine looks incoherent. In sect.
2, I articulate and defend an alternative interpretation of divine
simplicity, one calculated to avoid the problems plaguing the
standard interpretation and to establish the doctrine’s coherence. In
sect. 3, I show how my preferred interpretation can be extended to
deal with what is perhaps the chief objection to simplicity from
within traditional theism itself—namely, that it appears to exclude
the possibility of contingent divine volition and knowledge.

1. THE DOCTRINE OF DIVINE SIMPLICITY

The doctrine of divine simplicity arises from philosophical re ection

on a traditional conception of God—one that is common both to
philosophers of antiquity and to adherents of the world’s three great
monotheistic religions: Judaism, Christianity, and Islam. In order to
clarify this doctrine, as well as to prepare for a discussion of the
main objections to it, it will be useful to approach simplicity via the
traditional theistic conception from which it arises.

1.1. Personhood, Aseity, and Simplicity

Traditional theism has many ingredients, including among others

that God is an omnipotent, omniscient, perfectly good (or loving),
eternal, necessarily existing, divine being. Obviously, this list is not
exhaustive. But it is su cient to highlight one of the most important
assumptions of traditional theism—namely, that God is a person, at
least in the broad sense of an entity possessing the sorts of mental
states generally regarded as constitutive of personhood.4 For as the
foregoing list makes clear, the God of traditional theism possesses
intellectual states such as knowledge (in virtue of which he is
omniscient), and appetitive states such as desires or volitions (in
virtue of which he is perfectly good or loving).

Traditional theists disagree among themselves about how exactly

to think of divine personhood. Some ancient Greeks, for example,
claim that God possesses each of his mental states essentially, so
that he could not have known or willed anything other than he
actually knows or wills. By contrast, other traditional theists,
including most orthodox Jews, Christians, and Muslims, insist that
God possesses at least some of his mental states contingently, so that
he could, for example, have chosen to create a universe di erent
from the actual one—or none at all. Even so, these di erences occur
within the traditional conception of God as a person in the broad
sense sketched above.

There is another ingredient of traditional theism that is

especially important for our discussion in what follows. In addition
to thinking of God in broadly personal terms, traditional theists also
habitually think of him as an absolutely independent being—that is,
as a being who is rst or primary in the sense that he does not
depend on anything distinct from himself. Such a being, it is often
said, exists entirely from himself (a se). Hence, traditional theists
often characterize God’s absolute independence in terms of his
aseity.

Here again, there are disagreements to be noted among

traditional theists. Some, in uenced by the philosophical theology
of Aristotle, regard divine aseity as among the most basic or
fundamental features of our conception of God. Aquinas’s famous
‘ ve ways’, for example, are speci cally designed to establish the
existence of a being who is primary in just the sense described
above—a being which, he says, ‘everyone calls God’.5 Others,
in uenced more by Neoplatonic considerations, take divine
goodness or perfection to be God’s most basic feature, arguing that
divine aseity follows directly from it. Thus, both Augustine and
Anselm defend divine aseity on the grounds that dependency on
another is always an imperfection, and hence must be excluded from
our conception of God. But here again the di erences among
traditional theists arise from a common commitment—in this case,
to the view that divine aseity is somehow essential to God.

The reason that divine aseity is especially important for our

purposes here is that it provides divine simplicity’s chief motivation
historically. As Aquinas points out, in the section of his Summa
Theologiae immediately following the ‘ ve ways’, it is a very short
step from aseity to simplicity: ‘Every composite thing is posterior to
its components and dependent on them. But, as was shown above,
God is the rst being [and hence not dependent on anything]’
(Summa Theologiae 1. 3. 7). The basic pattern of reasoning that
Aquinas invokes here helps to explain why, on the traditional
doctrine of divine simplicity, God must lack not only proper parts or
constituents, but even distinct properties or attributes. If God exists
entirely a se, he cannot depend on anything in any way at all, not
even in the way that a subject depends on its properties (in order to
exemplify them). As Aquinas says in his other Summa: ‘In every
simple thing, its being and that which it is are the same. For if the
one were not the other, simplicity would be removed. As we have
shown, however, God is absolutely simple. Hence, in God, being
good is not anything distinct from him; he is his goodness’ (Summa
Contra Gentiles 1. 38).

This same basic pattern of reasoning pervades traditional

philosophical theology, and lies behind not only Aquinas’s
Aristotelian-based formulation of divine simplicity, but also
standard Neoplatonic formulations of the doctrine. Thus, Augustine,
inspired by the very same considerations, says in a passage much
quoted in the contemporary literature:

We speak of God in many ways—as great, good, wise, blessed, true,

and whatever else does not seem unworthily said of him.
Nonetheless, God is identical with his greatness, which is his
wisdom (since he is not great by virtue of quantity, but by virtue of
power); and he is identical with his goodness, which is his wisdom
and his greatness; and he is identical with his truth, which is all of
these things. For in him it is not one thing to be blessed and another
to be great, or wise, or true, or to be good, or to be altogether
himself. (De Trinitate 6. 7. 8)

Following Augustine in this regard, Anselm says in a well-known

passage of his Proslogion: ‘Life, wisdom, and all the rest are not parts
of you, but all are one, and each of them is the whole of what you
are and the whole of what the others are’ (Proslogion 18). And
similar such remarks can be found in the works of other traditional
theists, not only in the Latin-speaking Christian west, but also in the
Arabic-speaking Jewish and Islamic east. In short, it can be found
wherever there is support for the traditional understanding of divine
aseity.6

1.2. Simplicity and Its Interpretation

On the standard contemporary interpretation, the doctrine of divine

simplicity requires that God is identical with each of his intrinsic
properties. Nor is it hard to see why. What else could it mean to say
that God is identical with his nature, goodness, power, and so on?
The problem with this standard interpretation, however, is that it
appears to lead directly to incoherence. If God is identical with each
of his properties, then God must himself be a property. But that
seems absurd. As we have seen, one of the most obvious things
about God, on traditional theism, is that he is a person. But no
person can be a property. For properties are, by their very nature,
exempli able—that is, things that can be possessed, instantiated, or
had. But no person could be a thing of that sort. Indeed, insofar as
divine simplicity requires God to be a property, it appears to be not
merely absurd, but guilty of a category mistake—that of placing a
non-exempli able thing (namely, God) into the category of
exempli ables (namely, properties).7
 If we take a closer look at traditional formulations of divine
simplicity, however, we can see that there is nothing in them that
requires the identi cation of God with a property. On the contrary,
all they require is that if a predication such as ‘God is good’ is true,
then there exists an entity, God’s goodness, that is identical with God;
likewise, if ‘God is powerful’ is true, then God’s power exists and is
identical with God; and so on for all other such divine predications.
More precisely, what traditional formulations of the doctrine require
is the following:

(DS) If an intrinsic predication of the form ‘God is F’ is true, then

God’s F-ness exists and is identical with God.

So understood, the doctrine of divine simplicity takes no stand

whatsoever on the precise nature of the entities with which it
identi es God. It does assume that there are (or at least could be)
entities corresponding to expressions such as ‘God’s goodness’,
‘God’s power’, and ‘God’s wisdom’.8 Nonetheless, it says nothing
about the speci c ontological category to which they belong. But
this just goes to show that the claim that God is identical with a
property results not from the doctrine of divine simplicity itself, but
rather from its conjunction with something like the following
‘property account’ of predication and abstract reference:

(PA) If an intrinsic predication of the form ‘a is F’ is true, then a’s

F-ness exists, where this entity is to be understood as a
property.9
 Although contemporary defenders of divine simplicity often
recognize that something like PA is at the root of contemporary
di culties with the doctrine, they are extremely reluctant to
abandon it. Indeed, they prefer (almost to a person) to defend the
coherence of the standard interpretation (i.e. the conjunction of DS
and PA) rather than develop an account of predication and abstract
reference in terms of something other than properties.10 This seems
to me to be a mistake—indeed, one that explains the general failure
of contemporary defenses of divine simplicity. Even so, there are at
least two things that explain the contemporary resistance to
abandoning PA.

First, this account strikes many as extremely intuitive. We

habitually speak as if for any true (atomic) predication, there is a
subject of predication (e.g. Socrates), there is property (e.g.
Socrates’s wisdom or wisdom in general), and the subject
exempli es the property (e.g. Socrates is wise). In fact, this account
is so intuitive that many nd it di cult to imagine how expressions
of the form ‘a’s F-ness’ could refer to anything but properties.11

Second, many traditional proponents of divine simplicity say

things that appear to commit them to PA. To take just one example,
consider the following passage from Anselm’s Monologion, which
involves a comparison of divine and human justice:

A human being cannot be his justice, though he can have his justice.
For the same reason, a just human being is not understood as being
his justice (existens iustitia), but as having his justice. By contrast, it
is not properly said that the supreme nature has its justice, but is its
justice. Hence when the supreme nature is called just, it is properly
understood as being its justice, rather than as having its justice.12
(Monologion 16)

Here Anselm suggests that, if we want to explain the justice of a

human being, we must appeal to a property exempli ed by that
human being. Thus, if Socrates is just, this is because he has his
justice. Like other traditional proponents of simplicity, therefore,
Anselm seems to take for granted that at least some creaturely
predications of the form ‘a is F’ entail the existence of properties,
which can in turn be referred to by expressions of the form ‘a’s F-
ness’. But if expressions of the form ‘a’s F-ness’ refer to properties in
the case of creatures, we might expect them to behave similarly in
the case of God. To the extent that traditional proponents of divine
simplicity suppose that at least some creaturely predications imply
the existence of properties, therefore, it is natural to suppose they
do so because they accept the property account at PA.

There are, then, some considerations that make it tempting to

accept the standard interpretation of divine simplicity in terms of
PA. Even so, I maintain, we must not give in to this temptation. For
the claim that God is identical with a property really is absurd, and
hence any interpretation of simplicity that requires its truth must be
rejected as incoherent. I realize, of course, that there are some who
want to resist this conclusion.13 But since I have defended it at
length elsewhere, I will not repeat my arguments for it here.14
Instead, I will simply present my own preferred alternative to the
standard interpretation, and then argue that it not only renders the
doctrine of divine simplicity perfectly coherent, but also ts well
with the view that there is contingent divine volition and
knowledge.

2. SIMPLICITY AND COHERENCE

The standard interpretation of divine simplicity assumes that

expressions such as ‘God’s nature’, ‘God’s goodness’, and ‘God’s
power’ refer to properties. If we are to reject the standard
interpretation, then we must nd an alternative interpretation
according to which these expressions refer to entities of some other
type. But what other type of entities can they plausibly be taken to
refer to? The answer, I suggest, is to entities of a broadly functional
type—namely, truthmakers.15

2.1. Truthmakers, Predication, and Ontological Neutrality

‘Truthmaker’ is something of a term of art in contemporary

philosophy, but the idea behind it is perfectly intuitive. Many of the
predications we make about the world are true. That much seems
obvious. But many of us also think it is obvious that when such
predications are true, their truth must be a function of the way the
world is. That is to say, when a predication of the form ‘a is F’ is
true, there must be something that makes it true—or better, some
thing (or plurality of things) which explains its truth or in virtue of
which it is true. As these quali cations indicate, the notion of
‘making’ at work here is not causal, but explanatory.16

How exactly are we to understand the notion of explanation

involved in truth-making? Most contemporary philosophers take it
to be a form of broadly logical necessitation or entailment, so that if
an entity E is a truthmaker for a predication P, then E necessitates
that P.17 Although truthmaking and entailment are closely allied
notions, we must be careful not to identify them, since that would
lead to obvious absurdities, including the claim that truths
expressed by necessary predications—such as ‘2 is an even
number’—have anything whatsoever for their truthmaker. But, then,
if truthmaking is not to be identi ed with entailment, how exactly
are we to understand it? This is a notoriously di cult question to
answer. My own view is that the truthmaking is a primitive or sui
generis form of necessitation, one that does not admit of (non-
circular) analysis or de nition.18 But other answers to the question
have been given in the literature.19 Fortunately, we needn’t insist on
the correctness of any of the answers here. For our purposes, it will
su ce to note that truthmaking must involve some form of broadly
logical necessitation, so that even if E’s necessitating that P does not,
by itself, guarantee that E is P’s truthmaker, it does make E a
candidate—perhaps even a prima facie good candidate—for playing
this role.
 Whatever else we say about the notion of a truthmaker, it should
be clear that it is intended to be an ontologically neutral notion. To
characterize an entity as a truthmaker, as we’ve just seen, is to
characterize it in terms of a certain metaphysical function or role it
plays—that of necessitating (in a certain way) the truth of the
predications it makes true. But such a ‘functional’ characterization
places no restriction on the speci c nature or ontological category to
which a truthmaker can belong. Indeed, it leaves open the
possibility that truthmakers can belong to ontological categories of
very di erent kinds, including both concrete individuals (such as
persons) and properties.

Consider, for example, an essential predication such as ‘Socrates

is human.’ In this particular case, it is possible to identify its
truthmaker with the subject of the predication itself. For the
existence of the concrete individual, Socrates, is by itself su cient
for the truth of ‘Socrates is human,’ and hence a candidate for its
truthmaker. Generalizing on this sort of case, some philosophers
have been led to think that concrete individuals are the truthmakers
for all of their true essential predications—that is, for each
predication whose truth they entail.20 But that seems too strong.
Entailment, as we’ve seen, is necessary but not su cient for
truthmaking. Even so, it does seem plausible to think that a concrete
individual can be the truthmaker for a proper subset of its true
essential predications—namely, each of its true intrinsic essential
predications. Thus, Socrates himself, just in virtue of being the
concrete individual he is, can be regarded as the truthmaker for
‘Socrates is human,’ ‘Socrates is an animal,’ ‘Socrates is a material
object,’ ‘Socrates exists,’ ‘Socrates is identical with himself,’ and so
on.

What if we turn now to contingent predications, such as

‘Socrates is wise’ or ‘Socrates is just’? Here things get more
complicated. The mere existence of Socrates, it would seem, is not
su cient for the truth of these predications; hence he cannot
himself be their truthmaker. But then what are we to say about their
truthmakers? There is more than one way to answer this question,
but the two most common ways of answering it both appeal to
properties. First of all, we can say, as David Armstrong does, that
the truthmakers for contingent predications are facts (or concrete
states of a airs) that include properties as constituents.21 In that
case, the truthmaker for ‘Socrates is just’ will be the fact that
Socrates is just, which includes the property justice as a constituent.
Alternatively, we can say, as C. B. Martin does, that the truthmakers
for contingent predications are non-transferable tropes (or concrete
individual properties essentially dependent on the subjects of which
they are the properties).22 In that case, the truthmaker for ‘Socrates
is just’ will not be the fact that Socrates is just, but Socrates’s justice—
an entity such that, in all possible worlds in which it exists, Socrates
exists and is just.

2.2. Truthmakers and Divine Simplicity

The relevance of all this for divine simplicity is perhaps already
obvious. As we have seen, this doctrine requires us to identify God
with each of the things that can be intrinsically predicated of him:

(DS) If an intrinsic predication of the form ‘God is F’ is true, then

God’s F-ness exists and is identical with God.

As we have also seen, however, the doctrine takes no stand on the

precise nature of the entities with which it identi es God. Thus, if
we are to understand what the doctrine amounts to, we must adopt
an interpretation that takes such a stand (but does so without also
rendering the doctrine incoherent). Suppose, therefore, we adopt the
following ‘truthmaker account’ of predication, modeled after the
property account given earlier:

(TA) If an intrinsic predication of the form ‘a is F’ is true, then a’s

F-ness exists, where this entity is to be understood as the
truth maker for ‘a is F’.

Interpreted in light of TA, the doctrine of divine simplicity entails

that God is identical with the truthmaker of each of the true
intrinsic predications that can be made about him. Thus, if God is
divine, he is identical with that which makes him divine; if he is
good, he is identical with that which makes him good; and so on in
every other such case. On this interpretation, therefore, divine
simplicity just amounts to the claim that God is the truthmaker for
each of his true intrinsic predications.23
 It should be clear already that the truthmaker interpretation goes
a considerable distance toward rendering the doctrine of divine
simplicity coherent. On this interpretation, for example, the doctrine
does not require that God is identical with each of his properties,
and hence is himself a property. In fact, it does not even require that
God has any properties at all (in the ontologically loaded sense of
exempli ables). On the contrary, all the doctrine requires is that, for
every true intrinsic divine predication, there is a truthmaker and
God is identical with that truthmaker. But there is nothing obviously
absurd about that. Indeed, on the assumption that each of God’s
intrinsic predications is also essential, this interpretation renders the
doctrine quite plausible in certain respects (more on this below.)

Finally, we should note that the truthmaker interpretation allows

us to make sense of the claim, endorsed by traditional proponents of
simplicity, that abstract expressions such as ‘a’s justice’ can refer
both to concrete individual persons (in the case of God) and to
properties or exempli ables (in the case of creatures). For according
to TA, expressions of this form will refer to whatever it is that
makes predications such as ‘a is just’ true. But in the case of
creatures, unlike that of God, such predications will often be
plausibly regarded as contingent. In order to supply a truthmaker
for it, therefore, we must appeal to something like particular
properties (or non-transferable tropes). And this, I think, is exactly
what traditional proponents of simplicity such as Augustine,
Anselm, and Aquinas do.24
 For all these reasons, it should be clear that the truthmaker
interpretation provides us with an account of divine simplicity that
is at least prima facie coherent. Of course, even if this interpretation
avoids the standard di culty with the doctrine, there might still be
concerns about its own coherence or plausibility. In order to
complete my defense, therefore, I need to respond to the objections
most likely to threaten TA itself. In decreasing order of strength,
these would seem to be the following: (1) truths do not have
truthmakers; (2) truthmakers can’t serve as referents for expressions
of the form ‘a’s F-ness’; (3) a single thing cannot be the truthmaker
for many distinct truths; and (4) a simple God cannot be the
truthmaker for all his true intrinsic predications.25

Objection 1: In the course of developing my preferred

interpretation of divine simplicity, I have appealed to the thesis that
truths have truthmakers—that is, to the claim that when a given
predication is true, there is some entity (or plurality of entities) in
the world that explains its truth. Although initially intuitive, this
appeal might seem problematic, since it runs into di culty when
applied to predications such as ‘Homer is blind’ and ‘Unicorns are
not real,’ whose logical structure seems to be of the form ‘a is not F’
and ‘there are no Fs’. Indeed, predications of this sort are often
thought to be the undoing of truthmaker theory, since the only
candidate truthmakers for them appear to be negative facts such as
a’s not being F and there not being any Fs, and yet negative facts strike
many as extremely implausible. As David Lewis says: ‘It seems,
o hand, that [such predications] are true not because things of
some kind do exist, but rather because counterexamples don’t exist’
(Lewis 1999: 204).

It is important to emphasize, however, that my interpretation of

simplicity does not require the general thesis that all truths have
truthmakers. On the contrary, it requires only a restricted version of
it—namely, the thesis that all positive (atomic) predications of the
form ‘a is F’ have truthmakers (cf. TA above). This more restricted
thesis, however, is extremely plausible. It appears, for example, to
be what motivates the traditional problem of universals, which is
plausibly regarded as that of specifying the truthmakers for positive
atomic predications of the form ‘a is F’.26 Indeed, this understanding
of the problem of universals appears to be the one driving the
‘property account’ of predication (at PA above), in terms of which
divine simplicity is standardly interpreted. The property account
takes for granted that whenever a predication of the form ‘a is F’ is
true, there is a subject, there is a property, and the subject
exempli es the property. But this just appears to be a speci cation
of the restricted truthmaker thesis, a way of saying that the
truthmakers for positive (atomic) predications share a common
structure or ontology.

But perhaps it will be worried that once we abandon the general

truthmaker thesis, we lose all motivation for holding the restricted
version of it. ‘If any truths can lack truthmakers,’ it might be said,
‘then surely some positive atomic truths can, however initially
plausible it seemed to deny this.’ But if that’s right, then my
interpretation will be committed to the general truthmaker thesis
after all.

It is not clear to me how serious this worry is. A number of

philosophers have argued that a principled account can be given of
why certain truths (including negative existentials) lack
truthmakers, even though all positive atomic truths require them.27
But even if their argument fails, and no such principled account can
be given, it isn’t obvious what follows for my interpretation. For
other philosophers have argued equally strenuously for the general
truthmaker thesis, insisting in particular that we are perfectly within
our rights to allow truthmakers even for negative existentials.28

Obviously, these issues are too large to be resolved here. Let us

simply grant, therefore, that it’s a cost of my interpretation that it
requires the success of truth-maker theory in some form or other. In
the present context, this seems to me a perfectly acceptable cost. In
fact, I would regard it as a major advance for the doctrine of divine
simplicity if contemporary philosophers came to regard its
plausibility as on a par with truthmaker theory.

Objection 2: Although truthmaker theory has received

considerable attention from contemporary philosophers, it is not
typically developed in connection with any particular view of
abstract reference in mind. On my interpretation, however,
truthmakers are required not only to explain the truth of true
predications of the form ‘a is F’, but also to serve as the referents for
their abstract counterparts—that is, expressions of the form ‘a’s F-
ness’. This might seem problematic. For it is natural to assume that
such abstract expressions can only refer to properties.29 Indeed, I
suspect it’s the naturalness of this assumption that has always made
the property interpretation so tempting.

As far as I can tell, however, there is nothing to prevent us from

rejecting this assumption, despite its naturalness, and simply
stipulating, as my interpretation does, that expressions of the form
‘a’s F-ness’ are technical terms whose referents are the truthmakers
for the corresponding predications (in this case, ‘a is F’). Of course,
it might be objected that this stipulation is arbitrary. But such an
objection would seem to have little force. A truthmaker theorist
will, presumably, need some way of referring to truthmakers for
particular predications and TA gives us just what is needed. To refer
to the truthmaker for ‘a is F’, qua truthmaker, it tells us, simply
construct the corresponding sentential nominalization, ‘a’s F-ness’.

Again, it might be objected that in adopting such a stipulation

we are departing from common sense, or at least the common
practice of a large number of philosophers. As we’ve seen, many
traditional philosophers have reserved expressions of the form ‘a’s F-
ness’ for properties, and hence for things that can be exempli ed.
And the same could be said of many contemporary philosophers.
But even granting this, it’s not clear that the objection has much
force. As we have seen, the stipulation in question allows—indeed,
is designed to allow—that expressions of the form ‘a’s F-ness’ still
refer to properties in many cases (e.g. Socrates’s justice). Indeed,
insofar as philosophers have traditionally taken these sorts of cases
as paradigmatic, it should come as no surprise that they would think
all such expressions refer to properties. In any case, we shouldn’t
balk at some regimentation of ordinary (philosophical) language,
especially in the service of a large-scale metaphysical view such as
truthmaker theory.30

Objection 3: But even if we allow that all predications of the form

‘a is F’ have truthmakers, and that such truthmakers can be referred
to by expressions of the form ‘a’s F-ness’, there might still appear to
be something problematic about my interpretation of simplicity. For
if this interpretation is true, then God himself turns out to be the
truthmaker for a large number of conceptually distinct predications
—including ‘God is divine,’ ‘God is good,’ ‘God is powerful,’ and
‘God is wise.’ But can a single thing really be the truthmaker for
many conceptually distinct truths?

This objection would have some force if truthmakers were

intended to provide the conceptual content or meaning of
predications. But they are not. Indeed, it is part-and-parcel of
contemporary truthmaker theory to deny that truthmakers stand in
a one-to-one correspondence with the predications they make true.
And such a denial is plausible in particular cases. Consider, for
example, the following predications:

(S1) Socrates is human.

(S2) Socrates is an animal.

 (S3) Socrates is a material object.

(S4) Socrates exists.

(S5) Socrates is identical with himself.

As noted earlier, it seems perfectly plausible to suppose that

Socrates himself is the truthmaker for each of these, despite the fact
that they di er in meaning and logical form. But if that is right,
then there can be no objection in principle to saying that God is the
truthmaker for many distinct predications.

Objection 4: But even if a single thing can, in principle, be the

truthmaker for many distinct predications, it might be said that my
interpretation stretches credulity insofar as it requires not only this,
but also that an absolutely simple thing can make true the speci c
variety of predications that are supposed to be true of God on
traditional theism. Let us consider what can be said in response to
this objection.

To begin, let us note that there doesn’t appear to be anything

particularly objectionable about saying that even a simple thing can
be the truthmaker for a number of distinct predications. Suppose
there exists a particular property or trope of Socrates—say,
Socrates’s whiteness. Now consider the following set of predications
that can be made about it, modeled on those stated above (at S1-
S5):

(T1) Socrates’s whiteness is a whiteness trope.

(T2) Socrates’s whiteness is a color trope.

 (T3) Socrates’s whiteness is a trope.

(T4) Socrates’s whiteness exists.

(T5) Socrates’s whiteness is identical with itself.

Here, as above, it seems plausible to say that a single thing—

namely, Socrates’s whiteness—is the truthmaker for all these
predications, despite the fact that Socrates’s whiteness is a trope and
tropes are standardly regarded to be simple beings. If there is
something objectionable about the truthmaker interpretation,
therefore, it must have something to do with the speci c variety of
predications it requires a simple being to make true.
Initially, there does appear to be something worrisome about the
speci c variety of predications for which the God of traditional
theism is supposed to be the truthmaker. In order to see why,
compare the two sets of predications we’ve considered so far (at S1-
S5 and T1-T5). In each case, we have three predications that
subsume their subject under increasingly general sortals (namely,
‘human’, ‘animal’, and ‘material object’ in S1-S3, and ‘whiteness
trope’, ‘color trope’, and ‘trope’ in T1-T3), and two predications
whose truth appears to follow from purely formal features of reality
(namely, S4-S5 and T4-T5, since everything is such that it both
exists and is identical with itself). Given the relationship between
the members of these two sets, it seems plausible to say that
anything that makes true the rst member (S1 or T1) will thereby
make true the others (S2-S5 or T2-T5).
Now contrast these two sets of predication, where it is plausible to
suppose we have a single truthmaker, with the following sorts of
predication, which are supposed to be true of God on traditional
theism:

(G1) God is divine.

(G2) God is good.

(G3) God is powerful.

(G4) God is wise.

(G5) God is just.

Unlike the members of the other two sets of predications we’ve

considered, G1-G5 do not appear to involve sortals related as
general to speci c; nor do any of them appear to be true solely in
virtue of purely formal features of reality. We could, of course,
always add further predications to this e ect, but that wouldn’t help
with an explanation of how God can be the truthmaker for each of
the predications already listed.

A little re ection, however, reveals that G1-G5 are more closely

connected than they might initially seem. Traditional theists
standardly derive the intrinsic divine attributes (or better, the truth
of predications involving them) from their understanding of the
divine nature. That is to say, they take God to be not only good,
powerful, wise, and just, but to be all these things in virtue of being
divine.31
 Of course, there is a real question whether the list of divine
attributes we arrive at following either of these two procedures will
include all the things traditional theists have wanted to say about
God. For our purposes, however, that is neither here nor there.
Provided we insist—as traditional theists have, and as certainly
seems coherent to do—that all God’s (non-formal) intrinsic attributes
derive from the divine nature, then it will follow that predications
such as those at G1-G5 are related in roughly the way that
predications subsuming their subject under increasingly general
sortals are related. That is to say, they will be related in such a way
that anything making the rst predication true will also make the
others true.

In the end, therefore, there would appear to be nothing absurd

or incoherent about asserting any of the following claims: (1) the
truths expressed by positive atomic predications have truthmakers;
(2) that such truthmakers can be referred to by the corresponding
abstract terms; (3) that a single thing is the truthmaker for many
distinct truths; and (4) a simple God is the truthmaker for all his
true intrinsic predications, provided that the truth of these
predications either follows from purely formal features of reality (as
in the case of ‘God exists’ or ‘God is identical with himself’) or can
be explained in terms of the truthmaker of the predication ‘God is
divine.’ And since the coherence of these claims is all that’s required
to secure the coherence of the truthmaker interpretation of divine
simplicity, I conclude that this interpretation provides us with all we
need to resolve the chief contemporary di culty with the doctrine.

3. SIMPLICITY AND CONTINGENCY

Apart from questions about its coherence, the main contemporary

objection to divine simplicity has focused on its apparent exclusion
of contingent divine volition and knowledge.32 As indicated earlier,
most traditional Jews, Christians, and Muslims believe that God
possesses certain of his mental states contingently—more
speci cally, that he could have chosen to create a universe di erent
from the actual one (or none at all), and hence could have known
di erent things than he actually knows. But this seems incompatible
with divine simplicity. For in order to will or know something
di erent, it would seem that God himself would have to be
di erent. But divine simplicity requires that he be the same across
all possible worlds.

It is important to see that this problem is not one that the

truthmaker interpretation can, by itself, resolve. If God possesses
certain acts of will or knowledge contingently, then it would seem
to follow that there are true divine predications that are both
intrinsic and contingent:

(G6) God freely chooses to create the universe.

(G7) God knows that p, where p is a contingent truth.

Insofar as these predications are intrinsic, my interpretation requires
God himself to be their truthmaker. But insofar as these same
predications are contingent, it would also seem that God cannot be
their truthmaker. For as already indicated, truthmaking is a form of
broadly logical necessitation. Hence, if an entity E is a truthmaker
for a predication P, then E must necessitate that P.

Let us call this ‘the problem of contingency’. In this section, I

respond to the problem by denying that predications such as G6 and
G7 must be regarded as genuinely intrinsic. As will become clear,
once this has been denied the truthmaker interpretation ts well
with the assumption that God is capable of contingent volition and
knowledge.

3.1. Contingent Divine Volition

To begin, let us focus on predications such as G6, which involve the

ascription of contingent divine volitions. It is natural to think of
such volitions as intrinsic features of God, especially if we are
thinking of them on the model of contingent human volitions. But
this is not the only way to think of them. In order to see why, let us
approach the matter somewhat indirectly—via a comparison of two
di erent accounts of human volition.

Suppose some agent A freely chooses on some occasion to

perform a speci c action, even though A could just as easily have
chosen to do something else. We might describe this situation by
saying that, in the actual world, w1, A has volition V1, even though
in another possible world, w2, A has volition V2—where V1 and V2,
let us grant, are distinct intrinsic states of A. Now, if we think about
such volitions on a libertarian—or more speci cally, an agent-causal
—model according to which volitions are the irreducible products of
free agents, it will turn out that w1 and w2 share the same causal
history and laws of nature, and yet di er because in w1 A directly
causes V1, whereas in w2 A directly causes V2 (see Fig. 5.1). No
doubt this understanding of the di erence between w1 and w2, like
the agent-causal model of volition it presupposes, is controversial.
But we can, I think, safely presuppose its coherence here. For once
again, it would be a major advance for the doctrine if critics were
forced to admit that its plausibility is on a par with that of agent
causation.

Figure 5.1. Agent-causal model of volition

How is all this relevant to the contingency problem? Note rst of

all that on the agent-causal model, there are no intrinsic di erences
in virtue of which A causes its volitions. That is to say, if we want to
distinguish worlds w1 and w2, we have no choice but to appeal to
the immediate e ects of A’s agent-causal activity in these worlds—
namely, volitions V1 and V2. Now up to this point I’ve been
assuming that V1 and V2 are intrinsic features of A—that is,
properties or states possessed by but distinct from A. Given this, the
model is obviously inapplicable to a simple God. Nonetheless, it is
easy enough to modify the model so as to make it applicable: simply
‘kick away’ the distinct volitions and think of our agents as directly
causing the e ects themselves.33 Initially, such a kicking away
might seem implausible. Since our volitions are not ordinarily
su cient to produce their immediate e ects, we have no choice but
to distinguish the direct objects of our agent-causal activity (the
volitions) from the e ects that indirectly (and perhaps only
partially) come about via them. But let us suppose—as seems
plausible at least in the case of certain agents, such as God—that the
connection between the objects of our agent-causal activity and
their e ects is non-contingent. In that case, it will be possible to
kick away the volitions and thus to revise our original account of
the di erence between worlds w1 and w2 as follows: in w1, whereas
in w2 A directly causes E2, where E1 and E2 are what we might
earlier have described as the immediate e ects of V1 and V2 (see
Fig. 5.2).

On this revised agent-causal model, there are no distinct

volitions (or for that matter, any distinct intrinsic features
whatsoever) in virtue of which A causes its e ects. Even so, it would
be a mistake to deny that A freely brings about its e ects on this
model—especially if we add, as often seems plausible, that A has
some reasons that partly explain (in the only way that reasons can
explain actions on an agent-causal theory) the speci c e ects it
brings about.34 Rather than deny that A has any volitions or choices
on this model, therefore, we can instead simply identify A’s volitions
or choices with A’s direct acts of agent causing.35

Figure 5.2. Revised agent-causal model of volition

By now it should be clear that the revised agent-causal model

provides a straightforward way of handling contingent divine
volition, one that violates neither the spirit nor the letter of divine
simplicity as I’ve interpreted it. Consider again our representative
example of a predication involving contingent divine volition:

(G6) God freely chooses to create the universe.

On the revised agent-causal model, this predication turns out to be
straightforwardly extrinsic. To say that God freely chooses to create
the universe is just to say that he stands in a certain relation to
something wholly distinct from himself—namely, the relation of
agent causation. But as we have seen, this is a relation in which an
agent can stand to di erent things in di erent worlds, despite any
lack of intrinsic di erences across those worlds.36 In any case, once
G6 is taken to be extrinsic, there is no pressure to interpret God
alone as its truthmaker. Indeed, as far as my interpretation of
simplicity is concerned, there is no pressure to say it has any
truthmaker at all (since my interpretation requires truthmakers only
for positive atomic predications of the form ‘a is F’). Even so, there
is a truthmaker ready to hand, if one is wanted—namely, God, the
universe, and the relation of agent causation. Nor is there any
danger of this truthmaker’s violating divine aseity, since both the
universe and God’s relation to it contingently depend for their
existence (or obtaining) on God rather than the other way around.
Assuming, therefore, as seems plausible, that other predications
involving contingent divine volition can be handled similarly to G6,
we appear to have a perfectly general solution to the problem of
contingency for divine volitions.

3.2. Contingent Divine Knowledge

But we are not quite out of the woods yet. For as we have seen, the
problem of contingency arises for ascriptions of contingent divine
knowledge as well as volition. Initially, the problem here might
seem no more worrisome than in the case of divine volition. Just as
we can treat predications involving contingent divine volition as
extrinsic, so too it might seem that we can do the same in the case
of contingent divine knowledge. Consider a speci c instance of the
type mentioned at G7 above:

(G7a) God knows that human beings exist.

The truth of a predication like G7a would seem to require only two
things: (1) that God stand in a certain cognitive relation to the
contingent truth that human beings exist; and (2) that human beings
do in fact exist. But if that’s right, then predications such as G7a will
be extrinsic after all. And assuming our treatment of this sort of
predication generalizes, we would appear to have a general solution
to the problem of contingency for divine knowledge as well as for
divine volition.37

But note that there is a question that arises here that does not
arise for our treatment of divine volition—one that has to do with
the consistency of this account with divine aseity. For whether or
not predications such as G7a are extrinsic, they appear to make God
dependent on something distinct from himself—namely, the objects
of his contingent knowledge. After all, if God knows that human
beings exist, he seems to know this because they exist.

Now, in the case of certain predications, including G7a itself, the

appearance of dependency can perhaps be avoided by noting that
the objects of God’s contingent knowledge depend for their
existence on God himself. For example, since human beings exist (at
least in the rst place) only because God has freely chosen to create
a universe containing them, we can perhaps say that, strictly
speaking, what God depends on for his knowledge that human
beings exist is not anything distinct from himself (namely, human
beings themselves), but only his own free acts of will or choice.

This strategy will not obviously work, however, for all

predications involving contingent divine knowledge. Suppose that
instead of G7a, we consider the following predication:

(G7b) God knows that Smith is freely choosing to mow his lawn.

It is considerably more di cult to treat this predication in the way

we treated G7a above. For unlike the existence of human beings in
general, it’s hard to see how Smith’s free choice to mow his lawn
could depend solely on the divine will—unless, of course, we are
prepared to accept some form of compatibilism about human
freedom. If we are, then we can treat G7b in the same way we
treated G7a. For in that case, we can adopt a compatibilist model of
human volition—that is, one according to which free choices are not
the irreducible products of agents, but the causal consequences of
deterministic processes. Then we can say that Smith’s free choice to
mow his lawn depends on God in roughly the same way that the
existence of human beings in general does—namely, by depending
on God’s choice to create a certain kind of universe (in this case, one
having a certain causal history and laws of nature). And likewise for
all the free choices of all human agents. Of course, it might seem
odd to adopt a compatibilist account of human volition, but not of
divine volition. But perhaps not (or perhaps the compatibilist will
want to extend the account to divine volitions as well). In any case,
once we adopt a compatibilist account of human volition, the
problem vanishes.

But what if we are not prepared to accept any form of

compatibilism? In that case, it is hard to avoid the conclusion that
the truth of G7b violates divine aseity. For in that case, we certainly
won’t be able to say that Smith’s choice depends entirely on God’s
will—since it is a contradiction to say that God causes an agent to
freely choose (in the incompatibilist or libertarian sense) to perform
some action. But what else, besides Smith’s free choice itself, could
God’s knowledge in G7b be said to depend on? The only obvious
suggestion here is that made by the so-called Molinists (i.e. the
followers of the sixteenth-century Spanish Jesuit, Luis de Molina):
God’s knowledge here depends not only on his will (or more
speci cally, on his choice to create a universe with Smith in his
current circumstances), but also on his knowledge of what Smith
would freely do in any circumstances in which he is created and left
free. Rather than solve our problem with G7b, however, it merely
pushes it back a step. For now we have a new predication involving
contingent divine knowledge that needs explaining:

(G7c) God knows that Smith would freely choose to mow his lawn, if
created and placed in his current circumstances.
And here we seem to run up against a predication whose truth really
does violate divine aseity. For what Smith would freely do, if
created and placed in various circumstances, is generally taken by
Molinists to be a brute contingent truth about Smith (or better,
about Smith’s individual essence, which is an abstract object
existing independently of God). Evidently, therefore, the truth of
G7c, and hence the truth of G7b, ultimately requires God to be
dependent on something distinct from himself.38

What all of this shows is that a full solution to the problem of

contingency—that is, one that not only takes care of worries about
contingent divine volition and knowledge, but also is fully
consistent with the chief motivation for divine simplicity historically
—carries a signi cant cost. It requires the truth of compatibilism
(about human, though not divine freedom). Is this cost prohibitive?
No doubt, many will think it is, seeing in it su cient grounds if not
for rejecting divine simplicity wholesale, then at least for detaching
it from its traditional moorings in divine aseity. But we mustn’t be
too hasty here. Such an attitude would be justi ed if compatibilism
were obviously incoherent, absurd, or false. But it isn’t.
Compatibilism has a rich history of supporters, both within and
outside traditional philosophical theology. Indeed, the former often
see compatibilism as a natural consequence of theological doctrines
just as well established as that of creation (most notably,
providence, foreknowledge, predestination, and election). If we are
going to evaluate divine simplicity, therefore, especially in the
context of its original motivation, then we must engage in the kind
of complicated cost-bene t analysis that we have come to expect for
the evaluation of any large-scale metaphysical theory. In such an
analysis, compatibilism will certainly be one, but it will not be—or
at least, shouldn’t be—the only deciding factor.

4. CONCLUSION

My aim in this chapter has been to show that the doctrine of divine
simplicity deserves further consideration than it has yet received
from contemporary philosophers and theologians. If I have achieved
my aim, it will be clear that the standard objections to this doctrine
can all be answered: the doctrine is neither incoherent nor
incompatible with contingent divine volition and knowledge. Of
course, this by itself does not give us reason to accept the doctrine
as true. But it does, I hope, go considerable distance toward showing
its acceptability. Moreover, insofar as negative attitudes toward
simplicity have contributed to the neglect of the considerations
motivating it historically, I hope that my defense will also
encourage their reevaluation. Contemporary philosophers and
theologians often reject traditional views about divine aseity
precisely because they lead to the doctrine of divine simplicity. But
this is a mistake. Indeed, quite apart from simplicity, such views
deserve our attention insofar as they lie at the very heart of the
traditional theism, which is still widely taken for granted, both by
those who follow Anselm in thinking of God as ‘that than which
nothing greater can be conceived’ and by those who follow Aquinas
in thinking of him as ‘that which explains all motion, change, and
contingency’.

NOTES

A portion of this chapter was presented at a meeting of the

International Society of Thomas Aquinas in Prague 2006. I am
grateful to the audience on that occasion for stimulating
comments and discussion. I am also grateful to Michael
Bergmann, Susan Brower-Toland, Thomas Flint, Wesley
Morriston, Daniel Novotny, Alexander Pruss, and Michael Rea
for helpful comments and criticism on earlier drafts. Finally, I
owe a special debt of gratitude to Wesley Morriston for
extended correspondence that resulted in substantial
clari cation of the views developed and defended here.

1. Let us say that an intrinsic predication characterizes things ‘in

virtue of the way they themselves are’, whereas an extrinsic
predication characterizes them ‘in virtue of their relations or
lack of relations to other things’ (cf. Lewis 1986: 61). This
distinction, though pervasive in philosophy, is notoriously
resistant to analysis. In what follows, therefore, I shall have to
rely on the intuitive understanding of it. I would only add that
it must be kept separate from two other familiar distinctions—
namely, (a) that between essential and contingent predication,
and (b) that between non-relational and relational predication.
 Intrinsic predications can be either essential (e.g. ‘God is wise’)
or contingent (e.g. ‘Socrates is wise’). And, even if most
intrinsic predications are non-relational, there are some
intrinsic relational predications (e.g. ‘God is identical with
himself,’ ‘Socrates chooses to go to the marketplace,’ ‘Socrates
has parts’).

2. Martin 1976: 40.

3. Q. Smith 1988: 524 n. 3. The locus classicus for contemporary

di culties with simplicity is Plantinga 1980. There are,
however, a number of works that have contributed in
important ways to the contemporary understanding of these
di culties, including the following: Mann 1982, 1983; Morris
1985; and Stump and Kretzmann 1985.

4. In certain contexts, it might be better to speak of the God of

traditional theism as an entity possessing the sorts of mental
state generally regarded as constitutive of a personal being
rather than a person, so as to leave open the possibility,
insisted on by traditional Christian theists, that there is more
than one person in God. I shall ignore this complication here,
since it is not relevant for my purposes, but cf. Michael Rea’s
contribution to this volume, Ch. 18.

5. Summa Theologiae 1. 2. 3.

6. Cf. Jordan 1983 for further discussion and references,

especially to Plotinus, who is an important source for thinkers
 in the east and west.

7. Again, cf. Plantinga 1980 for the locus classicus for this sort of
objection.

8. Thus, traditional formulations of the doctrine are inconsistent

with certain forms of nominalism—namely, all those that deny
that expressions of the form ‘a’s, F-ness’ can functionas
genuinely referring devices.

9. Since proponents of divine simplicity typically speak of God

being identical with his goodness, his power, and his wisdom
(rather than with goodness, power, wisdom tout court), I have
stated the property account in such a way that it introduces
properties using expressions of the form ‘a’s F-ness’ (rather
than of the form ‘F-ness’). Those philosophers who think of
properties as particular (i.e. the so-called trope theorists) will,
of course, have no di culty with this, but those who think of
properties as universals may nd it puzzling. It should be
noted, however, that strictly speaking PA is neutral as regards
the speci c nature of the properties. If Socrates and Plato are
both human, then according to PA they must each have a
property of humanity. Even so, PA takes no stand on the
question of whether Socrates’s humanity is identical with
Plato’s (and hence whether properties are particular or
universal).

10. Cf. e.g. Leftow 1999; Mann 1982, 1983; Rogers 1996; Stump
and Kretzmann 1985; Vallicella 1992.
11. It is sometimes suggested that these expressions could refer to
states of a airs—that is, to the exempli cation of the
properties by their subjects (cf. Plantinga 1980). But this
suggestion is of little use, from the point of view of divine
simplicity, since identifying God with a state of a airs seems
just as absurd as identifying him with a property. For further
discussion, cf. Brower 2008.

12. The possessive pronouns do not explicitly occur in the Latin,

but I think they are best understood as implicit here (as they
often are in Latin)—especially in light of Anselm’s insistence
that God’s goodness, power, wisdom, and so on must be
conceived of as distinct from our own. For discussion of
Anselm’s theory of properties, in the case of both God and
human beings, see Brower 2004.

13. Cf. again the references cited in n. 10.

14. Cf. Brower 2008.

15. This answer is apparently one whose time has come. Cf. e.g.
Bergmann and Brower 2006; Pruss forthcoming; and Oppy
2003. For a defense of the claim that truthmakers are not only
su cient, but also necessary for making sense of divine
simplicity, cf. Brower 2008.

16. It would, perhaps, be better to speak of truth-explainers rather

than truthmakers, but the terminology of truthmaking has
 become so well entrenched that to do so would require a
departure from standard usage.

17. Assuming, of course, that P exists. For simplicity’s sake, I shall

hereafter ignore this complication and speak of truthmakers as
entities that actually make predications true rather than as
entities that would make such predications true, if they existed.
In doing so, I don’t mean to beg any questions about the
ultimate nature of truth-bearers or to assume that truthmakers
require the existence of corresponding truths.

18. Here I follow Rodriguez-Pereyra 2002: 34, who argues that the
best we can do is to say that an entity E is a truthmaker for a
predication P if and only if E is an entity in virtue of which P is
true, and then illustrate what we mean by ‘being true in virtue
of’ with examples.

19. E.g. Restall 1996 suggests that we can de ne truthmaking in

terms of the notion of non-classical or ‘relevant’ entailment,
and B. Smith 1999 and 2002 claim that we can de ne it in
terms of representation or projection, which is the dual of
necessitation.

20. E.g. Bigelow 1988:128.

21. Cf. Armstrong 1997, 1989.

22. Cf. Armstrong 1989, esp. 116–19. For a more complete

development and defense of this view, cf. Mulligan, Simons,
and Smith 1984.
23. Following Pruss forthcoming, we can put this point more
precisely by saying that, on the truthmaker interpretation, God
is the minimal truthmaker for each of his true intrinsic
predications—where an entity E is a minimal truthmaker for a
predication P just in case E is such that no proper part of it
also makes P true. As Pruss points out, this quali cation is
needed since on some theories of truthmaking, if E is a
truthmaker for P, then so is anything of which E is a part.
Once the quali cation is added, however, the absolute
simplicity of God follows immediately. For if God had any
proper parts, there would be true intrinsic divine predications
(namely, about these parts) whose minimal truthmakers would
not be God (but the parts).

24. Cf. Fox 1987 for defense of the claim that medieval
philosophers operate within some form of truthmaker account
of predication. Cf. Klima 2004 for a defense of the claim that
they also typically adopt some form of trope theory of
properties.

25. For discussion of some of the other objections facing TA,

including worries about the relationship between divine and
human ‘attributes’ on TA, cf. Brower 2008.

26. Cf. Rodriguez-Pereyra 2000 and 2002, ch.1.

27. Cf. e.g. Bigelow 1988: 128–34 and Lewis 2001 for an account
in terms of the super-venience of truth on being.
28. Cf. e.g. Armstrong 1997 and Rodriguez-Pereyra 2005.

29. Again, once we’ve ruled out the possibility of their referring to
states of a airs. Cf. n. 11 above.

30. It may be wondered how, on the truthmaker interpretation,

such expressions as ‘a’s F-ness’ are related to more general
expressions of the form ‘F-ness’, which are also typically taken
to refer to properties (or better, universals). The most natural
answer, it seems to me, is to say that whereas such expressions
as ‘a’s F-ness’ refer to truthmakers for speci c predications—
namely, predications of the form ‘a is F’—expressions such as
‘F-ness’ refer to truthmakers for some predication or other of
the form ‘x is F’. In that case, the truthmaker interpretation
will allow us to say not only that God is identical with his
goodness (i.e. the truthmaker for ‘God is good’), but also that
he is identical with goodness—or better, a goodness (i.e. a
truthmaker for some statement of the form ‘x is good’).

31. As I mentioned earlier, traditional theists di er among

themselves about how exactly the divine nature is to be
conceived. Some, such as Anselm, conceive of the divine
nature in terms of absolute perfection, as that than which
nothing greater can be conceived. For such theists, the
predication at G1 will be shorthand for the claim that God is
an absolutely perfect being, and the predications at G2-G5, as
well as any other intrinsic divine predications, will involve
speci cation of particular perfections. Others, such as Aquinas,
 conceive of the divine nature in terms of aseity. For theists of
this sort, the predication at G1 will be shorthand for
something like the claim that God is an absolutely
independent being, and the derivation of particular divine
attributes will be more indirect (e.g. Aquinas himself derives
complete actuality from independence, and derives the other
attributes from this, arguing that a being who is completely
actual will have all perfections, without limit, and hence be all
good, powerful, wise, and just).

32. Cf. e.g. Craig 2001; O’Connor 1999; Pruss forthcoming; Stump
and Kretzmann 1985; and Stump 2003. This objection is also
taken seriously by traditional proponents of simplicity. Cf. e.g.
Aquinas’s discussion in Summa Theologiae 1.19. 3.

33. According to McCann 2005 (esp. 144–5), something like this

model is required to make sense of God’s creating anything at
all.

34. I’m assuming that in God’s case his reasons for creating remain
the same across all possible worlds, and hence that the only
thing that varies is the particular set of reasons he acts on.

35. For a theory of reasons explanation that ts with agent-

causation in general, cf. O’Connor 1995; and for an attempt to
modify this theory to t with something like the revised agent-
causal model, cf. O’Connor 1999 and also Pruss forthcoming.
36. This is perhaps the idea behind the traditional description of
the di erence between worlds in which God creates di erent
creatures (or none at all) as a di erence in the e ect.

37. Indeed, assuming (as I shall throughout) that divine beliefs can
be explained along the same lines as divine knowledge, we
would appear to have a solution to the problem of contingency
for them (and perhaps all contingent divine mental states) as
well. For a di erent account of divine beliefs, one which treats
them di erently from divine knowledge and requires a fairly
radical form of externalism about content, cf. Pruss
forthcoming.

38. One could try to avoid this problem by adopting a form of what
is known as ‘the-istic activism’—that is, the view that all
abstracta (including individual essences) are mental states of
God. In that case, what Smith would freely do, if created and
placed in various circumstances, would be a brute contingent
truth about God (or God’s mental states). Although this might
seem to help from the point of view of divine aseity (though
cf. Bergmann and Brower 2006 for reservations even about
this), it would be of no use for preserving divine simplicity.
For such brute truths would appear to involve true
predications of God (or God’s ideas) that are not only
contingent but also intrinsic. As we have seen, however, these
are precisely the sorts of predications about God that divine
simplicity excludes.
REFERENCES

ARMSTRONG, DAVID (1989). Universals: An Opinionated Introduction.

Boulder: Westview

——— (1997). A World of States of A airs. Cambridge: Cambridge

University Press.

BERGMANN, MICHAEL, and BROWER, JEFFREY E. (2006). ‘A Theistic

Argument against Platonism (and in Support of Truthmakers and
Divine Simplicity)’, Oxford Studies in Metaphysics 2: 357–86.

BIGELOW, JOHN (1988). The Reality of Numbers: A Physicalist’s

Philosophy of Mathematics. Oxford: Clarendon.

BROWER, JEFFREY E. (2004). ‘Anselm’s Ethics’, in Brian Davies and

Brian Leftow (eds.), The Cambridge Companion to Anselm.
Cambridge: Cambridge University Press, 222–56.

——— (2008). ‘Making Sense of Divine Simplicity’, Faith and

Philosophy 25: 3–30.

CRAIG, WILLIAM L. (2001). God, Time, and Eternity. Dordrecht: Kluwer

Academic.

FOX, JOHN (1987). ‘Truthmaker’, Australasian Journal of Philosophy

65:188–207.

JORDAN, MARK D. (1983). ‘The Names of God and the Being of

Names’, in Alfred J. Freddoso (ed.), The Existence and Nature of
God. Notre Dame: University of Notre Dame Press, 161–90.
KLIMA, GYULA (2004). ‘The Medieval Problem of Universals’, in
Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy
(Winter 2004 Edition), <http://plato.stanford.
edu/archives/win2004/entries/universals-medieval/>, accessed
13 May 2008.

LEFTOW, BRIAN (1999). ‘Is God an Abstract Object’, Noûs 24: 581–98.

LEWIS, DAVID (1986). On the Plurality of Worlds. Oxford: Blackwell.

——— (2001). ‘Truthmaking and Di erence-Making, Noûs 35: 602–

15.

MCCANN, Hugh J. (2005). ‘The Author of Sin?’, Faith and Philosophy

22:144–59.

MANN, WILLIAM (1982). ‘Divine Simplicity, Religious Studies 18: 451–

71.

——— (1983). ‘Simplicity and Immutability in God’, International

Philosophical Quarterly 23: 267–76.

MARTIN, C. B. (1976). ‘God, the Null Set and Divine Simplicity’, in

John King-Farlow (ed.), The Challenge of Religion Today. New
York: Science History Publications, 138–43.

MORRIS, THOMAS V. (1985). ‘On God and Mann: A View of Divine

Simplicity’, Religious Studies 21: 299–318.

MULLIGAN, KEVIN, SIMONS, PETER, and SMITH, BARRY (1984).

‘Truthmakers’, Philosophy and Phenomenological Research 44:
287–321.
O’CONNOR, TIMOTHY (1995). ‘Agent Causation’, in Timothy O’Connor
(ed.), Agents, Causes, and Events: Essays on Indeterminism and Free
Will. New York: Oxford University Press.

——— (1999). ‘Simplicity and Creation’, Faith and Philosophy 16:

405–12.

OPPY, GRAHAM (2003). ‘The Devilish Complexities of Divine

Simplicity’, Philo 6:10–22.

PLANTINGA, ALVIN (1980). Does God Have a Nature? Milwaukee:

Marquette University Press.

PRUSS, ALEXANDERR. (forthcoming). ‘On Two Problems of Divine

Simplicity’, in Jonathan Kvanvig (ed.), Oxford Studies in
Philosophy of Religion, i.

RESTALL, GREG (1996). ‘Truthmakers, Entailment and Necessity’,

Australasian Journal of Philosophy 74: 331–40.

RODRIGUEZ-PEREYRA, GONZALO (2000). ‘What is the Problem of

Universals?’ Mind 109: 255–73.

——— (2002). Resemblance Nominalism: A Solution to the Problem of

Universals. Oxford: Oxford University Press.

——— (2005). ‘Why Truthmakers?’ in Helen Beebee and Juliane

Dodd (eds.), Truthmakers: The Contemporary Debate. Oxford:
Clarendon.

ROGERS, KATHERIN (1996). ‘The Traditional Doctrine of Divine

Simplicity’, Religious Studies 32: 165–86.
SMITH, BARRY (1999). ‘Truthmaker Realism’, Australasian Journal of
Philosophy 77: 274–91.

——— (2002). ‘Truthmaker Realism: Response to Gregory’,

Australasian Journal of Philosophy 80: 231–4.

SMITH, QUENTIN (1988). ‘An Analysis of Holiness’, Religious Studies 24:

511–27.

STUMP, ELEONORE (2003). Aquinas. London: Routledge.

——— and KRETZMANN, NORMAN (1985). ‘ABSOLUTE Simplicity’,

Faith and Philosophy 2:353–82.

VALLICELLA, WILLIAM (1992). ‘Divine Simplicity: A New Defense’, Faith

and Philosophy 9: 508–25.
 CHAPTER 6

OMNISCIENCE

EDWARD WIERENGA

OMNISCIENCE is the divine attribute of possessing complete or

unlimited knowledge. In this chapter we will examine motivations
for taking such a property to be a divine attribute, attempts to
de ne or analyze omniscience, possible limitations on the extent of
divine knowledge, and, nally, objections either to the coherence of
the concept or to its compatibility with other divine attributes or
with widely accepted claims.

SOURCES OF THE IDEA THAT GOD IS OMNISCIENT

Among the principal sources of the idea that God is omniscient are
the numerous biblical passages that attribute vast knowledge to
God. Thomas Aquinas, in his discussion of the knowledge of God
(1945 Summa Theologiae (ST) I. q. 14), cites such texts as Rom. 11:
33 (‘O the depths of the riches and wisdom and knowledge of
God!’), Job 12: 13 (‘With God are wisdom and strength; he has
counsel and understanding.’), and Heb. 4: 13 (‘And before him no
creature is hidden, but all are naked and laid bare to [his] eyes…’).
 Another source of the attribution of omniscience to God derives
from ‘perfect being theology’, usually attributed to St Anselm.
Anselm’s ontological argument attempts to prove the existence of a
being than which no greater can be thought. Re ection on the
concept of a greatest conceivable or a greatest possible being yields
the conclusion that such a being possesses all perfections. In
Anselm’s formulation, ‘God is whatever it is better to be than not’
(1998b: 5), and although the intellectual attribute Anselm derived
from this formula is that God is wise (1998a: 15), later philosophers
included complete knowledge among the perfections.

The requirements of one or another theological doctrine provide

a third source of the idea that God is omniscient. Prominent among
these is the doctrine of divine providence. This doctrine holds that
God has a plan for the world according to which all things are in his
care and work out according to his good will. In Thomas Flint’s
formulation, ‘to see God as provident is to see him as knowingly and
lovingly directing each and every event involving each and every
creature toward the ends he has ordained for them’ (1998: 12). It is
hard to see how God could order things according to such a detailed
plan without having vast, if not unlimited, knowledge.

DEFINING OMNISCIENCE

Standard accounts of omniscience take it to be knowledge of all

truths. That lends itself to the following straightforward de nition:
 (D1) S is omniscient =df for every proposition p, if p is true, then
S knows p.

Several philosophers have proposed variants of this de nition. For

example, Alvin Plantinga (1974: 68) and Stephen Davis (1983: 26)
implicitly or explicitly accept

(D2) S is omniscient =df for every proposition p, if p is true then S

knows p, and if p is false then S does not believe p.

Unless it is possible to know all truths and yet believe some

falsehood, (D2) is equivalent to (D1). But it seems highly
implausible that someone could have knowledge of all truths and
nevertheless believe the denial of one of the propositions he or she
knows to be true. Believing the denial of a proposition would seem
to undermine knowledge of it, and in any event, it seems
implausible to think that someone could know some proposition p,
know that he or she knows, and, hence, believes p, and also know
that he or she believes not-p. But this would have to be possible in
order for someone to know all truths and yet believe a falsehood.
Another variant is stated by Linda Zagzebski (2007: 262):

(D3) S is omniscient =df for every proposition p, either S knows

that p is true or S knows that p is false.

(D3) is also equivalent to (D1), because if S knows all true

propositions, then for any false proposition, p, S knows that p is false;
in other words, if S knows all true propositions and p is false, then S
knows not-p and so knows that p is false.
 ADDITIONAL PROPERTIES OF GOD’S KNOWLEDGE

Another recent proposed de nition of omniscience turns out not to

be equivalent to the ones we have been considering, but it is useful
for what it suggests about another feature of God’s knowledge. Peter
van Inwagen (2006: 26) suggests

(D4) S is omniscient =df for every proposition p, S either believes

p or the denial of p and it is metaphysically impossible that
there is a proposition q such that S believes q and q is false.

(Van Inwagen actually endorses a restricted version of (D4).) The

condition that it be impossible for an omniscient being to believe
mistakenly adds a modal element; a being who believed all and only
truths but who in some non-actual situation believed a falsehood
would not satisfy this condition. Believing, or better, knowing, all
and only truths would seem to constitute complete knowledge, so
the modal clause seems not to be required for omniscience. This sort
of immunity from error, infallibility, does seem to be an epistemic
perfection, however. So a complete account of perfect knowledge
would attribute not only omniscience but infallibility as well.

There is a related attribute that also seems to be part of perfect

knowledge, namely, essential omniscience. Someone who is
essentially omniscient is not merely omniscient but in addition
could not possibly exist without being omniscient. Infallibility does
not entail essential omniscience, because, at least in theory,
someone could be infallible while failing to believe all truths. But
essential omniscience does entail infallibility; someone who is
essentially omniscient could not possibly have a false belief.

Another question that arises about God’s knowledge is whether it

is all occurrent knowledge or whether some of his knowledge is
dispositional. Knowledge of a proposition is occurrent just in case the
knower has that proposition in mind or is currently aware of it.
Knowledge of a proposition is dispositional, very roughly, if the
subject knows the proposition but is not currently thinking of it or
aware of it. Dispositional knowledge is thus a sort of standing
knowledge of which the subject need not be aware. Aquinas
recognized two kinds of discursion in knowledge, the rst of which is
related to dispositional knowledge. One kind of discursion, then, is
‘according to succession only, as when we have actually understood
anything, we turn ourselves to understand something else’ (1945 ST
I.q. 14 a. 7). But he denies that God’s knowledge is like this, because
‘God sees all things together and not successively’ (ibid.). So
Aquinas denies that God’s knowledge is ever merely dispositional,
and that seems to have been the view of most philosophers who
have considered the matter. Recently, however, David Hunt (1995)
has argued that taking God’s knowledge of the future to be
dispositional can help with the problem of reconciling divine
foreknowledge and human free action (see below). But it is di cult
to see how God could fail to be aware of what he knows and it is
not clear why the standard arguments for the incompatibility of
foreknowledge and free action cannot be adapted to the case of
dispositional knowledge.

Aquinas identi ed a second way in which knowledge can be

dispositional, namely, ‘according to causality, as when through
principles we arrive at the knowledge of conclusions’. Aquinas held
that from the fact that God’s knowledge is not discursive in the rst
sense, of rst attending to one proposition and then to another, it
follows that God’s knowledge is not discursive in the second, either.
Two quali cations are in order. First, if God knows propositions,
they certainly stand in logical relations to one another, and that
includes standing in the relation of premises to conclusion.
Aquinas’s claim is that God does not arrive at a conclusion by
deriving it from premises. Second, Aquinas allows that God can
know e ects by knowing a cause, but this is not so much deriving
the e ect from the cause as having such complete knowledge of a
cause that it can be ‘resolved’ into its e ects. More puzzling,
however, is Aquinas’s further claim that God can know everything
there is to know through knowledge of his own essence. On this
view there is a certain unity to God’s knowledge—it is not
fragmented into di erent pieces—and he does not obtain his
knowledge by being the passive recipient of causal factors in the
world. But it is di cult to see how his essence, which presumably is
the same in every possible world, could nevertheless re ect or
reveal di erent contingent facts in di erent possible worlds.

IS GOD’S KNOWLEDGE JUSTIFIED TRUE BELIEF?

The traditional accounts of omniscience treat it as a special case of
knowledge, although perhaps a very special sort of knowledge,
including infallibility, being essential, and being global or uni ed in
a way no human knowledge is. Despite this focus on knowledge,
however, very little philosophical attention has been devoted to
exploring the application of traditional accounts of knowledge to the
divine case. For example, knowledge is traditionally thought to
involve justi ed true belief, plus a ‘fourth condition’ to avoid
counterexamples. (See, for example, Chisholm 1977.) More recent
accounts of justi cation or of warrant (justi cation plus something
else) have emphasized reliably produced beliefs or the proper
functioning of one’s epistemic faculties in conditions for which they
were designed. (See Alston 1989b and Plantinga 1993, respectively.)
But, with only a few exceptions, philosophers have not inquired into
the status of God’s beliefs or the nature of his justi cation.

One exception is William Alston (1989a), who defends the view

that God does not have beliefs, that beliefs are not ‘constituents of
divine knowledge’. Alston considers two ways of developing this
claim. According to the rst, God ‘is in no state that embodies the
complexity of [a] proposition’, so God does not believe any
propositions. But on this account, God does not know any
propositions, either. So, unless belief but not knowledge is
essentially propositional, Alston’s rst suggestion seems less a
departure from the traditional account of knowledge than an
implicit departure from the traditional account of omniscience.
Alston’s second proposal holds that ‘knowledge is a di erent
psychological state from belief (judgment); it is not a belief that
meets certain further conditions’ (ibid. 188). Knowledge, on this
view, is ‘intuitive’. It is ‘immediate awareness’ of a fact. This sort of
knowledge does apparently yield propositional knowledge, since
Alston writes that ‘God’s immediate awareness that p is itself His
knowledge that p.’ Alston thinks that such knowledge need not
involve belief, that it need not be of the ‘true belief + variety.’ It
could just as easily be the case, however, that the psychological
state of having this direct awareness is a state of belief. What makes
it special is that the other conditions for knowledge—truth,
justi cation, etc.—are automatically satis ed in the case of this kind
of belief. In any event, most philosophers have not followed Alston
in this regard and have continued to conceive of God’s knowledge as
including belief.

A second, less explicit, respect in which philosophers have

treated God’s knowledge as satisfying the traditional account is by
asking how God arrives at his knowledge. This may be understood
as a way of asking what justi es God in his beliefs. Alfred Freddoso
inquires with respect to future contingents (propositions about the
future that are contingently true), ‘what is the source of and
explanation for the fact that God knows future contingents...?’
(1988: 1). Freddoso’s constraint on an answer to this question,
stated in his development of the views of Aquinas, Molina, and
Bañez, but presumably one he accepts, is that ‘an adequate answer
to the source-question must essentially appeal to God’s active causal
role in the created world and must eschew even the faintest
suggestion that God’s knowledge of e ects produced in the created
world is causally dependent on the activity of His creatures’. So
what justi es God in his beliefs, on this view, is quite di erent from
what justi es us in many of our beliefs. God is not causally
impinged upon and thus is not justi ed in his beliefs by being
a ected by anything other than himself.

THE EXTENT OF OMNISCIENCE AND PHILOSOPHICAL OBJECTIONS

Questions may be asked about the extent of God’s omniscience,

about exactly what is included in knowledge of all truths, whether
this knowledge leaves anything out, and whether it is compatible
with other attributes of God or with other claims we take to be true.

Knowledge de re

Knowledge (and belief) can be knowledge with respect to a thing

that it has a certain property. Someone can have this de re
knowledge without having the corresponding propositional or de
dicto knowledge. For example, not recognizing the Provost in a dark
alley, Ralph may know of the Provost that he appears suspicious
without believing, and hence not knowing, that the Provost appears
suspicious. In his discussion of omniscience, Arthur Prior (1962)
called attention to the phenomenon of de re knowledge and gave
several examples ‘which a believer in God’s omniscience would wish
to maintain’. If omniscience should thus extend to knowledge de re,
the question arises as to whether our de nition (D1), which only
requires knowledge of true propositions, captures such knowledge.

Philosophers have disagreed as to what exactly is required to

acquire de re belief with respect to an object, but typically they
assume that it is by bearing a relation of acquaintance with the
object, to be epistemically en rapport with it, and then to have some
de dicto belief attributing a property to the object in a way
appropriately connected to that relation. Proposals for bearing the
right relation of acquaintance range from having an appropriate
concept of the object to being in the right causal relation to the
object. (See Chisholm 1976; Lewis 1979; Kaplan 1968.) On all these
accounts, knowledge de re is a species of knowledge de dicto
together with a relation of acquaintance. It might be tempting to
think that whatever that relation of acquaintance should turn out to
be, God bears it to every thing and, thus, has complete de re
knowledge. But that would not be right, in view of the claim in the
last section, if the relevant relation requires being causally impinged
upon by the object. Nevertheless, a related claim seems plausible:
for any object x, whatever relation is such that standing in it to x is
su cient (together with the appropriate de dicto belief) for having a
de re belief about x, God bears a relation to x that is at least as
intimate, one that is at least as good for supporting a de re belief.
God’s creating, sustaining, and direct awareness of contingent things
put him epistemically en rapport with them. So if de re knowledge is
reducible to de dicto knowledge, then God’s satisfying (D1) gives
him complete de re knowledge.

Of course, if it turns out that de re knowledge is not thus

reducible to de dicto knowledge, it would be a simple matter to
revise (D1) so as explicitly to extend omniscience to knowledge de
re, namely, by adding the following second clause to (D1), ‘and for
any x and property P, if x has P, then S knows of x that x has P’.

Indexicals, Temporal Knowledge, and First-Person Knowledge

What we know changes as time goes by. And what we know when
we use presenttense sentences, or sentences with such temporal
indexicals as ‘now’ or ‘today’, often seems particularly eeting or
ephemeral. We can use a sentence such as

(1) It is sunny today

to express a truth one day, and a falsehood the next. Moreover,

what we express with (1) on a given occasion of use seems not to be
equivalent to anything we could express with the corresponding
sentence with the time-indication made explicit, for example, with

(2) It is sunny on 16 June 2008.

If omniscience requires knowledge of all truths, then God must

know whatever proposition we express with a given use of (1), as
well as that expressed by (2), if those propositions are true. So, our
rst question is whether omniscience extends to knowledge de
praesenti, the kind of knowledge we have when we use (1) to express
a truth.

In addition, many philosophers have claimed that the

phenomenon of temporal knowledge can be used to show that if
God is omniscient then he lacks one of the divine attributes
traditionally ascribed to him. For example, Norman Kretzmann
(1966) argued that omniscience is incompatible with immutability,
Nicholas Wolterstor (1975) claimed that omniscience is
incompatible with divine eternity, and Stephen Davis (1983) held
that omniscience is incompatible with divine timelessness. Other
philosophers who urge a version of this claim include Arthur Prior
(1962), Anthony Kenny (1979), and Patrick Grim (1985).

It is instructive that a parallel issue arises for rst-person

knowledge or knowledge of propositions expressed by sentences
containing such rst-person indexicals as ‘I’ or ‘me’. To cite an
example given by Kretzmann (1966), there is a di erence between
the proposition Jones expresses by the sentence

(3) ‘I am in hospital’

and the proposition

(4) Jones is in hospital.

If Jones is hospitalized for amnesia, he might well believe the

proposition expressed by his use of (3) without believing (4). And,
similarly amnesiac but ignorant of his location, Jones can believe
(4) without believing the proposition he would express by (3), if he
reads a newspaper account of his hospitalization. So propositions we
express by sentences using the rst-person personal pronoun (‘ rst-
person propositions’) do not seem to be equivalent to the
corresponding propositions we express without using a rst-person
indexical.

Kretzmann took the rst-person proposition Jones knows when

he knows the proposition his use of (3) expresses to be one that no
one other than Jones could know. This leads to the objection that if
omniscience requires knowledge of all true propositions, then the
existence of an omniscient God is incompatible with the existence of
any other person with self-knowledge or knowledge de se.
Aternatively, in a formulation closer to that given by Patrick Grim
(1985), given that we do have knowledge de se, there is no God who
knows all true propositions.

As in the case of knowledge de praesenti, we face the question of

whether omniscience extends to knowledge de se, in particular to
such rst-person knowledge of others. And there is the further
question of whether facts about knowledge de se lead to good
objections to there being an omniscient God.

The structural similarity between the two types of knowledge, as

well as the sorts of problems they are supposed to create for
omniscience, suggests that they might be given similar rejoinders,
and indeed they have. From the fact that the propositions we know
change over time, it does not follow that the propositions we know
when we know what day it is are not eternally true or that they
cannot be known at other times. Similarly, from the fact that we
know certain propositions about ourselves as ourselves, it does not
follow that the propositions in question could not be known by
someone else. In each case, what is missing in the objection is an
account of exactly what the objects of knowledge and belief are, an
account according to which there are special propositions believable
only at certain times or by certain individuals.

Jonathan Kvanvig (1986) has given an account of the objects of

knowledge according to which what is unique about knowledge de
praesenti and knowledge de se is not a special proposition graspable
only at certain times or by certain individuals but rather a special
kind of access to or grasping of ordinary propositions. According to
Kvanvig, what is involved in knowledge of the present or of oneself
is a ‘direct grasp’ of a proposition—which leaves it open that God
could believe the same propositions without ending up with
knowledge de praesenti or knowledge de se of someone other than
himself.

Edward Wierenga (1989) has made a slightly di erent proposal

according to which sentences containing the rst-person personal
pronoun typically express a proposition involving the haecceity or
individual essence of the person using the sentence. Believing a
proposition involving one’s own haecceity gives one belief de se.
Although, as a matter of fact, we never grasp the propositions
involving the haecceities of others, nothing prevents God from
grasping or believing them. He would end up with a belief de se in
the process only if the haecceity in question were his own. Similarly,
if somewhat less plausibly, sentences with present-time indexicals
express, at a given time of use, propositions involving the haecceity
or individual essence of that time. Believing such a proposition at its
time gives one a belief de praesenti. Although we never grasp such
propositions at any other time, nothing prevents God from grasping
or believing them. He would acquire a de praesenti belief only if he
were present at the time in question. On this proposal, whether God
is in time or timeless, or whether he is immutable, is not settled by
whether he is omniscient but instead by whether he has his beliefs
in time.

Another account of the objects of knowledge and belief is

available, one that does not appeal to a special kind of grasping or
an exotic type of proposition. It is the claim that many propositions
are perspectival. A proposition can be true at one time and false at
another. More generally, a proposition can be true at one index,
consisting of a person and a time (and perhaps a place and a world),
and false at another. Consider

(5) I am in hospital.

On the present proposal, (5) might be true at the index <Jones,

4.30 p.m. on 16 June 1967>, when amnesiac Jones is in hospital,
but false at the index <Jones, 10.30 a.m. on 14 October 2007> by
which time Jones has been cured and discharged. (See Sosa 1983
and Kaplan 1989.)
 If propositions are perspectival in this way, there is a distinction
between believing that a proposition is true at an index and believing at
an index that a proposition is true. An example of the former is
believing

(6) I am in hospital is true at <Jones, 4.30 p.m. on 16 June

1967>.

An example of the latter is believing (5) at the index <Jones, 4.30

p.m. on 16 June 1967>. Anyone can have the former belief; only
someone at the index in question can have the latter. In particular,
only Jones can have beliefs at that index, and even that he can only
do at the time of the index. In our story, Jones did believe (5) at
<Jones, 4.30 p.m. on 16 June 1967>, but he did not believe (6)
then, for at that time he did not know that he was Jones. Belief de
praesenti is thus belief in any proposition at an index, and belief de
se is belief in a rst-person proposition at an index; both are
instances of believing a perspectival proposition at an index.

Wierenga (2002) has shown how this account may be applied to

omniscience. God should not be expected to have beliefs impossible
for him to have. Thus, he should not be expected, even though
omniscient, to believe (5) at <Jones, 4.30 p.m. on 16 June 1967>,
because that is possible only for Jones. Omniscience would require
him to know only those perspectival propositions true at whatever
indices he is at. Thus, we could revise (D1) as
 (D4) S is omniscient =df for any proposition p and perspective
<x, t>, (i) if p is true at <x, t> then S knows that p is
true at <x, t>, and (ii) if S is at <x, t> and p is true at
<x, t>, then at <x, t> S knows that p.

According to (D4), God can be omniscient without having the de se

beliefs of others, and it would follow from his omniscience that he is
in time only on the additional assumption that he has his beliefs at
temporal indices. But of course if it could be shown that God has
beliefs at temporal indices, it would follow straightaway that he is
in time, and arguing for that conclusion by way of appeal to
omniscience is simply otiose.

Foreknowledge and Free Action

If omniscience requires knowledge of all truths and there are truths

about the future, then omniscience includes foreknowledge. The
simplest argument for this conclusion, considered by both Augustine
and Boethius, can be stated as:

(7) If God knows in advance that a person will perform a certain

action, then that person must perform it.

(8) If a person must perform an action, then the person does not
do so freely.

(9) If God knows in advance that a person will perform a certain

action, then the person does not do so freely.
Although this argument enjoys occasional support, its aw had been
identi ed by Aquinas (1945: i 14, 13, ad 3). The rst premise is
ambiguous. It may be taken as

(7′) Necessarily, if God knows that a person will perform a

certain action, then the person will do so,

which is true, but so taken the conclusion does not follow. The
conclusion does follow if the premise is

(7″) If God knows that a person will perform a certain action,

then the proposition that the person will perform that
action is necessarily true,

but this premise is false. So neither interpretation yields an

argument that is both logically valid and has true premises.

A considerably more di cult argument for the incompatibility of

divine foreknowledge and free human action appeals to the
apparent xity or necessity of the past, what William of Ockham
called the accidental necessity of the past. This argument requires
three principles about accidental necessity:

(10) True contingent propositions reporting God’s past beliefs are

forever after accidentally necessary.

(11) A contingent proposition entailed by an accidentally

necessary proposition is itself accidentally necessary.

(12) If a proposition is accidentally necessary at a time, then no

one is able at any later time to make it false.
Suppose, to take an example introduced into the literature by
Nelson Pike (1965), that eighty years ago God believed that Jones
will mow his lawn tomorrow. Since this is a contingently true
proposition reporting a past belief of God’s, by (10), it is
accidentally necessary, that is,

(13) It is accidentally necessary that eighty years ago God

believed that Jones will mow his lawn tomorrow.

At this point the argument needs a slightly stronger assumption than

that God is omniscient. It requires the assumption either that God is
essentially omniscient or that he is infallible (see sect. 3 above).
Inasmuch as perfect knowledge includes essential omniscience, let
us make this assumption. Then the proposition that eighty years ago
God believed that Jones will mow his lawn tomorrow entails that
Jones will mow his lawn tomorrow. So by (11), the principle that
accidental necessity is closed under (contingent) entailment, it
follows that

(14) It is accidentally necessary that Jones will mow his lawn

tomorrow.

But from (14) and (12) it follows that there is nothing anyone can
do to make (14) false. In particular, there is nothing Jones himself
can do to make (14) false. Mowing his lawn tomorrow is a chore he
is unable to shirk. So Jones is not free with respect to mowing his
lawn tomorrow. This argument may, of course, be generalized to
cover any human action that God has foreknown.
 The philosophical literature on this topic is voluminous. Many
important papers are collected in Fischer (1989); see also Hasker
(1989), Zagzebski (1991), and Wierenga (1989) for engagement
with and further references to that literature. Responses fall into
several di erent categories. One response is to accept the conclusion
that divine foreknowledge is incompatible with human free action
but to attempt to mitigate it by claiming that God’s omniscience
does not include foreknowledge. This was the response of Boethius
(1999), who held that God’s mode of existence was that of eternity,
‘the complete possession all at once of illimitable life’ (see Stump
and Kretzmann, 1981). Thus, although God knows everything that
ever happens, it is not foreknowledge from his eternal perspective.
Aquinas, too, seemed to have thought that the correct response to
the problem was to appeal to divine eternity. More recently, Stump
and Kretzmann (1991) speak of ‘the eternity solution’. But, as
several philosophers have argued (Plantinga 1986; Zagzebski 1991;
Wierenga 1989), whatever the merits of the doctrine of divine
eternity, it does not by itself solve the problem raised by this
argument. The reason is that an exactly analogous argument can be
constructed in terms of the past truth of God’s eternal knowledge of
propositions about our future. The defender of divine eternity would
need to say something about that argument, so we should nd out
what that is so that we can see whether it applies to the present
argument.
 Another response also accepts the conclusion of the argument
but denies both that God has foreknowledge and that he has any
other way of knowing the future free actions of creatures. This view
has become increasingly popular in recent years. A limited version
of it was defended by Peter Geach (1977), who held that God lacks
foreknowledge of future free actions of creatures unless those
actions result from ‘present trends and tendencies’. Swinburne
(1993) agrees, at least on the assumption that God is contingent,
that he does not have knowledge of future free actions. Ho man
and Rosenkrantz (2002) develop a detailed analysis of omniscience,
intentionally limiting God’s knowledge of the contingent future to
those truths that are ‘causally inevitable’. Finally, William Hasker
(1989) and others associated with the ‘Open Theism’ movement (see
Pinnock et al. (1994)) are insistent that God leaves the future open
so as to leave room for human freedom. Nevertheless, denying that
God has foreknowledge (without substituting eternal knowledge or
some other kind of cognition for it) seems a retreat from the
traditional conception of the God of the philosophers, and it
certainly seems di cult to see how a signi cant doctrine of
providence could be developed without including knowledge of the
future in omniscience.

Appealing to divine eternity and abandoning foreknowledge of

future free actions are two options open to philosophers who accept
the argument above. Philosophers who do not want to accept the
conclusion of that argument will have to deny one or more of its
assumptions (10), (11), and (12). Ockhamism denies that
propositions reporting God’s past beliefs about the future need be
accidentally necessary. Ockhamists notably include, in addition to
William of Ockham (1969) himself, Alvin Plantinga (1986). They
hold that propositions reporting God’s past beliefs may be partly
about the future, just as

(15) Twenty years ago Ralph correctly believed that the sun will
rise on 1 January 2010

is not wholly about the past. It has turned out to be remarkably

di cult to give a satisfactory account of the key concepts of
accidental necessity or of being wholly about the past. Nevertheless,
if God’s past beliefs about the past are not now accidentally
necessary, the Ockhamist has found a premise of the argument to
deny with some plausibility.

Molinists deny (11), the claim that accidental necessity is closed

under entailment of contingent propositions. Molinism, to be
discussed in the next section, defends the claim that God has ‘middle
knowledge’, and it provides a plausible explanation of how God
could know the future free actions of human beings without causing
those actions or otherwise destroying their freedom. This central
feature of Molinism seems independent of any of the premises of the
argument under review, so it is not clear why Molinism picks (11)
to deny. Nevertheless, Freddoso (1988: 58), himself a convert to
Molinism, has found a passage in Molina that pretty clearly seems to
deny (11). There seems to be a good reason to accept (11), however.
If accidental necessity is indeed a well-behaved modality it must
obey two principles. The rst is that

(16) Any proposition that is logically equivalent to an

accidentally necessary proposition is itself accidentally
necessary.

If a proposition p is xed and unalterable in virtue of what has

happened, if there is nothing anyone can do that would render p
false, then it is hard to conceive how a proposition q that is logically
equivalent to p is not similarly xed and beyond anyone’s ability to
render false. The second principle is that

(17) Any conjunction that has as a conjunct a contingent

proposition that is not accidentally necessary is itself not
accidentally necessary.

Suppose there is something you could do that would make p false. If

you were to do it, then for any proposition q, the conjunction p & q
would be false, too; so the truth of the conjunction is not xed or
unalterable. If accidental necessity satis es these two modest and
intuitive principles, then it is closed under (contingent) entailment.
To see this, suppose that p is accidentally necessary and that it
entails a proposition q (where q is possibly false). Since p entails q, p
is logically equivalent to the conjunction of p & q. By (16), then,
this conjunction is accidentally necessary. But if this conjunction is
accidentally necessary, then by (17) it does not have any contingent
conjuncts that are not accidentally necessary. By assumption, q is a
contingent conjunct of the conjunction p & q. So q is accidentally
necessary. Hence, if p is accidentally necessary and entails a
contingent proposition q, then q is accidentally necessary. So
accidental necessity is closed under (contingent) entailment.

The nal option is to deny (12), the claim that no one can make
an accidentally necessary proposition false. More carefully, it is
open to hold that from the fact that there is nothing Jones can do to
make

(14) It is accidentally necessary that Jones will mow his lawn

tomorrow

false, it does not follow that he is not free with respect to mowing
his lawn tomorrow. This response will be congenial to compatibilists
—those who hold that free action is compatible with causal
determinism. But it will be much less palatable to those who think
that free will is libertarian.

Middle Knowledge

A nal question about the extent of God’s knowledge is whether it

extends to include middle knowledge. This is a question that
requires some explanation. Luis de Molina (Freddoso 1988) and
other late-medieval philosophers divided God’s knowledge into
several categories. Part of what God knows is natural knowledge,
that is, knowledge of necessary truths, of possibilities, and of the
natures of things. Two features of natural knowledge are that it is of
propositions that are both necessarily true and independent of God’s
will or decision to create. Another category is free knowledge,
knowledge of propositions true because God decides to have them
be true. So this category includes propositions that are both
contingent and dependent upon God’s will. An especially interesting
category of propositions are those that describe what free agents
would freely do in a variety of situations (‘counterfactuals of
freedom’). Molina’s insight was to categorize these propositions as
intermediate between the other two categories, in two dimensions:
they are contingently true but independent of God’s will. Hence,
knowledge of propositions such as these is naturally thought of as
middle knowledge, in the middle of natural and free knowledge.

Thomas Flint (1998) has described how on Molinism God’s

knowledge ‘grows’ through four ‘logical moments’. First, God has
‘natural knowledge’, that is, knowledge of necessary truths. Second,
he has middle knowledge, knowledge, that is, of contingent truths
beyond his control, preeminently, counterfactuals of creaturely
freedom. Third, God decides upon a particular divine creative act of
will, and thus knows what his direct contribution to the world is.
Finally, by using his knowledge of his own creative action as well as
his knowledge of what would happen if he were to do that action,
God arrives at knowledge of all remaining contingent truths. This
nal category includes propositions describing what free agents will
do in the future. But it is clear from the way God arrives at this
knowledge that his knowledge exerts no causal or determining
in uence on the actions of free agents.

Attributing middle knowledge to God has been employed in

ways that have been enormously fruitful in recent philosophy of
religion, ranging from a response to the problem of evil (see Ch.
15), to an account of God’s knowledge of future contingents, to a
developed account of divine providence (see Ch. 12). But the claim
that omniscience includes middle knowledge has been controversial.
Molina’s sixteenth-century opponents denied that counterfactuals of
freedom really were independent of God’s will. Contemporary
opponents deny that they are true, or that they are true soon enough
to be of use to God in deciding what to create (Adams 1977; Hasker
1986; Kenny 1979). Philosophers who hold that the counterfactuals
of freedom needed to guide God’s creation decision are not true
typically claim that they are not ‘grounded’, that prior to creation
there is nothing that would make them true (Adams 1977).
Proponents often deny that all contingent propositions need a
ground of truth. But consideration of this dispute is beyond the
scope of this chapter.

CONCLUSION

Omniscience may be understood as perfect knowledge and, thus, as

knowledge of all truths. Perfect knowledge would also have
additional features, such as, infallibility, being essential, and being
whole or non-discursive. It is a matter of controversy as to whether
omniscience extends to such particular cases as knowledge de
praesenti, knowledge de se, knowledge of the future, including future
free actions, and middle knowledge, especially knowledge of what
free creatures would do in alternative circumstances. On these
topics, philosophers have taken opposing sides and debate
continues.

REFERENCES

Adams, R. M. (1977). ‘Middle Knowledge and the Problem of Evil’,

American Philosophical Quarterly 14: 109–17.

ALSTON, W. P. (1989a). ‘Does God Have Beliefs?’, in id., Divine

Nature and Human Language. Ithaca: Cornell University Press.

——— (1989b). Epistemic Justi cation: Essays in the Theory of

Knowledge. Ithaca: Cornell University Press.

ANSELM (1998a). Monologion, in B. Davies and G. R. Evans (eds.),

Anselm of Canterbury: The Major Works. Oxford: Oxford
University Press.

——— (1998b). Proslogion, in B. Davies and G. R. Evans (eds.),

Anselm of Canterbury: The Major Works. Oxford: Oxford
University Press.

AQUINAS, THOMAS (1945). The Basic Writings of Thomas Aquinas, ed. A.

C. Pegis, 2 vols. New York: Random House.
BOETHIUS (1999). The Consolation of Philosophy, trans. P. G. Walsh.
Oxford: Oxford University Press.

CHISHOLM, R. (1976), ‘Knowledge and Belief: “De Dicto” and “De Se”’
Philosophical Studies 29: 1–20.

——— (1977). Theory of Knowledge, 2nd edn. Englewood Cli s:

Prentice-Hall.

DAVIS, S. (1983). Logic and the Nature of God. Grand Rapids: Wm. B.
Eerdmans.

FISCHER, J. M. (1989). God, Foreknowledge, and Freedom. Stanford:

Stanford University Press.

FLINT, T. (1998). Divine Providence: The Molinist Account. Ithaca:

Cornell University Press.

FREDDOSO, A. (1988). ‘Introduction’ to Luis Molina, On Divine

Foreknowledge: Part IV of the Concordia. Ithaca: Cornell
University Press.

GEACH, P. (1977). Providence and Evil. Cambridge: Cambridge

University Press.

GRIM, P. (1985). ‘Against Omniscience: The Case from Essential

Indexicals’, Noûs 19:151–80.

HASKER, W. (1986). ‘A Refutation of Middle Knowledge’, Noûs 20:

545–57.

——— (1989). God, Time, and Knowledge. Ithaca: Cornell University

Press.
HOFFMAN, J., and ROSENKRANTZ, G. (2002). The Divine Attributes.
Oxford: Blackwell, ch. 6.

HUNT, D. (1995). ‘Dispositional Omniscience’, Philosophical Studies

80: 243–78.

KAPLAN, D. (1968). ‘Quantifying In’, Synthese 29: 178–214.

——— (1989). ‘Demonstratives’ (with ‘Afterthoughts’), in J. Almog,

J. Perry, and H. Wettstein (eds.), Themes from Kaplan. New York:
Oxford University Press.

KENNY, A. (1979). The God of the Philosophers. Oxford: Clarendon.

KRETZMANN, N. (1966). ‘Omniscience and Immutability’, Journal of

Philosophy 63: 409–21.

KVANVIG, J. (1986). The Possibility of an All-Knowing God. New York:

St Martin’s.

LEWIS, D. (1979). ‘Attitudes De Dicto and De Se’, Philosophical Review

88: 513–43.

PIKE, N. (1965). ‘Divine Omniscience and Voluntary Action’,

Philosophical Review 74: 27–46.

PINNOCK, C., et al. (1994). The Openness of God. Downers Grove:

InterVarsity Press.

PLANTINGA, A. (1974). God, Freedom, and Evil. rpt. Grand Rapids:

Wm. B. Eerdmans, 1977.
——— (1986). ‘On Ockham’s Way Out’, Faith and Philosophy 3: 235–
69.

——— (1993). Warrant and Proper Function. New York: Oxford

University Press.

PRIOR, A. N. (1962). ‘Formalities of Omniscience’, in id., Papers on

Time and Tense. Oxford: Oxford University Press.

SOSA, E. (1983). ‘Consciousness of the Self and of the Present’, in J.

Tomberlin (ed.), Agent, Language, and the Structure of the World.
Indianapolis: Hackett.

STUMP, E., and KRETZMANN, N. (1981). ‘Eternity’, The Journal of

Philosophy 78: 429–58.

——— (1991). ‘Prophecy, Past Truth, and Eternity’, in J. Tomberlin

(ed.), Philosophical Perspectives 5: 395–424.

SWINBURNE, R. (1993). The Coherence of Theism. rev. edn. Oxford:

Clarendon, ch. 10.

VAN INWAGEN, P. (2006). The Problem of Evil: The Gi ord Lectures

Delivered in the University of St Andrews in 2003. Oxford:
Clarendon.

WIERENGA, E. (1989). The Nature of God. Ithaca: Cornell University

Press.

——— (2002). ‘Timelessness out of Mind’, in G. Ganssle and D.

Woodru (eds.), God and Time: Essays on the Divine Nature. New
York: Oxford University Press.
WILLIAM of OCKHAM (1969). Predestination, God’s Foreknowledge, and
Future Contingents, trans. M. M. Adams and N. Kretzmann. New
York: Appleton-Century-Crofts.

WOLTERSTORFF, N. (1975). ‘God Everlasting’, in C. Orlebeke and L.

Smedes (eds.), God and the Good: Essays in Honor of Henry Stob.
Grand Rapids: Wm. Eerdmans.

ZAGZEBSKI, L. (1991). The Dilemma of Freedom and Foreknowledge.

New York: Oxford University Press.

——— (2007). ‘Omniscience’, in C. Meister and P. Copan (eds.), The

Routledge Companion to Philosophy of Religion. London:
Routledge, 261–9.
 CHAPTER 7

DIVINE ETERNITY

WILLIAM LANE CRAIG

‘GOD’, declares the prophet Isaiah, ‘is the high and lofty One who
inhabits eternity’ (Isa. 57:15). But being a prophet and not a
philosophical theologian, Isaiah did not pause to re ect up on the
nature of divine eternity. Minimally, to be eternal means to exist
without beginning and end. To say that God is eternal means
minimally that he never came into being and will never go out of
being. To exist eternally is to exist permanently.1

There are, however, at least two ways in which something could

exist eternally. One way would be to exist omnitemporally
throughout in nite time. In this case God would have an
immemorial and everlasting temporal duration. The other way in
which a being could exist eternally would be by existing timelessly.
In this case God would completely transcend time, having neither
temporal location nor temporal extension. He would simply exist in
an undi erentiated, timeless state.

If we take Scripture as our guide in matters of theology, the

initial question we must ask is: does biblical teaching on divine
eternity favor either one of these views? The question turns out to
be surprisingly di cult to answer. On the one hand, it is
indisputable that the biblical writers typically portray God as
engaged in temporal activities, including foreknowing the future
and remembering the past, and when they speak directly of God’s
eternal existence they do so in terms of beginningless and endless
temporal duration. The data are not wholly one-sided, however.
There is some evidence, at least, that when God is considered in
relation to creation he must be thought of as the transcendent
creator of time and the ages and therefore as existing beyond time
(Gen. 1: 1; Prov. 8: 22–3; 1 Cor. 2: 7; 2 Tim. 1: 9; Titus 1: 2–3; Jude
25). So the biblical data are underdeterminative, and one seems
forced to conclude with James Barr that ‘if such a thing as a
Christian doctrine of time has to be developed, the work of
discussing it and developing it must belong not to biblical but to
philosophical theology’.2

At issue here is God’s relationship to time: Does God exist

temporally or atemporally? God exists temporally if and only if he
exists in time, that is to say, if and only if his duration has phases
that are related to each other as earlier and later. In that case, God,
as a personal being, has experientially a past, a present, and a
future. No matter what moment in time we pick, given God’s
permanence, the assertion, ‘God exists now’, were we to make it,
would be literally true.
 By contrast, God exists atemporally if and only if he is not
temporal. This de nition makes it evident that temporality and
timelessness are contradictories: an entity must exist one way or the
other and cannot exist both ways at once. If, then, God exists
atemporally, he has no past, present, and future. At any moment in
time it would be true to assert, ‘God exists’, in the tenseless sense of
‘exists’, as when one says, ‘The natural numbers exist’, but not true
to assert, ‘God exists now’.

Philosophical theologians have been sharply divided with respect

to God’s relationship to time. What are the principal arguments they
have o ered for divine timelessness and temporality?

ARGUMENTS FOR DIYINE TIMELESSNESS

Argument from Simplicity or Immutability

Consider rst the view that God exists timelessly. Traditionally

Christian theologians such as Thomas Aquinas argued for God’s
timelessness on the basis of his absolute simplicity and immutability
(Summa theologiae (ST) Ia q. 10 a. 3). The argument can be simply
formulated. As a rst premise, we assume either

(1) God is simple

or

(1′) God is immutable.

Then we add
 (2) If God is simple or immutable, then he is not temporal,

from which it follows that

(3) Therefore, God is not temporal.

Since temporality and timelessness are contradictories, it follows

that

(4) Therefore, God is timeless.

Is this a cogent argument? Consider (2). The doctrine of divine

simplicity implies not merely that God does not have parts, but that
he does not possess even distinct attributes. In some mysterious way
his omnipotence is his goodness, for example. He stands in no
relations whatsoever. His nature or essence is not even distinct from
his existence. He is, Aquinas tells us, the pure act of existing (ST Ia
q. 3 a. 4). Now if God is simple in the way described, it obviously
follows that he cannot be temporal. For a temporal being is related
to the various times at which it exists: it exists at t1 and at t2, for
example. But a simple being stands in no real relations. Moreover, a
temporal being has phases of its life that are not identical but rather
are related to one another as earlier and later. But an absolutely
simple being could not stand in such relations and so must have its
life, as Boethius put it, ‘all at once’ (totum simul) (Philosophiae
consolationis 5 pr. 6. 9–11).

Similarly, if God is immutable, then even if he is not simple, he

cannot be temporal. Like simplicity, the immutability a rmed by
the medieval theologians is a radical concept: God cannot change in
any respect. God not only cannot undergo intrinsic change, but he
also cannot change extrinsically by being related to changing
things.3 But obviously a temporal being undergoes at least extrinsic
change in that it exists at di erent moments of time and, given the
reality of the temporal world, coexists with di erent temporal
beings as they undergo intrinsic change. Even if we relax the
de nition of ‘immutable’ to mean ‘incapable of intrinsic change’ or
the even weaker concept ‘intrinsically changeless’, an immutable
God cannot be temporal. For if God is temporal, God will be
constantly changing in his knowledge, knowing rst that ‘It is now
t1’ and later that ‘It is now t2.’ God’s foreknowledge and memory
must also be steadily changing, as anticipated events transpire and
become past. God will constantly be performing new actions, at t1
causing events at t1 and at t2 causing events at t2. Thus, a temporal
God cannot be changeless. It follows, then, that if God is immutable,
he is timeless.

Thus, God’s timelessness can be deduced from either his

simplicity or immutability. Is this a good reason for thinking that
God is timeless? That depends on whether we have any good reason
to a rm (1) or (1′). Here we run into severe di culties. For
doctrines of divine simplicity and immutability that are su ciently
strong to support divine timelessness are even more controverted
than the doctrine of divine timelessness itself. These strong
doctrines nd no explicit support in Scripture, which at most speaks
of God’s immutability in terms of his faithfulness and unchanging
character (Mal. 3: 6; Jas.1: 17), and philosophically there seem to be
no good reasons to embrace these doctrines but weighty objections
against them.4 These cannot be discussed here; the point is that (1)
and (1′) are even more di cult to prove than (4), so that they do
not constitute good grounds for believing (4). Thus, while we may
freely admit that a simple or immutable God must be timeless, we
have even less reason to think God simple or immutable than to
think him timeless and so can hardly hold that he is timeless on the
basis of those doctrines.

Argument from Divine Knowledge of Future Contingents

In contrast to divine simplicity and immutability, divine

omniscience is clearly a great-making property and enjoys
considerable Scriptural warrant. An argument for divine
timelessness predicated upon God’s omniscience would therefore
have a more secure theological foundation. Many thinkers have
argued that God’s knowledge of future contingent events, for
example, future human free actions, implies divine timelessness. The
reasoning seems to go as follows:

(5) A temporal being cannot know future contingent events.

(6) God knows future contingent events.

(7) Therefore, God is not a temporal being.

Again, if God is not a temporal being, then (4) follows as before.

 Despite the denial of (6) on the part of a wide range of
contemporary thinkers from process theologians to ‘open’ theists, a
biblical doctrine of divine omniscience makes (6) incumbent upon
an orthodox theologian.5 The argument hinges, therefore, on the
truth of (5). On behalf of (5) it is usually claimed that contingent
events, not being deducible from present causes, can be known only
insofar as they are real or existent. Given (6), it follows that future
contingent events are real or existent for God. Defenders of divine
timelessness such as Boethius, Anselm, and Aquinas thus typically
maintained that all events in time are real to God and therefore can
be known by him via his scientia visionis (knowledge of vision).

How can we make sense of this claim? The most plausible move
for the defender of divine timelessness to make will be to hold that
the four-dimensional spacetime manifold exists tenselessly and that
God transcends that manifold. A good many physicists and
philosophers of time and space embrace such a tenseless view of
time (spacetime realism). Such a view makes sense of the traditional
claim that all events in time are present to God and therefore known
to him via his scientia visionis.

The drawback is that there is a high price to be paid

philosophically and theologically for such a tenseless theory of
time.6 Therefore, the claim that contingent events can be known
only insofar as they are real or existent comes at a considerable cost
for the biblical theist. One is therefore inclined to be sceptical of the
argument on behalf of (5).
 Moreover, (5) can be directly challenged as well. In assessing the
question of how God knows truths about temporal events, we may
distinguish two models of divine cognition: the perceptualist model
and the conceptualist model. The perceptualist model construes
divine knowledge on the analogy of sense perception: God looks and
sees what is there. Such a model patently underlies the classic
doctrine of scientia visionis and is implicitly assumed when people
speak of God’s ‘foreseeing’ the future. The perceptualist model of
divine cognition does encounter di culty concerning God’s
knowledge of future contingents, for, if future events do not exist,
there is nothing there to perceive.7

By contrast, on a conceptualist model of divine knowledge, God

does not acquire his knowledge of the world by anything like
perception. His knowledge of the future is not based on his ‘looking’
ahead and ‘seeing’ what lies in the future (a terribly
anthropomorphic notion in any case). Rather God’s knowledge is
more like a mind’s knowledge of innate ideas. It is therefore
inappropriate to speak of God’s acquiring knowledge at all. Rather as
an omniscient being, God has essentially the property of knowing all
truths; there are truths about future events; ergo, God knows all
truths concerning future events. So long as we are not seduced into
thinking of divine foreknowledge on the model of perception, it is
no longer evident why knowledge of future contingents should be
impossible.
 We can go further, however. For the doctrine of middle
knowledge (scientia media) is a version of the conceptualist model
that allows us to say considerably more about the basis of God’s
foreknowledge of future contingents. Divine foreknowledge is based
on (1) God’s middle knowledge of what every creature would freely
do under any circumstances and (2) his knowledge of the divine
decree to create certain sets of circumstances and to place certain
creatures in them. Given middle knowledge and the divine decree,
foreknowledge follows automatically as a result without any
perception of the created world. This complex and interesting
doctrine must be pursued in an independent discussion (see Ch. 12,
Divine Providence).

In sum, while the argument from God’s knowledge of future

contingents has some force in motivating a doctrine of divine
timelessness, that force is mitigated by the availability of viable
alternatives and the high price exacted by a tenseless theory of time.

Argument from Special Relativity

A third argument for divine timelessness arises from the concept of

time in Einstein’s Special Theory of Relativity (STR). According to
this theory there is no unique, universal time and so no unique,
worldwide ‘now’. Each inertial frame has its own time and its own
present moment, and there is no overarching, absolute time in
which all these diverse times are integrated into one. So if God is in
time, then the obvious question raised by STR is: Whose time is he
in?

The defender of divine timelessness maintains that there is no

acceptable answer to this question. We cannot plausibly pick out
some inertial frame and identify its time as God’s time because God
is not a physical object in uniform motion, and so the choice of any
such frame would be wholly arbitrary. Moreover, it is di cult to see
how God, con ned to the time of one inertial frame, could be
causally sustaining events that are real relative to other inertial
frames but are future or past relative to God’s frame. Similarly,
God’s knowledge of what is happening now would be restricted to
the temporal perspective of his frame, leaving him ignorant of what
is actually going on in other frames. In any case, if God were to be
associated with a particular inertial frame, then surely, as God’s
time, the time of that frame would be privileged. It would be the
equivalent of the privileged frame of the aether in classical physics.
So long as we maintain, with Einstein, that no frame is privileged,
then we cannot identify the time of any inertial frame as God’s time.

Neither can we say that God exists in the ‘now’ associated with
the time of every inertial frame, for this would obliterate the unity
of God’s consciousness. In order to preserve God’s personal
consciousness, it must not be fragmented and scattered among the
inertial frames in the universe. But if God’s time cannot be
identi ed with the time of a single frame or of a plurality of frames,
then God must not be in time at all; that is to say, he exists
timelessly.

We can summarize this reasoning as follows:

(8) STR is correct in its description of time.

(9) If STR is correct in its description of time, then if God is

temporal, he exists in either the time associated with a
single inertial frame or the times associated with a plurality
of inertial frames.

(10) Therefore, if God is temporal, he exists in either the time

associated with a single inertial frame or the times
associated with a plurality of inertial frames.

(11) God does not exist in either the time associated with a
single inertial frame or the times associated with a plurality
of inertial frames.

(12) Therefore, God is not temporal.

What can be said in response to this argument? Although it may

come as something of a shock to many, the most dubious premise of
the argument is (8). For STR’s concept of time rests upon decrepit
epistemological foundations. Einstein’s rede nition of simultaneity
in terms of clock synchronization by light signals simply assumes
that the time light takes to travel between two relatively stationary
observers A and B is the same from A to B as from B to A in a
roundtrip journey. That assumption presupposes that A and B, while
at relative rest, are not both in absolute motion, or in other words
that neither absolute space nor a privileged inertial frame exists.
What justi cation did Einstein have for so radical a presupposition?
The answer, in a word, is veri cationism. It is empirically
impossible to distinguish uniform motion from rest relative to such a
frame, and Einstein believed that if absolute space and absolute
motion or rest are undetectable empirically, they therefore do not
exist (and may even be said to be meaningless). Historians of
science have shown that at the philosophical roots of Einstein’s
theory lies a veri cationist epistemology, mediated to the young
physicist chie y through the in uence of Ernst Mach, which comes
to expression in Einstein’s analysis of the concepts of time and
space.8

The untenability of veri cationism is so universally

acknowledged that it will not be necessary to rehearse the
objections against it here.9 Veri cationism provides no justi cation
for thinking that Newton erred, for example, in holding that
absolute time, grounded in God’s sempiternal duration, exists
independently of our physical measures of it and may or may not be
accurately registered by them. With the demise of veri cationism,
the philosophical underpinnings of STR have collapsed. In short,
there is no reason to think that (8) is true.

But what about (9)? The di culty with this premise is that it
fails to take into account the fact that STR is a restricted theory of
relativity and therefore is correct only within prescribed limits. It is
a theory that deals with uniform motion only. The analysis of non-
uniform motion, such as acceleration and rotation, is provided by
the General Theory of Relativity (GTR). STR cannot therefore be
expected to give us the nal word about the nature of time and
space; indeed, within the context of GTR a new and important
conception of time emerges.

GTR serves to introduce into Relativity Theory a cosmic

perspective, enabling us to draft cosmological models of the
universe governed by the gravitational eld equations of GTR.
Within the context of such cosmological models, the issue of time
resurfaces dramatically. All contemporary cosmological models
derive from Russian physicist Alexander Friedman’s 1922 model of
an expanding, material universe characterized by ideal homogeneity
and isotropy. Although GTR does not itself mandate any formula for
how to slice up spacetime into a temporally ordered foliation,
nevertheless certain models of spacetime, such as the Friedman
model, have a dynamical, evolving spatial geometry whose natural
symmetries guide the construction of a cosmic time; in order to
ensure a smooth development of this geometry, it will be necessary
to construct a time parameter based on a preferred slicing of
spacetime. Now as a parameter independent of spatial coordinates,
cosmic time measures the duration of the universe as a whole in an
observer-independent way; that is to say, the lapse of cosmic time is
the same for all observers.

Based on a cosmological, rather than a local, perspective, cosmic

time serves to restore the classical notions of universal time and
absolute simultaneity that STR denied. The defender of divine
temporality may accordingly hold that God exists in cosmic time.10
Hence, contrary to (9), it does not follow from the correctness of
STR that if God is in time, then he is in the time of one or more
inertial frames. Because space itself is expanding, there is no
universal inertial frame with which God can be associated, even
though there does exist a preferred foliation of space time and so a
cosmic time in which God can be conceived to exist.11

Argument from the Incompleteness of Temporal Life

An important argument in favor of divine timelessness is based on

the incompleteness of temporal life. Brian Leftow, as well as
Eleonore Stump and Norman Kretzmann, argues that the eeting
nature of temporal life is incompatible with the life of a most perfect
being such as God. A temporal being is unable to enjoy what is past
or future for it, possessing only the eeting present. The passage of
time thus renders it impossible for any temporal being, even God, to
possess all its life at once. By contrast a timeless God lives all his life
at once because he literally has no past or future and so su ers no
loss. Therefore, since God is the most perfect being, he is timeless.

We can formulate this argument as follows:

(13) God is the most perfect being.

(14) The most perfect being has the most perfect mode of
existence.
 (15) Therefore, God has the most perfect mode of existence.

(16) Temporal existence is a less perfect mode of existence than

timeless existence.

(17) Therefore, God has a timeless mode of existence.

The key premise here is (16), which rests on very powerful

intuitions about the irretrievable loss that arises through the
experience of temporal passage, a loss that intuitively should not
characterize the experience of a most perfect being. Some
philosophers of time might try to avert the force of this
consideration by adopting a tenseless view of time according to
which things and events do not in fact come to be or pass away. The
di erence between past, present, and future is a subjective illusion
of consciousness. On this view of time, no temporal being ever
really loses its past or has not yet acquired its future; it (or its
temporal parts) just exists tenselessly at its various temporal
locations. A temporal God would exist at all temporal locations
without beginning or end and so would not lose or acquire portions
of his life.

The problem with this escape route is that it fails to appreciate

that the argument is based on the experience of temporal passage,
rather than on the objective reality of temporal passage itself. Even
if the future never becomes and the past is never really lost, the fact
remains that for a temporal person the past is lost to him and the
future is not accessible to him. For this reason, it would be futile to
attempt to elude the force of this argument by postulating a
temporal deity in a tenseless time.

Perhaps, however, the realization that the argument is

essentially experiential in character opens the door for a temporalist
alternative. When we recall that God is perfectly omniscient and so
forgets nothing of the past and knows everything about the future,
then time’s passage is not so tragic for him. His past experiences do
not fade as ours do, and he has perfect prescience of what the future
holds. So it is far from obvious that the experience of temporal
passage is so melancholy an a air for an omniscient God as it is for
us. Moreover, the life of a perfect person may have to be
characterized by the incompleteness that would in other contexts be
considered an imperfection. There is some evidence that
consciousness of time’s ow can actually be an enriching
experience, as in music appreciation.12 Timelessness may not be the
most perfect mode of existence of a perfect person. All this goes to
call into question (16). Still, this last argument, like the argument
from divine foreknowledge, does have some force and so needs to be
weighed against whatever arguments can be o ered on behalf of
divine temporality.

ARGUMENTS FOR DIYINE TEMPORALITY

Argument from the Impossibility of Atemporal Personhood

 What arguments, then, might be o ered for divine temporality?
One argument frequently raised in the literature is that timelessness
and personhood are incompatible. Some philosophers have denied
that a timeless God can be a self-conscious, rational being because
he could not exhibit certain forms of consciousness that we normally
associate with personal beings (namely, ourselves). For example,
Robert Coburn has written,

Surely it is a necessary condition of anything’s being a person that it

should be capable (logically) of, among other things, doing at least
some of the following: remembering, anticipating, re ecting,
deliberating, deciding, intending, and acting intentionally. To see
that this is so one need but ask oneself whether anything which
necessarily lacked all of the capacities noted would, under any
conceivable circumstances, count as a person. But now an eternal
being would necessarily lack all of these capacities in as much as
their exercise by a being clearly requires that the being exist in
time…. Hence, no eternal being, it would seem, could be a person.13

Since God is essentially personal, he therefore cannot be timeless.

We can formulate this argument as follows (using x, y, z to

represent certain properties allegedly essential to personhood):

(18) Necessarily, if God is timeless, he does not have the

properties x, y, z.
 (19) Necessarily, if God does not have the properties x, y, z, then
God is not personal.

(20) Necessarily, God is personal.

(21) Therefore, necessarily, God is not timeless.

The defender of divine timelessness may attempt to turn back

this argument either by challenging the claim that the properties in
question are necessary conditions of personhood or by showing that
a timeless God could possess the relevant properties after all. With
respect to the second strategy, even if Coburn were correct that a
personal being must be capable of exhibiting the forms of
consciousness he lists, it does not follow that a timeless God cannot
be personal. For God could be capable of exhibiting such forms of
consciousness but be timeless just in case he does not in fact exhibit
any of them. In other words, the hidden assumption behind
Coburn’s reasoning is that God’s being timeless or temporal is an
essential property of God. But that assumption seems dubious.
Suppose, for the sake of argument, that God is in fact temporal. Is it
logically impossible that God could have been timeless instead?
Since God’s decision to create is free, we can conceive of a possible
world in which God alone exists. If he is unchanging in such a
world, then on any relational view of time God would be timeless.
In such an atemporal world God would lack certain properties
which we have supposed him to have in the actual world—for
example, the property of knowing what time it is or the property of
coexisting with temporal creatures—and he would have other
properties that he lacks in the actual world—for example, the
property of being alone or of knowing that he is alone—but none of
these di erences seems signi cant enough to deny that God could
be timeless yet still be the same being. But then it seems that there
are possible worlds in which God exists temporally and possible
worlds in which he exists timelessly. God’s temporal status is thus
plausibly a contingent rather than essential property. So (apart from
highly controversial claims on behalf of divine simplicity or
immutability) there seems to be no reason to think that God is
either essentially temporal or essentially timeless.

So if timelessness is a merely contingent property of God, he

could be entirely capable of remembering, anticipating, re ecting,
and so on; only were he to do so, then he would not be timeless. So
long as he freely refrains from such activities he is timeless, even
though he has the capacity to engage in those activities. Thus, by
Coburn’s own lights God must be regarded as personal.

At a more fundamental level, it is in any case pretty widely

recognized that most of the forms of consciousness mentioned by
Coburn are not essential to personhood—indeed, not even the
capacity for them is essential to personhood. Take remembering, for
example. Any temporal individual who lacked memory would be
mentally ill or less than human. But if an individual exists
timelessly, then he has no past to remember. Similarly with regard
to anticipation: since a timeless individual has no future, there just
is nothing to anticipate. Nevertheless, given his omniscience, God
would still know what takes place (tenselessly) at every time.

As for re ecting and deliberating, these are ruled out not so

much by God’s time-lessness as by his omniscience. An omniscient
being cannot re ect and deliberate because he already knows the
conclusions to be derived. Even if God is temporal, he does not
engage in re ection and deliberation. But he is surely not
impersonal as a result.

What about deciding, intending, and acting intentionally? All

these forms of consciousness are exhibited by a timeless God. With
respect to deciding, again omniscience alone precludes God’s
deciding in the sense of making up his mind after a period of
indecision. Even a temporal God does not decide in that sense. But
God does decide in the sense that his will inclines toward one
alternative rather than another and does so freely. It is up to God
what he does; he could have willed otherwise. This is the strongest
sense of libertarian freedom of the will. In God’s case, because he is
omniscient, his free decisions are either sempiternal or timeless
rather than preceded by a period of ignorance and indecision.

As for intending or acting intentionally, there is no reason to

think that intentions are necessarily future-directed. One can direct
one’s intentions at one’s present state. God, as the Good, can
timelessly desire and will his own in nite goodness. Such a
changeless intention can be as timeless as God’s knowing his own
essence. Moreover, in the empty world we have envisioned, God
may timelessly will and intend to refrain from creating a universe.
God’s willing to refrain from creation should not be confused with
the mere absence of the intention to create. A stone is characterized
by the absence of any intention to create but cannot be said to will
to refrain from creating. In a world in which God freely refrains
from creation, his abstaining from creating is a result of a free act of
the will on his part. Hence, it seems that God can timelessly intend,
will, and choose what he does.

In short, the argument for divine temporality based on God’s

personhood cannot be deemed a success. On the contrary, a timeless
God can be plausibly said to be a self-conscious, rational individual
endowed with freedom of the will and therefore a person.

Argument from Divine Action in the World

In our thought experiment above, we abstracted from the actual

existence of the temporal world and considered God existing alone
without creation and asked whether he could exist timelessly. But,
of course, the temporal world does exist. The question therefore
arises whether God can stand in relation to a temporal world and
yet remain timeless. It is very di cult to see how he can. Imagine
once more God existing changelessly alone without creation, but
with a changeless determination of his will to create a temporal
world with a beginning. Since God is omnipotent, his will is done,
and a temporal world comes into existence. Can God remain
untouched by the world’s temporality? It seems not. For at the rst
moment of time, God stands in a new relation, one in which he did
not stand before (indeed, there was no ‘before’). Even if in creating
the world God undergoes no intrinsic change, he at least undergoes
an extrinsic change. For at the moment of creation, God comes into
the relation of sustaining the universe or, at the very least, of
coexisting with the universe, relations in which he did not stand
before. Since he is free to refrain from creation, God could have
never stood in those relations, had he so willed. But in virtue of his
creating a temporal world, God comes into a relation with that
world the moment it springs into being. Thus, even if it is not the
case that God is temporal prior to his creation of the world, he
nonetheless undergoes an extrinsic change at the moment of
creation that draws him into time in virtue of his real relation to the
world. So even if God is timeless without creation, his free decision
to create a temporal world also constitutes a free decision on his
part to exist temporally.

The argument can be summarized as follows:

(22) God is creatively active in the temporal world.

(23) If God is creatively active in the temporal world, God is

really related to the temporal world.

(24) If God is really related to the temporal world, God is

temporal.

(25) Therefore, God is temporal.

This argument, if successful, does not prove that God is essentially
temporal, but that if he is a creator of a temporal world—as he in
fact is—then he is temporal.

One way to escape this argument is to deny (23). This might not
appear to be a very promising strategy, since it seems obvious that
God is related to his creatures insofar as he sustains them, knows
them, and loves them. Remarkably, however, it was precisely this
premise that medieval theologians such as Aquinas denied. Thomas
agrees with (24). On his view, relational properties involving God
and creatures, like God’s being Lord, rst begin to exist at the
moment at which the creatures come into being (ST Ia q. 13 a. 7).
Hence, if God stands in real relations to his creatures, he acquires
those relational properties de novo at the moment of creation and
thus undergoes change. And anything that changes, even
extrinsically, must be in time. Thomas escapes the conclusion that
God is therefore temporal by denying that God stands in any real
relation to the world. Since God is absolutely simple, he stands in no
relations to anything, for relations would introduce complexity into
God’s being. Aquinas holds, paradoxically, that while creatures are
really related to God, God is not really related to creatures. The
relation of God to creatures exists only in our minds, not in reality.
On Aquinas’s view, then, God undergoes no extrinsic change in
creating the world. He just exists, and creation is creatures’ coming
into existence with a real relation to God of being caused by God.
 This is certainly an extraordinary doctrine. Wholly apart from its
reliance on divine simplicity, the doctrine of no real relations is very
problematic. God’s sustaining the world is a causal relation rooted
in the active power and intrinsic properties of God as First Cause.
Thus, to say the world is really related to God by the relation is
sustained by, but that God is not really related to the world by the
relation is sustaining, seems unintelligible. It is to say that one can
have real e ects without a real cause—which seems self-
contradictory or incomprehensible.

Moreover, God is surely really related to his creatures in the

following sense: in di erent possible worlds, God’s will, knowledge,
and love are di erent than they actually are. For example, if God
had not chosen to create a universe at all, he would surely have a
di erent will than that which he has (for he would not will to create
the universe); he would know di erent truths than the ones he
knows (for example, he would not know The universe exists); he
would not love the same creatures he actually loves (since no
creatures would exist). It is the implication of Aquinas’s view,
however, that God is perfectly similar in every possible world: he
never wills di erently, he never acts di erently, he never knows
di erently, he never loves di erently. Whether the world is empty
or chock-full of creatures of every sort, there is no di erence in God.
But then it becomes unintelligible why this universe or any universe
exists rather than just nothing. The reason cannot lie in God, for he
is perfectly similar in all possible worlds. Nor can the reason lie in
creatures, for we are asking for some explanation of their existence.
Thus, on Aquinas’s view there just is no reason for why this universe
or any universe at all exists. Therefore, Aquinas’s attempt to evade
the present argument by denying (23) is implausible.

Recent defenders of timeless eternity have turned their guns on

(24) instead. They have tried to craft theories of divine eternity that
would permit God to be really related to the temporal world and yet
to exist timelessly.

For example, Eleonore Stump and the late Norman Kretzmann

attempted to craft a new simultaneity relation, which they believed
would allow a timeless God to relate to his creation.14 They propose
to treat modes of existence as analogous to reference frames in STR
and to construct a de nition of ET-simultaneity in terms of two
reference frames (timelessness and temporality) and two observers
(one in eternity and one in time). Their basic idea is: take some
eternal being x and some temporal being y. These two are ET-
simultaneous just in case relative to some hypothetical observer in
the eternal reference frame x is eternally present and y is observed
as temporally present, and relative to some hypothetical observer in
any temporal reference frame y is temporally present and x is
observed as eternally present.15

On the basis of their de nition of ET-simultaneity, Stump and

Kretzmann believe themselves to have solved the problem of how a
timeless being can be really related to a temporal world. For relative
to the eternal reference frame, any temporal entity that exists at any
time is observed to be present, and relative to any moment of time
God is observed to be present. The metaphysical relativity
postulated by ET-simultaneity implies that all events are present to
God in eternity and therefore open to his timeless causal in uence.
Every action of God is ET-simultaneous with its temporal e ect.

Unfortunately, as many critics have pointed out, the language of

observation employed in the de nition is wholly obscure.16 In STR
very speci c physical content is given to the notion of observation
through Einstein’s operational de nitions of distant simultaneity.
But in the de nition of ET-simultaneity, no hint is given as to what
is meant, for example, by x’s being observed as eternally present
relative to some moment of time. In the absence of any procedure
for determining ET-simultaneity, the de nition reduces to the
assertion that relative to the reference frame of eternity x is
eternally present and y is temporally present and that relative to
some temporal reference frame y is temporally present and x is
eternally present—which is only a restatement of the problem!
Worse, if y is temporally present to God, then God and y are not ET-
simultaneous at all, but temporally simultaneous. Thus, God would
be temporally simultaneous with every temporal event, which is to
sacri ce divine timelessness.

To their credit, Stump and Kretzmann later revised their

de nition of ET-simultaneity so as to free it from observation
language.17 Basically, their new account tries to de ne ET-
simultaneity in terms of causal relations. On the new de nition, x
and y are ET-simultaneous just in case relative to an observer in the
eternal reference frame, x is eternally present and y is temporally
present, and the observer can enter into direct causal relations with
both xand y; and relative to an observer in any temporal reference
frame, x is eternally present and y is at the same time as the
observer, and the observer can enter into direct causal relations
with both x and y.

The fundamental problem with this new account of ET-

simultaneity is that it is viciously circular. For ET-simultaneity was
originally invoked to explain how a timeless God could be causally
active in time; but now ET-simultaneity is de ned in terms of a
timeless being’s ability to be causally active in time. Our original
problem was to explain how God could be both timeless and yet
creatively active in the world. That is hardly explained by saying
that a timeless God is ET-simultaneous with his e ects in time and
then de ning ET-simultaneity in terms of the ability of a timeless
being to be causally related to temporal e ects. This amounts to
saying that God can be causally active in time because he can be
causally active in time. Since their rst de nition was explanatorily
vacuous and their second de nition viciously circular, Stump and
Kretzmann must be judged to have failed in their attempt to
undercut (24).

Leftow has o ered a di erent account of divine eternity in order

to refute (24).18 On the Stump-Kretzmann model, there is no
common reference frame or mode of existence shared by timeless
and temporal beings. As a result, Stump and Kretzmann were unable
to explain how such beings could be causally related. The essence of
Leftow’s proposal is to remedy this defect by maintaining that
temporal beings do exist in eternity; they share God’s mode of
existence and so can be causally related to God. But, he insists, this
does not imply that time or temporal existence is illusory, for
temporal beings also have a temporal mode of existence.

How can it be shown that temporal beings exist in timeless

eternity? Leftow’s argument is based on three theses:

(I) The distance between God and everything in space is zero.

(II) Spatial things do not change in any way unless there is a

change of place (a motion involving a material thing).

(III) If something is in time, it is also in space.

On the basis of these theses Leftow argues as follows: there can be

no change of place relative to God because the distance between
God and everything in space is zero. But if there is no change of
place relative to God, there can be no change of any sort on the part
of spatial things relative to God. Moreover, since anything that is
temporal is also spatial, it follows that there are no temporal, non-
spatial beings. The only temporal beings there are exist in space,
and none of these changes relative to God. Assuming, then, some
relational view of time, it follows that all temporal beings exist
timelessly relative to God. Thus, relative to God all things are
timelessly present and so can be causally related to God.

The problem with this reasoning is that all three of its

foundational theses seem false, some obviously so. (I), for example,
rests pretty obviously on a category mistake. When we say that
there is no distance between God and creatures, we do not mean
that there is a distance and its measure is zero. Rather we mean that
the category of distance does not even apply to the relations
between a non-spatial being such as God and things in space. What
about (II)? This thesis is false if time is ‘tensed’. For then spatial
things can change even if there is no spatial motion by changing in
their temporal properties. For example, some spatial object can
change by being one year old and then becoming two years old,
even if no change of place has occurred. Even most relationalists are
today willing to admit that time can go on during periods of spatial
changelessness. So even if the entire universe were frozen into
immobility, there would still be change relative to God, namely,
change of temporal properties. Thus, if time is tensed—and Leftow
allows that may be—then his theory is nulli ed. Finally, consider
(III). Leftow needs this thesis, lest someone say that there are non-
spatial, temporal beings such as angels that are changing relative to
God. Such beings would (on Leftow’s analysis) have a zero distance
from God and yet not be changeless relative to God. Thus, they
would not exist in eternity. So in order to sustain his claim that
temporal beings exist in eternity, Leftow has to get rid of such
beings. But we have every reason to reject this radical thesis. Even
in the absence of a physical universe, God could choose to entertain
a succession of thoughts or to create an angelic being that
experiences a stream of consciousness, and such a series of mental
events alone is su cient for such entities’ being in time. Thus, all
Leftow’s key theses are at least dubious, if not clearly false. We have
little choice but to conclude that he has given no good grounds for
thinking that temporal beings exist in timeless eternity.

In summary, it seems that we have here a powerful argument for

divine temporality. Classical attempts such as Aquinas’s to deny that
God is really related to the world and contemporary attempts such
as those of Stump, Kretzmann, and Leftow to deny that God’s real
relation to the world involves him in time all appear in the end to
be less plausible than the premises of the argument itself.

Argument from Divine Knowledge of Tensed Facts

We have seen that God’s action in the temporal world gives us good
grounds for concluding God to be temporal in view of the extrinsic
change he undergoes through his changing relations with the world.
But the existence of a temporal world also seems to entail intrinsic
change in God in view of his knowledge of what is happening in the
temporal world. For since what is happening in the world is in
constant ux, so also must God’s knowledge be in constant ux.
Defenders of divine temporality have argued that a timeless God
cannot know certain tensed facts about the world—for example,
what is happening now—and therefore, since God is omniscient, he
must be temporal.

We can formulate the argument as follows:

(26) A temporal world exists.

(27) God is omniscient.

(28) If a temporal world exists, then if God is omniscient, God

knows tensed facts.

(29) If God is timeless, he does not know tensed facts.

(30) Therefore, God is not timeless.

Again, this argument does not prove that God is essentially

temporal, but, if successful, it does show that if a temporal world
exists, then God is temporal.

Defenders of divine timelessness have attempted to refute this

argument either by arguing that a timeless God can know tensed
facts or by arguing that God may still qualify as omniscient even if
he is ignorant of tensed facts.

Let us look rst at the plausibility of denying (29). Can a

timeless God know tensed facts? Although Jonathan Kvanvig,
Edward Wierenga, and Leftow have all argued that God can know
the facts expressed by tensed sentences, an analysis of their
respective positions reveals that in the end they all embrace the
view that the factual content expressed by tensed sentences is
tenseless.19 Despite rst appearances to the contrary, they all accept
the truth of (29). Kvanvig, Wierenga, and Leftow’s accounts are the
most sophisticated attempts to explain how a timeless God can
know the facts expressed by tensed sentences, yet they all nally
deny that God knows tensed facts. Thus, (29) seems secure.

Defenders of divine timelessness have no recourse, then, but to

deny (28). They must deny that omniscience entails a knowledge of
tensed facts. They can do this either by revising the traditional
de nition of omniscience or else by maintaining that tense, while an
objective feature of time, does not strictly belong to the factual
content expressed by tensed sentences. Let us examine each strategy
in turn.

The general problem with the strategy of revising the traditional

de nition of omniscience is that any adequate de nition of a
concept must be in line with our intuitive understanding of that
concept. We are not free simply to ‘cook’ the de nition just to solve
some problem under discussion. According to the traditional
de nition, a person S is omniscient if and only if, for every fact, S
knows that fact and does not believe its contradictory. On such a
de nition, if there are tensed facts, an omniscient person must know
them. What plausible alternative de nition of omniscience might
the defender of divine timelessness o er?

Wierenga o ers a revised account of omniscience that would not

require an omniscient person to know tensed facts. Some facts, he
says, are facts only from a particular perspective. They must be
known to an omniscient being only if he shares that particular
perspective. Thus, a person is omniscient if and only if, for every
fact and every perspective, if something is a fact from a certain
perspective, then that person must know that it is a fact from that
perspective, and if he shares that perspective, then he must know
the fact in question. Wierenga treats moments of time as
perspectives relative to which tensed facts exist. So while a temporal
person existing on 8 December 1941 must (if he is omniscient) know
the fact Yesterday the Japanese attacked Pearl Harbor, a timeless
person must know only that from the perspective of 8 December
1941, it is a fact that Yesterday the Japanese attacked Pearl Harbor.
On this de nition God’s being omniscient does not require that he
know the tensed fact, but only the tenseless fact that from a certain
perspective a certain tensed fact exists.

Wierenga’s revised de nition of omniscience seems to be

unacceptably cooked. Wierenga is not denying that there are tensed
facts. Rather he wants to allow that there really are tensed facts but
to maintain that an omniscient being need not know them. This
claim seems quite implausible. On Wierenga’s view temporal
persons know an incalculable multitude of facts about the world of
which a supposedly omniscient being is ignorant. Temporal persons
know that the Japanese attack on Pearl Harbor is over; God has no
idea whether it has occurred or not. Since he does not know what
time it actually is, he does not know any tensed facts. This is an
unacceptably limited eld of knowledge to qualify as omniscience.
 Leftow also entertains the idea of revising the de nition of
omniscience in such a way that omniscience does not entail
knowledge of all truths. He argues, in e ect, that there are many
sorts of truths that God cannot know, so there is no harm in
admitting one more class of truths (namely, tensed truths) of which
God is ignorant. But, again, such a consideration should not a ect
the de nition of ‘omniscience’ as such. Besides, should it turn out
that there are truths God cannot know, that is no reason for further
eroding the extent of his knowledge by denying him knowledge of
tensed truths. In any case, does Leftow succeed in showing that
there are truths which God cannot know? It seems not. His examples
of things God cannot know include how it feels to be oneself a
failure or a sinner. But Leftow has confused knowing how with
knowing that. Knowing how does not take truths as its object. God can
know such truths as Being a failure feels lousy, Sinners feel guilty and
hopeless, and so on. God’s not knowing how it feels to be himself a
failure or a sinner is not an example of truths he fails to know and
so does not constitute a restriction on his omniscience. Leftow
furnishes no example of any truth which might be conjoined with
‘knows that’ such that we cannot say, ‘God knows that—,’ where the
blank is lled by the truth in question. Therefore, he has not
adequately motivated denying that knowledge of tensed truths
properly belongs to omniscience.

It seems, therefore, that no adequate grounds have been given

for thinking that someone could be omniscient and yet not know
tensed truths. The traditional de nition of omniscience requires it,
and we have no grounds that do not involve special pleading for
revising the usual de nition.

So what about the second strategy for denying (28), namely,

maintaining that tense does not, strictly speaking, belong to the
factual content expressed by tensed sentences, even though tense is
an objective feature of the world? Tense might be analyzed as a
feature of the mode in which the factual content is presented to
someone expressing it, or of the way in which a person grasps the
factual content, or of the context of someone’s believing the factual
content. Alternatively, tense could be understood in terms of a
person’s ascribing to himself in a present-tense way the property of
being such as the factual content expressed by the sentence
speci es. On such analyses, an omniscient being could be timeless
because omniscience is traditionally de ned in terms of factual
knowledge and tense is not part of the factual content of tensed
sentences. Tense is an objective feature of the world, but since it
does not belong to the factual content of a sentence, a being which
knew only tenseless facts could on the traditional de nition count as
omniscient.

Even though such analyses are plausible and attractive, they do

not ultimately save the day for the defender of divine timelessness.
For as the greatest conceivable being, God is not merely factually
omniscient, but also maximally excellent cognitively. On the
theories under discussion a merely factually omniscient God would
know such things as God is omnipotent, God loves his creatures, God
created the universe, and so on. But he would not have to possess any
rst-person indexical beliefs such as ‘I am omnipotent’, ‘I love my
creatures’, ‘I created the universe’, and so forth. A machine could
count as omniscient under such analyses. But such a God or machine
would clearly not have maximum cognitive excellence. In order to
qualify as maximally excellent cognitively, God would have to
entertain all and only the appropriate, true rst-person beliefs about
himself. This would furnish him with knowledge de se ( rst-person
self-knowledge) in addition to mere knowledge de re (knowledge of
a thing from a third-person perspective). In order to be maximally
excellent cognitively, God would not have to possess all knowledge
de se in the world, but only such knowledge de se as is appropriate
to himself. It would be a cognitive defect, not a perfection, for God
to have the belief ‘I am Napoleon’, though for Napoleon such a
belief would be a perfection. The point is that omniscience (on these
theories) is not enough for perfect-being theology; God must be
maximally excellent cognitively.

Now in the same way, it is a cognitive perfection to know what

time it is, what is actually happening in the universe. A being whose
knowledge is composed exclusively of tenseless facts is less excellent
cognitively than a being who also knows what has occurred, what is
occurring, and what will occur in the world. This latter person
knows in nitely more than the former and is involved in no
cognitive defect in so knowing. On the analogy of knowledge de se,
we can refer to such knowledge as knowledge de praesenti
(knowledge of the present). A being that lacks such knowledge is
more ignorant and less excellent cognitively than a being that
possesses it. Therefore, if we adopt views according to which tense
is extraneous to the factual content expressed by a tensed sentence,
we should simply revise premise (28) to read

(28′) If a temporal world exists, then if God is maximally

excellent cognitively, then God has knowledge de praesenti

and, with appropriate revisions, the argument goes through as

before.

The attempt to deny (28) thus seems to fare no better than the
e ort to refute (29). If God is omniscient, then given the existence
of a temporal world, he cannot be ignorant of tensed facts. It follows
that God is not timeless, which is to say, he is temporal. So in
addition to the argument from divine action in the world, we now
have a second powerful argument based on God’s changing
knowledge of tensed facts for thinking that God is in time.

ETERNITY AND THE NATURE OF TIME

On the basis of our foregoing discussion, we have seen

comparatively weak grounds for a rming divine timelessness but
two powerful arguments in favor of divine temporality. It would
seem, then, that we should conclude that God is temporal. But such
a conclusion would be premature. For there does remain one way of
escape still open for defenders of divine timelessness. The argument
based on God’s action in the world assumed the objective reality of
temporal becoming, and the argument based on God’s knowledge of
the temporal world assumed the objective reality of tensed facts. If
one denies the objective reality of temporal becoming and tensed
facts, then the arguments are undercut. For in that case, nothing to
which God is related ever comes into or passes out of being, and all
facts exist tenselessly, so that God undergoes neither extrinsic nor
intrinsic change. He can be the immutable, omniscient Sustainer and
Knower of all things and, hence, exist timelessly.

In short, the defender of divine timelessness can escape the

arguments for divine temporality by embracing the tenseless theory
of time. It is noteworthy, however, that almost no defender of divine
timelessness has taken this route. Virtually the only proponent of
timeless eternity to embrace consciously the tenseless theory of time
in defending God’s timelessness is Paul Helm.20

It seems, then, that in order to adjudicate the question of the

nature of divine eternity and God’s relationship to time,
philosophical theologians have no choice but to grapple with a
further question, one of the most profound and controverted issues
of metaphysics: is time tensed or tenseless? This is di cult and
mysterious territory. But we have no choice: if we are to understand
eternity, we must rst understand time.

NOTES
1. For an analysis of what it means to be permanent, see Brian
Leftow, Time and Eternity, Cornell Studies in the Philosophy of
Religion (Ithaca: Cornell University Press, 1991), 133; cf.
Quentin Smith, ‘A New Typology of Temporal and Atemporal
Permanence’, Noûs 23 (1989), 307–30. According to Leftow an
entity is permanent if and only if it exists and has no rst or
last nite period of existence, and there are no moments
before or after it exists.

2. James Barr, Biblical Words for Time (London: SCM, 1962), 149.

3. An intrinsic change is a non-relational change, involving only

the subject. For example, an apple changes from green to red.
An extrinsic change is a relational change, involving
something else in relation to which the subject changes. For
example, a man becomes shorter than his son, not by
undergoing an intrinsic change in his own height, but by being
related to his son as his son undergoes intrinsic change in his
height.

4. See discussion in Thomas V. Morris, Anselmian Explorations

(Notre Dame: University of Notre Dame Press, 1987), 98–123;
Christopher Hughes, On a Complex Theory of a Simple God,
Cornell Studies in Philosophy of Religion (Ithaca: Cornell
University Press, 1989).

5. I take for granted that there are contingent events such as

human free acts; see Ch. 12 in this volume on divine
providence. See also my The Only Wise God: The Compatibility
 of Divine Foreknowledge and Human Freedom (Grand Rapids:
Baker Bookhouse, 1987).

6. See my The Tenseless Theory of Time: A Critical Examination,

Synthese Library 294 (Dordrecht: Kluwer Academic, 2000).

7. Notice, however, that if we think of statements or facts as

within God’s perceptual purview, then even on a perceptualist
model, God must know the future, so long as the Principle of
Bivalence holds for future-tense statements. For he perceives
which future-tense statements presently have the property of
truth inhering in them or which future-tense facts presently
exist. Thus, by means of his perception of presently existing
realities he knows the truth about the future.

8. See especially Gerald J. Holton, ‘Mach, Einstein and the Search

for Reality’, in Ernst Mach: Physicist and Philosopher, Boston
Studies in the Philosophy of Science 6 (Dordrecht: D. Reidel,
1970), 165–99; id., ‘Where Is Reality? The Answers of
Einstein’, in Science and Synthesis, ed. UNESCO (Berlin:
Springer-Verlag, 1971), 45–69;and the essays collected
together in id., Thematic Origins of Scienti c Thought
(Cambridge: Harvard University Press). See also Lawrence
Sklar, ‘Time, Reality, and Relativity’, in Richard Healey (ed.),
Reduction, Time and Reality (Cambridge: Cambridge University
Press, 1981), 141.

9. Veri cationism proposed a criterion of meaning that was so

restrictive that it would consign vast tracts of apparently
 perfectly intelligible discourse to the trash heap of divine
eternity nonsense; moreover, the criterion seemed to be self-
refuting. See the excellent discussion in Frederick Suppe, ‘The
Search for Philosophical Understanding of Scienti c Theories’,
in id. (ed.), The Structure of Scienti c Theories, 2nd edn.
(Urbana: University of Illinois Press, 1977), 3–118.

10. Unless there exists a multiverse in which our observable

universe is but one domain, in which case God exists in the
global time of the multiverse.

11. Cosmic time is related to the local times of a special group of

observers called ‘fundamental observers’. These are
hypothetical observers, associated with the galaxies, who are
at rest with respect to the expansion of space itself. As the
expansion of space proceeds, each fundamental observer
remains in the same place, though his spatial separation from
fellow fundamental observers increases. Cosmic time relates to
these observers in that their local times all coincide with
cosmic time in their vicinity. Because of their mutual
recession, the class of fundamental observers does not serve to
de ne a global inertial frame, technically speaking, even
though all of them are at rest. But since each fundamental
observer is at rest with respect to space, the events he
calculates to be simultaneous will coincide locally with the
events that are simultaneous in cosmic time. One could say
that God exists in the time of the inertial frame of every
 fundamental observer; but then there is no problem, since all
their local times fuse into one cosmic time.

12. See the very interesting piece by R. W. Hepburn, ‘Time-

Transcendence and Some Related Phenomena in the Arts’, in
H. D. Lewis (ed.), Contemporary British Philosophy, 4th series,
Muirhead Library of Philosophy (London: George Allen &
Unwin, 1976), 152–73.

13. Robert C. Coburn, ‘Professor Malcolm on God’, Australasian

Journal of Philosophy 41 (1963), 155.

14. Eleonore Stump and Norman Kretzmann, ‘Eternity’, Journal of

Philosophy 78 (1981), 429–58.

15. A word of clari cation: by ‘eternal’ Stump and Kretzmann

mean ‘timeless’, and by ‘temporal reference frame’ they mean
‘moment of time’. It is also worth noting that this de nition is
not really analogous to simultaneity in STR at all. A better
analogy would be to say that x and y are ET-simultaneous just
in case they both exist at the same eternal present relative to
the eternal reference frame and both exist at the same moment
of time relative to the temporal reference frame. But then God
would be temporal relative to our mode of existence, which
Stump and Kretzmann do not want to say.

16. Stephen T. Davis, Logic and the Nature of God (Grand Rapids:
Wm. B. Eerdmans, 1983), 20; Delmas Lewis, ‘Eternity Again: A
Reply to Stump and Kretzmann’, International Journal for
 Philosophy of Religion 15 (1984), 74–6;Paul Helm, Eternal God
(Oxford: Clarendon, 1988), 32–3; William Hasker, God, Time,
and Knowledge (Ithaca: Cornell University Press, 1989), 164–6;
John C. Yates, The Timelessness of God (Lanham: University
Press of America, 1990), 128–30; Brian Leftow, Time and
Eternity, Cornell Studies in Philosophy of Religion (Ithaca:
Cornell University Press, 1991), 170–2; Garrett J. DeWeese,
God and the Nature of Time, Ashgate Philosophy of Religion
Series (Aldershot: Ashgate, 2004), 164.

17. Eleonore Stump and Norman Kretzmann, ‘Eternity, Awareness,

and Action’, Faith and Philosophy 9 (1992), 477–8.

18. Brian Leftow, ‘Eternity and Simultaneity’, Faith and Philosophy

8 (1991), 148–79;cf.id., Time and Eternity, ch.10.

19. Jonathan L. Kvanvig, The Possibility of an All-Knowing God (New

York: St Martin’s, 1986), 150–65;Edward R. Wierenga, The
Nature of God: An Inquiry into Divine Attributes, Cornell Studies
in Philosophy of Religion (Ithaca: Cornell University Press,
1989), 179–85;Leftow, Time and Eternity, 312–37. See also
Jonathan L. Kvanvig, ‘Omniscience and Eternity: A Reply to
Craig’, Faith and Philosophy 18 (2003), 369–76; Edward R.
Wierenga, ‘Omniscience and Time, One More Time: A Reply to
Craig’, Faith and Philosophy 21 (2004), 90–7.

20. Paul Helm, ‘Eternal Creation: The Doctrine of the Two

Standpoints’, in Colin Gunton (ed.), The Doctrine of Creation
(Edinburgh: T. & T. Clark, 1997), 42–3; Helm, Eternal God, 25–
7, 44, 47, 52, 79.
 CHAPTER 8

OMNIPOTENCE

BRIAN LEFTOW

THE doctrine that God is omnipotent takes its rise from Scriptural
texts:

Lord…nothing is too hard for you. (Jer. 32:17)

Our God is in heaven; He does whatever pleases Him. (Ps. 115: 3,
135: 6)
the Lord…can do all things. (Job 42: 2)
nothing is impossible with God. (Luke 1: 37)
with God all things are possible. (Matt. 19: 26).

These texts concern two linked topics. One is how much power God
has to put behind actions: enough that nothing is too hard, enough
to do whatever he pleases. Let’s call this ‘how strong God is’.1 The
other is how much God can do: ‘all things’. The link is obvious: we
measure strength by what tasks it is adequate to perform, and God is
so strong he can do all things. There are two distinct topics here
because strength does not consist in or entail abilities to accomplish
tasks. I might be physically and mentally strong enough to win a
chess match, but be unable to do so because I do not know the rules
or have not practiced. I might be strong enough to lift a skunk, but
unable because an allergic reaction causes me to pass out when I get
within range. Again, two people might be able to accomplish the
same tasks; if one does them easily and the other strains to equal the
rst, there is a di erence between them, which intuitively does not
consist in the range of tasks they are able to perform.2

The Christian philosophical theologian who seeks to explicate

omnipotence seeks a convincing account of the reality beneath the
‘phenomena’ of Scripture. I now look brie y at some historic
accounts of omnipotence. It emerges that the early history of the
concept emphasized strength more than range of action, with range
coming to prominence in Aquinas’s day. I next consider three recent
attempts to de ne omnipotence. I nd them all wanting, but draw
morals that help me hazard my own de nition. Treatments of
omnipotence typically test such de nitions against putative problem
cases, but space restrictions force me to forgo this for the present.

OMNIPOTENCE: EARLY ACCOUNTS

Some Patristic writers treat omnipotence simply as power over all

things. So Cyril of Jerusalem (c.315–86): ‘he is almighty who rules
all things, who has power over all things.’3 Origen (185–254) went
so far as to argue that ‘God cannot be…omnipotent unless there
exist those over whom He may exercise His power; and therefore,
that God may be…almighty, it is necessary that all things should
exist.’4 The eternity of some world or other, then, followed from the
claim that God is eternally omnipotent. One could avoid this by
treating omnipotence as the disposition to have power over
whatever else exists. Surprisingly, Origen saw it as conceivable or
even possible for there to be too many things for God to have power
over: God ‘made creatures according to some de nite number…
creatures have…a limit, because where there is no limit there can be
neither any comprehension nor any limitation. Now if this were the
case…created things could neither be restrained nor administered
by God.’5 To be omnipotent, then, is to have enough power
(strength) to control all other things. Augustine (354–430) wrote
that God ‘does whatever He wills: that is omnipotence’.6 Swinburne
recently echoed this: ‘God is omnipotent in that whatever he
chooses to do, he succeeds in doing.’7 There are problems here. On
this account, even a rather weak being could come out omnipotent
by carefully choosing to do only things at which it was guaranteed
to succeed. Perhaps what’s intended is that an omnipotent being is
incapable of choosing to bring something about and failing to do so,
but then a being able to choose to bring about only one thing, but
guaranteed to bring it about if it does, would come out omnipotent.
Again, nothing would seem to stop an ill-informed omnipotent being
from choosing to do something it is metaphysically impossible to do.

Augustine often discusses the paradox that a God who ‘can do all
things’ can’t do some things we can do. He writes that God
is omnipotent to make all things He may have willed to make. [But]
He cannot die [or] sin [or] deceive [or] be deceived…were He able
to do these, He would not be omnipotent.8 Neither is His power
diminished when we say that He cannot die or [err]—for this is in
such a way impossible to Him, that if it were possible for Him, He
would be of less power. But…He is rightly called omnipotent though
He can neither die nor [err]. For He is called omnipotent on account
of His doing what He wills, not on account of His su ering what He
wills not; for if that should befall Him, He would by no means be
omnipotent. Wherefore He cannot do something for the very reason
that He is omnipotent.9

The inabilities come from Scripture, which tells us that God is

immortal (1 Tim. 6: 16), omniscient about what we do and think
(Ps. 139), and cannot be so much as tempted to do evil (Jas. 1: 13).
Augustine tries to rationalize these. I suggest that he is working with
a strength account in these passages. He probably has such thoughts
as these in the background: ceteris paribus, no one wants to err. No
one perfectly rational and informed whose existence would on
balance be worth continuing would want to die. No one all-good
would want to sin or deceive. The only question for someone with
God’s other attributes, then, is whether he has strength enough to
have what he wants and not get what he does not want. Given God’s
other traits, only a lack of power could explain his sinning, erring,
etc., and so these are ‘in such a way impossible to Him, that if (they)
were possible for Him, he would be of less power’. So God is of
maximal strength, Augustine may think, only if he is so far from
having these befall him that they literally cannot happen to him.
The tie to strength became clearer when Anselm (1033–1109)
reworked Augustine in Proslogion 7:

He who can do these things [be corrupted, lie, etc.] can do what is
not good for himself and what he ought not to do. And the more he
can do these things…the less power he has against…adversity and
perversity. He, therefore, who can do such things, can do them not
by power but by impotence.10

With the strength-notion (and other texts of Augustine) in mind,

Peter Lombard (1095/1100–1160) made it a ‘sign’ of omnipotence
that God can su er nothing and cannot be impeded when he acts.11
In his Sentences, which became the standard theological textbook till
the 1300s, Lombard wrote that God is omnipotent because he ‘can
do all things being able to do which belongs to power’,12 having
previously suggested along Augustine’s line that lying, sinning,
dying, failing, being defeated, and being miserable are things
‘ability’ to do which is really a matter of impotence, not power.
William of Auxerre (1140–1231)—a household name then as now—
wrote that ‘God is omnipotent because He can do from Himself and
through Himself whatever He wills to do. By “whatever” is
understood a fullness which is not determined to some particular
e ect…” from Himself and through Himself” removes need of
extrinsic helps, and compulsion, and extrinsic impediment.’13
‘Whatever He wills to do’ harks back to Augustine. Here the note of
strength predominates, though the account of ‘whatever’ tries to
suggest at least that God can do other than he in fact does.14 The
denial of extrinsic impediment, present also in Lombard, becomes
important below.

The Summa Theologiae of Thomas Aquinas (1224–74) took over

from the Sentences as standard textbook. There we nd an
in uential range-centered account. Philosophers tend to misread it,
likely by focusing only on the paragraph in which Thomas states it
and ignoring his explications, arguments, and quali cations
elsewhere. With these taken into account, his de nition turns out to
be that

God is omnipotent at t = df. God can bring about all states of

a airs that can be produced at t,15

where something can be produced at t only if its being produced is

compatible with there having been the past there actually has
been.16 Thomas provides a reasonable gloss on the scriptural texts.
It’s not true that ‘all things are possible’ with God—God can’t make
round squares. Rather, Aquinas suggests, all products that are
producible as of t are ‘possible with God’—God has power enough to
actualize the whole range of the producible, whatever that is.
Aquinas pushes the strengthnotion to explain certain apparent lacks
of ability, writing that God can’t be moved because this would imply
lack of power, and so can’t walk: thus suggesting that God can’t
walk because he is omnipotent.17 This might suggest that one walks
because one lacks the power to stay still. But Aquinas’s thought may
instead be that being omnipresent, God has no place to go—he
already is everywhere. Thomas cashes out the claim that God is
everywhere inter alia in terms of his having complete power over all
places and their contents.18 If this is correct, God could move—be
rst only in some places and then only in others—only due to a lack
of power. Walking is an issue because God is omnipotent (and so,
we’d think, can walk) but lacks a body (and so can’t). Yet walking is
actually an easy case to handle. Either it is or it is not absolutely
possible that God take on a body. If it is, he has the power to walk,
even if he currently lacks a body. One has the power to walk if (so
to speak) one can just will it and be walking—if walking is just a
volition away.19 For God, this is so—but the volition would be to
make a walking body and be incarnate in it mid-stride. If it is not
absolutely possible that God take on a body, being unable to walk
doesn’t count against his omnipotence, as omnipotence doesn’t
require being able to bring about the impossible.20

Duns Scotus took Aquinas’s de nition to give a necessary but not

su cient condition of omnipotence: he added that something is
omnipotent only if it can produce all that it can produce
immediately rather than through some helping cause.21 This built
into the de nition a claim from William of Ockham that Aquinas in
fact held.22 Ockham gives Aquinas’s account in a Quodlibet,23 and
commits himself to Scotus’s addendum in another.24 Aquinas’s
de nition as modi ed by Scotus and Ockham became the consensus
medieval account of omnipotence, and has remained in uential.
Almost all recent attempts at de nition continue its stress on range
to the exclusion of strength, and most seek to hew closely to it.25 To
see what the consensus medieval account comes to, and whether it
is viable, I now turn to two recent attempts to de ne omnipotence
which try merely to update it.

WIERENGA

Edward Wierenga’s de nition is that

A being x is omnipotent in a world W at a time t = df. (i) it is true

in W that possibly x strongly actualizes some state of a airs at t, and
(ii) it is true in W that for every state of a airs A, if it is possible
that both there has been the initial segment of W up to t and x
strongly actualizes A at t, then at t x can strongly actualize A.26

Wierenga does not de ne ‘strongly actualizes’, telling us only that if

one causes a state of a airs to obtain, one strongly actualizes it.27
Nor does he de ne ‘initial segment’, saying only that an initial
segment of W up to t consists of all states of a airs in W that occur
before t save those that depend in the wrong way on events after t
(e.g. God’s foreknowing them at t).28 (i) seeks to keep things that
cannot strongly actualize anything from trivially satisfying the
de nition. As Wierenga does not specify that one strongly actualizes
A only if one directly, immediately causes A to obtain, this is
Aquinas’s account with one key di erence: Aquinas’s version of (ii)
would have ‘up to t and that there exist some y that strongly
actualizes A at t…’ Aquinas does not say by whom A is to be strongly
actualized.

In (ii), a quanti er within the scope of ‘it is true in W’ governs a

free-variabled sentence, (it is possible that both there has been the
initial segment of W up to t and x strongly actualizes A at t) D (at tx
can strongly actualize A).29 As it occurs within the scope of ‘it is
true in W’, this open sentence really amounts to

(1) (it is possible in W that both there has been the initial
segment of W up to t and x strongly actualizes A at t) ⊃ (at t
in W x can strongly actualize A).

(1)’s antecedent asserts that in W there is a possible world in which

there has been the initial segment of W up to t and x strongly
actualizes A at t. We might read (1)’s consequent in three ways:

(a) as that at t in W there is a possible world with W’s initial

segment up to t in which x strongly actualizes A. So read, (1)’s
consequent is identical with (1)’s antecedent, save for
asserting that this world is possible in W at one particular
time. Most would see this di erence as vacuous, as most
would think that what is metaphysically possible in a world
does not vary over the course of its history. So on this
reading, (ii) is vacuous.
 (b) as that at t in W there is a possible world in which x strongly
actualizes A. But this would be true if there were a world with
an initial segment other than W’s in which x does A at t. This
wouldn’t entail that x was able to do A in a world with W’s
initial segment—e.g. W itself. It would entail that at t in W, x
could have strongly actualized A, not that x can. If there is a
state of a airs S x could have strongly actualized but cannot,
x is less powerful than x might be, unless gaining the power
to actualize S would have cost x an at least equivalent amount
of power. But there is no reason to expect this to be true in
general, and if there is one state of a airs x could have
actualized for which it is not true, x has less power than x
might. Nothing with less than its greatest possible power is
omnipotent. So on this reading, (ii) is arguably not su cient
for omnipotence at t in W.

(c) as involving a thicker notion than the modal ‘◊,’ primitive

power-locution, entailing not that in some possible world x
does, but that in the world under discussion x has the power
to bring about. On this reading (i) doesn’t after all keep things
that can’t strongly actualize anything from satisfying the
de nition: for then it could be true both that possibly x
strongly actualizes some state of a airs at t and that x can’t—
i.e. lacks the power to—actualize anything at t in W.30

Wierenga’s clause (ii) also faces McEar problems.31 McEar is

essentially able only to scratch his left ear. As McEar has this ability
essentially, if McEar exists at t in W, he satis es (i). As this is
McEar’s only possible action, if he can do it at t in W, he also
satis es (ii). But if McEar has this ability essentially, he can scratch
his ear at t in W. So on Wierenga’s account, McEar is omnipotent.
Wierenga sees the di culty, and replies that if someone can scratch
his ear, then he can also do inde nitely many other things—move
his arm, move his hand, scratch his ear at t, from t to t*, etc.—and
so McEar isn’t a possible being.32 But to make it clear that McEar is
possible, we need only adjust his description: he’s essentially unable
to accomplish anything save as part of or as a logical consequence
of scratching his ear.33 Wierenga adds a general doubt about
essential limitations: surely, he suggests, for any agent who lacks
certain abilities, it is possible that God give it those abilities.34 But
the general doubt can be resisted. If some agents have non-trivial
essential properties (that is, essential properties other than those
like being self-identical or a being that are shared with all possible
things), and even one of these essential properties entails the lack of
some power, there are essentially limited agents. I shortly suggest it
to be plausible that both premises are true, but rst I must note that
for Christians, at least, problems crop up here from an unexpected
direction.

Christians believe that one human, Jesus, was also divine. The
orthodox Chalcedonian view has been that this entails that one
human was omnipotent; those with this sort of Christology, then,
can’t claim that being human is incompatible with being
omnipotent, and so, one might think, can’t say that being human
entails the lack of any power whatsoever. Further, the same will
hold for any nature God the Son might have taken on. However,
even if being human is compatible with being omnipotent, it does
not follow that just any human could have been omnipotent. In fact,
no human not actually assumed by a divine Person could have been.
For God is no body-snatcher. But had God the Son taken over the
body and soul that would otherwise have been mine to make
himself human, he would in e ect have stolen my natural
endowment. Further, actual humans are actually substances. Jesus’
body and soul do not on their own constitute a substance. They are
not independent existents; they exist only as adjuncts supported in
being by the substance who is God the Son. It is implausible that
being a substance is a contingent property—that something that is a
substance could have failed to be one. If this is correct, one can’t use
the Incarnation to argue against there actually being essentially
limited agents: no actual human could have been omnipotent in
virtue of having been the vehicle of a divine incarnation. One can
see this more simply still: had my body and soul been used for the
Incarnation, I would not have existed. The person in my body would
have been not me, but God the Son. So God’s becoming incarnate in
my body and soul would not entail that I was omnipotent.

Again, one strand of Christianity might see the claim that some
agents are essentially limited as controversial. Picking up on a few
New Testament texts, Eastern Orthodoxy believes that salvation
involves eventual ‘dei cation’. Richard Swinburne approvingly
interprets Maximus the Confessor’s version of this as implying that
the blessed in heaven become omniscient, omnipotent, and perfectly
good.35 But if God can make us so, the question arises of whether
there are any sorts of agent he can’t do this for: perhaps he could
deify rabbits, frogs…. Be this as it may, much turns on how the idea
of being made omnipotent is parsed out. A perfectly good God
would not grant omnipotence without also granting su cient
knowledge and goodness to use it properly or else hemming in its
use by a use of his own power: an omnipotent but evil, partly
corrupt, or partly ignorant being left free to act as it pleased would
be an evil too great to reconcile with God’s existence. On a weak
reading, becoming omnipotent could mean simply that God so
reforms the blessed’s characters and informs their knowledge that
they no longer make prayers whose granting would in the long run
go counter to the best good God has in mind, and God accordingly
grants all their prayers: anything they choose to bring about in this
way, they succeed in bringing about. This sort of ‘becoming
omnipotent’ would be limited to things possibly with su cient
cognitive complexity to pray, and would leave any essential intrinsic
limitations an agent had in place, since the agent’s intrinsic powers
would not be heightened one whit. And it is not in my view genuine
omnipotence at all: the agents in question remain intrinsically of a
sort to be frustrated by external limitations, and can succeed in
what they do only by external help. True omnipotence, I suggest
below, is not compatible with these things.
 On a strong reading, being made omnipotent would mean being
granted a new intrinsic property in virtue of which one henceforth
has that much power.36 Duns Scotus and others would argue that
this is impossible—that there cannot be two such omnipotent beings
—but I cannot delve into this here. Still, what could this intrinsic
property be? Being omnipotent is a matter of having powers. One
has powers in virtue of having an underlying categorical attribute
which supports or subvenes them.37 In God’s case, the underlying
categorical is deity. It is not clear or at all intuitive that any other
underlying nature could support omnipotence. But to bestow a case
of deity on a creature would be to make a created thing a God. It is
plausibly part of the divine nature to be uncreated. I therefore
suggest that God just cannot make another being intrinsically
omnipotent. It follows that every non-divine nature essentially
involves some limitation of power: even if there is no particular
power such that the nature guarantees that the being bearing it
lacks that one, the nature guarantees that there are some powers the
being does not have. It follows further that for every non-divine
being, there can be powers such that God cannot give that being
those powers: if I were improved up to (so to speak) one power
short of omnipotence, God couldn’t give me that last power.
Further, if I’m right in this last, every created being has at least the
non-trivial essential property of being non-divine, and this su ces,
as I’ve said, to involve essential limitation. But of course it’s
plausible that we have other, positive non-trivial essential
properties. I am human. So while any chicken embryo has the
ability to grow by natural processes into a chicken, I do not; while
any cat has the ability to clean its own cat-fur, I do not. Plausibly I
am essentially human; if I am, God can’t give either ability to me.
We can also generate essential limitations without substantive
essentialism about non-world-indexed properties. I have not just
abilities, but world-indexed abilities: if in W I can run, I also can-
run-in-W. All world-indexed attributes are essential. Thus any
world-indexed ability I lack, even God can’t give me. Finally,
Wierenga’s claim, again, is that for any agent who lacks certain
abilities, it is possible that God give it those abilities. But God, as
we’ve seen, lacks abilities to die and err. If he is essentially eternal
and omniscient, he cannot give himself these. Thus Wierenga’s
response fails, and on his de nition, McEar is omnipotent.

FLINT/FREDDOSO

Like Wierenga, Flint and Freddoso seek to explicate the consensus

medieval account. They suggest that

(FF) S is omnipotent at t in W if and only if for any state of a airs

p and world-type-for-S Ls such that p is not a member of Ls,
if there is a world W* such that

I Is is true in both W and W*, and

II W* shares the same history with W at t, and
III at t in W* someone actualizes p,
then S has the power at t in W to actualize p.38
A world-type is a set of subjunctive conditionals about how agents
would freely act in various circumstances—‘counterfactuals of
freedom’ (CFs) about those agents—such that for every CF P, either
P or ¬P is a member of the set.39 A true world-type is one all of
whose members are true. A world-type-for-S is a subset of a true
world-type, containing only CFs about agents other than S.40
Roughly, W and W* share the same history at t just if they share all
non-future-dependent states of a airs at t: that is, just if all that is
and was the case at t in W which is not a consequence of what will
be the case after t also is and was the case at t in W*. So if world-
types are non-future-dependent, as in standard Molinism, (I) is
redundant. Modulo the introduction of world-types, this is Aquinas’s
account, with ‘that can be produced’ speci ed (in (III)) to ‘that
someone can produce’. I read ‘someone’ as ‘at least one agent’, on
the assumption that it is the ordinary-language version of an
existential quanti er.

Ho man and Rosenkrantz o er the following counter-example to

(FF). Consider possible worlds W and W*, sharing the same history
up to t, such that in W a contingently omnipotent being is
omnipotent for the rst time at t. In W*, no omnipotent agent ever
exists. So, they suggest, someone in W* can bring it about at t that a
ball rolls and no omnipotent being ever exists. But the omnipotent
being in W can’t bring this about at t. So (they conclude) there is an
omnipotent being in W, and it can be the case that at t in W*
someone actualizes p, but it is false that the omnipotent being has at
t in W the power to actualize p. Now they infer that someone in W*
brings it about that a ball rolls and no omnipotent being ever exists
via the claim that

(HR) if S brings about p, q obtains, and ¬q is not in the power of

any one other than S, S brings it about that (p . q).41

But this principle is dubious. Booth brought it about that Lincoln no

longer exists. One hundred years later, let’s say, it came about
uncaused that q, where ¬q is necessarily uncaused. Is it plausible
that 100 years after his own death, Booth brought it about that
Lincoln no longer exists and q? If q was uncaused, it seems more
plausible that this conjunction has no full cause: that nothing
brought it about, though Booth brought about one conjunct.
According to Duns Scotus, God brought it about that 2 + 2 = 4, but
bringing about the opposite was and is in no one’s power. Suppose
for the nonce that Scotus was right. On (HR), it would then follow
that Booth brought it about that Lincoln no longer exists and 2 + 2
= 4. But this would be false were Scotus correct: the cause of a
conjunctive state of a airs both of whose conjuncts are caused is the
composite consisting of their causes. Again, suppose we live in a
physically deterministic universe. Then it obtains (say) that I will
eat eggs for breakfast tomorrow, and the opposite is in no one’s
power, even God’s. (Were God able to bring about the opposite, it
would not be the case that given the past (which includes the
physical laws’ obtaining) there is just one physically possible future,
and so the universe would not be physically deterministic.) Then on
(HR), Booth brought it about that Lincoln no longer exists and I will
eat eggs. But Booth would have had no e ect on my breakfast at all,
at any time; again, the cause of this conjunctive state of a airs is the
composite of the conjuncts’ causes. So I take it that (HR) is false,
and the example fails.

Oppy o ers the following counterexamples to (FF):

Consider the conjunctive state of a airs the cat comes in and if Jones
were in C at t, he would freely decide at t to let out the dog…suppose
now that S is omnipotent. Since [this] state of a airs…does not
belong to Ls, it follows from (FF) that an omnipotent being can
bring (it) about…though it cannot bring about one of the
conjuncts…consider…the disjunctive state of a airs if Jones were in
C at t, he would freely decide at t to let out the dog or if Jones were in C
at t, he would freely decide at t to go to the bathroom…(this) does not
belong to Ls, and so it follows from (FF) that an omnipotent being
can bring [it] about…even though (it) can bring about neither of
the disjuncts. Not good.42

The rst is a problem given the falsity of (HR), for this suggests that
something can bring about a conjunctive state of a airs only by
bringing about both conjuncts or something relevantly like bringing
about both conjuncts.43 One can address both by asking how CFs get
included in a true world-type. In standard counterfactual logics, if p.
q, then p q. So it seems that by making p . q true, one makes it true
that p q. Flint and Freddoso allow this in the case of CFs at one
point:

Any free being will have some say in determining which world-type
is true … since Jones is free to decide whether … to write … to his
wife, it is up to him whether the true world-type includes (that) if
Jones were in C at t, he would freely decide at t to (write) to his
wife … However, the vast majority of the counterfactuals which go
to make up a world-type relate to beings other than Jones, and
Jones … is powerless to make such counterfactuals true or false.44

Flint and Freddoso deny that anyone other than Jones can a ect the
truth-value of a CF about Jones.45 But they also allow that ‘a
person’s power is … in large measure … his ability to in uence the
free actions of others … e.g. by restricting their options … or by
persuading or dissuading them’.46 If so, then if there are CFs about
Jones, someone not Jones can bring it about that a CF about Jones
is true. Let’s distinguish complete from incomplete circumstances of
a free choice. A complete circumstance of a free choice K is the
entire history (in Flint and Freddoso’s sense) of the world prior to t,
plus those states of a airs obtaining at t whose obtaining in no way
depends on what is freely chosen in K. If complete circumstances
are sets of states of a airs, an incomplete circumstance for K is any
non-empty subset of its complete circumstance. There are CFs for
both complete and incomplete circumstances. Someone not Jones
can bring it about that a CF for an incomplete circumstance is true.
Let the incompleteness in the circumstance for K be that it excludes
everything I do so to persuade Jones, not in C, that

(2) were Jones in C, he would freely do A.

Suppose I give Jones good reasons R to do A if in C. It is possible

that he accept them voluntarily and freely. Perhaps, for instance,
part of his reason to accept them is that he trusts me on certain
matters, he would not accept them if he did not trust me, and his
experience of me is such as to leave this trust an attitude he freely
adopts. Again, if I give him evidence in favor of doing A in C, the
strength of the evidence may not be such as to coerce (as it were)
belief: he may be left with a decision about whether the reasons are
good enough. Again, even if Jones is caused to form the belief that
R are good reasons to do A, Jones is able not to adopt R and make
them his own reasons to do A. He must consent at least implicitly to
their in uencing his action, rather than (say) stubbornly giving
greater weight to reasons he had had against doing A, and this
consent can only be voluntary. Suppose, then, that Jones freely
accepts the reasons, nothing other than my giving them being
relevant to his doing so, I do not cause him to do so, but given that
he has accepted them, he now is such that (2) is true. Then the
following seems a legitimate chain of full explanations: (2) is true
because Jones freely accepted R, Jones freely accepted R because R
seemed good to him, and R seemed good to him because I presented
them persuasively. It seems to follow that (2) is true because I
presented R persuasively: which is to say that in some appropriate
sense of the term, I brought it about that (2) is true. This is not a
case of strongly or weakly actualizing the truth of (2), as Flint and
Freddoso de ne these terms, but their de nitions do not exhaust all
ways of bringing something about. Perhaps no one can bring it
about that a CF for a complete circumstance is true—at any rate
that’s a further issue. But where C does not include the persuading,
a persuader can a ect what Jones would do in C.47 Omnipotent
beings can doubtless be very persuasive. So it is not the case, I
submit, that an omnipotent being is unable to bring about the
problem conjuncts and disjuncts.

Oppy raises a question for (III):

Suppose that S is omnipotent at t in W (and) leaving S aside…no

agent in W…acting alone, can bring it about that p in any world W*
that shares its history with W, but…a group of agents in W…acting
together, can bring it about that p in…W*…Suppose (e.g.) that no
human acting alone can bring…it about…that a particular car is
raised one metre above the ground using nothing but human muscle
power…but that there are groups of four people who (can). As
things stand, (FF) would allow that S is omnipotent even if it cannot
bring it about that this particular car is raised one metre above the
ground using nothing but human muscle power. And this seems
wrong.48

The thought is that (III) only requires someone omnipotent to be

able to do what any one other agent can, not what groups of agents
can. I’m not sure this reads (III) correctly. Many, I think, would
render ‘someone’ in (III) as (∃x)(x actualizes…), which would in
turn translate as ‘at least one agent…’ Further, something Flint and
Freddoso say about strong actualization could help here. They rst
say that one who strongly actualizes it that P causally determines it
that P, then add ‘i.e. does something which in conjunction with
other operative causal factors constitutes a su cient causal
condition for p’s obtaining’.49 The second permits that one who
strongly actualizes it that P be only one causally contributing part of
the su cient causal condition, rather than on its own causally
determining it that P. On this account, any one of the four humans
can strongly actualize the car’s rising etc., the other humans
counting as ‘other operative causal factors’. So a being able to do
what any one other agent can do will be able to strongly actualize
this too—even absent any other causal factors. And there are no
states of a airs essentially able to be strongly actualized only by
many agents: whatever one agent can actualize in conjunction with
other agents, the one agent can in (FF)’s sense strongly actualize.
Oppy has noted in correspondence that this creates a problem: we
wind up saying that it is in an agent’s power to strongly actualize it
that P even if the agent requires signi cant help to do so—say, the
help of 20 million other voters in electing a government—and this is
a bit unintuitive. But only a bit: ‘strongly actualizes’ is an arti cial,
technical term, without an established sense to violate, and without
this, it’s not clear what intuitions we really have on the subject.
 Still, there are more-than-exegetical points to make here. The
humans don’t move the car solely by muscle-power. Their moving
the car also depends on non-muscular events such as the synaptic
rings involved in executively intending to lift the car. If that’s
compatible with its being the case that human muscles supply the
only lifting power involved, it’s not clear why an omnipotent being’s
volition that four humans move the car by muscle-power alone
could not play a parallel role. In this scenario, the causal chain
resulting in the car’s rising would have a further link, a divine
volition that sets o the executive intending, or perhaps
overdetermines its rst e ect. Or perhaps the divine volition could
just usurp the brain’s place and stimulate the muscles directly. We
might also suppose that God induces in himself a liability to be
a ected by the four’s volitions, so that when they will to lift the car,
he so wills as partly to usurp the brain’s place in the causal chain,
providing some link between their intentions and the motion of
their muscles. It would need some doing to show that such a causal
chain (particularly if lawlike) is incompatible with the lifting’s being
an action. The lifting that results would not be free in the rst case,
but might be so in the second—Frankfurt questions arise here. But
free or not, in the rst, second, and fourth cases the lifting would be
an action ascribable to the four, and in the third and fourth, the
omnipotent being would lift the car using only human muscles. In
none of these scenarios does an omnipotent being’s willing add any
lifting power other than human muscle. But then why couldn’t an
omnipotent being have four humans lift the car by muscle-power
alone?

One might reply that this is irrelevant, because what Oppy has in
mind is not whose the lifting power is, but rather the bringing-about
of the state of a airs four humans lift the car by their muscle power
and nothing, including an omnipotent being, helps them. But if an
omnipotent being is part of W’s history, as Oppy assumes, in no
possible world sharing W’s history at t do four humans bring this
about by themselves. Humans cannot block the in uence of
someone omnipotent: if S chose to help them, they could do nothing
about it. Thus four humans raise the car by their own muscle-power
only if S cooperates by leaving them alone—only if S’s refraining
from action is part of the causal history of their joint action. With S
lurking in the background, humans cannot e ect the second
conjunct—and if p obtains and one brings about q, that is not
su cient to entail that one brings it about that p . q. So this is not a
state of a airs they can bring about by human muscle-power alone.
If S cannot do so either, then, this counts nothing against S’s
omnipotence on (FF), for this state of a airs fails (III). Oppy replies:

When I am at the gym, I lift a weight unaided provided that no one

helps me…I can lift the weight using my muscle power alone even
though…my trainer could prevent me from lifting the weight…it
doesn’t seem right to me to say that since his consent is required—
he has to refrain from intervening to prevent me from lifting the
weight—I can’t lift the weight using my muscle power alone.50
But the question here is not whether only my muscles provide lifting
power but whether I can bring it about that only my muscles do so,
i.e. whether I can prevent my being helped. I can, at the gym: if
need be I can shoot the trainer. His consent is not required.

Again, Gellman asks us to consider S’s doing A while no one

other than S actualizes it that S does A. This, he claims, is something
S can actualize but an omnipotent being other than S cannot.51 But
this is a conjunctive state of a airs, and the claim that S can
actualize it requires either something like (HR) or that S itself be
able to bring it about that no one other than S actualizes that S does
A. If an omnipotent being coexists with S, S cannot on S’s own bring
it about that this being does not actualize S’s doing A. This is true
even if our basic actions consist in what we agent-cause and nothing
can cause an agent cause to cause whatever it causes. For in this
case it is the nature of agent causation and human agency that
accounts for the fact that the omnipotent being can’t bring it about
that S does A. And S cannot bring it about that this is the true story
about human agency. But if this story about agent causation is not
the reason S can’t bring it about that the omnipotent being does not
actualize S’s doing A, the omnipotent being must cooperate, and so
S can at most causally contribute to this.

Thus (FF) survives some arguments in the literature. But it still

has problems. Flint and Freddoso let part of an omnipotent being’s
power consist in power to weakly actualize states of a airs, where S
weakly actualizes it that P just if S strongly actualizes it that some
S* is in a circumstance C such that (S* is in C) > (S* freely strongly
actualizes it that P): again, ‘a person’s power is…in large measure…
his ability to in uence the free actions of others’.52 But suppose
Whitehead’s metaphysics were true. Then everything, even quarks,
would be conscious in some way, and open to divine persuasion. On
Whitehead’s account, God can strongly actualize only what is
necessary to persuade and to some extent limit options: he can
directly cause other things to apprehend possibilities, and make
some possibilities seem more attractive or relevant than others, and
this is all he can directly do. God must bring about everything else
he brings about through others’ help. Whitehead does not appear to
believe in CFs. But even if there were such CFs as let God get his
way in the end—that is, even if he o ers reasons R to Jones in C
knowing that were Jones in C and presented with R, Jones would do
A—would such a God be omnipotent? Surely it’s reasonable to
require an omnipotent being to be able to strongly actualize more
than this. For Whitehead’s God, in e ect, can’t do anything directly
or on his own but talk to others. Yet (FF) does not require this:
‘actualize’ is in both occurrences the disjunctive ‘strongly or weakly
actualize’.53 Thus if Whitehead had the necessary metaphysical truth
about the nature of things, his deity could well come out
omnipotent on (FF).54

Scotus’s addendum to Aquinas assures that omnipotence entails

the greatest possible range of power strongly to actualize at t. An
omnipotent being must have this. Suppose that God and Schmod are
both omnipotent per (FF), but God can strongly actualize more than
Schmod can, because God can bring about on his own some things
Schmod can bring about only by persuading free agents to do them
or perhaps manipulating them into doing them by the use of CFs. If
this is so and God and Schmod are otherwise equal, God seems more
powerful than Schmod. But nothing can be more powerful than an
omnipotent being. Thus only a being of maximal strong-
actualization range can be omnipotent, and (FF) fails for allowing
that something can be more powerful than someone omnipotent.
I’ve suggested that an omnipotent being can in uence the content of
its world-type. As this is so, but (FF) is entirely in terms of ability to
actualize states of a airs not included in world-types, (FF) lets one
omnipotent being be more powerful than another in a second way:
there could be two beings satisfying (FF), only one of which was
able to persuade Jones, and so one of which had the ability to
render true a CF the other could not.55 But again, nothing can be
more powerful than an omnipotent being.

Omnipotence, then, must entail maximal power to strongly

actualize. I now argue that such power exhausts it: that powers to
persuade or weakly actualize are not of the right sort to help
constitute a being as omnipotent. First, power to persuade: if this is
power over others’ libertarian-free actions, it seems the wrong sort
to gure in a de nition of omnipotence.56 ‘Persuade’ is a success
term: I persuade you only if I succeed in my attempt to do so. How
much persuading I can do depends on how hard the hearts are
around me. So power to persuade is not intrinsic: having it depends
on what is the case outside the agent. This is necessarily the case; it
is a matter of what it is to persuade. But omnipotence is an intrinsic
attribute. William of Auxerre held that an omnipotent agent needs
no help to do what its omnipotence gives it power to do. This
became part of the consensus medieval account, and is plausible.
But if a being can achieve an e ect only by exercising a power
whose possession must be partly determined extrinsically, it
necessarily needs the help of those extrinsic circumstances that help
give it the power to achieve that e ect. This is true even if the
circumstances obtain necessarily. If they obtain contingently, there
is another problem. Plausibly only the limits of logic, other
metaphysical necessities (e.g. that there be no ideas that sleep
furiously), the content of possibility (if this is not itself
metaphysically necessary), and perhaps its own decrees or the will
of another omnipotent being can impede the will of an omnipotent
being. But if a being can achieve an e ect only by exercising a
power whose possession must be partly determined extrinsically,
and the relevant circumstances can fail to obtain, lack of those
circumstances is another impeding factor to which it is subject.57 It
could seem to follow that no one can be omnipotent, since any
reasonable candidate for omnipotence would be able to persuade.
But another moral is that any being is omnipotent only if powers of
this sort are irrelevant to whether one is omnipotent—and they are
so only if their presence would not help constitute one as
omnipotent. An omnipotent being needs no help to do what powers
constituting it as omnipotent empower it to do. Persuaders need
help. So the power to persuade is not one of those powers.

Again, power to persuade is not guaranteed to produce the

desired e ect. If those I’m trying to persuade have libertarian
freedom, it is beyond my control whether I succeed, though it is in
my control (of course) greatly to raise the odds that I do. Suppose
that Zod and God o er the same blandishments to Smith in the
same way on two occasions, Smith being in the same relevant
internal state in both. Displaying the notorious persnicketyness of
free agents, Smith accedes to Zod but not God. It is unintuitive that
this shows Zod to have more power of any relevant sort, for Zod has
done only what God has. I’m not claiming that an omnipotent being
would not as such have the powers to speak, to o er reasons, and so
on. These are the purely intrinsic powers that are exercised in cases
of persuasion. But power to persuade is not intrinsic, not
independent of circumstances. If God is omnipotent, he doesn’t
depend for his omnipotence on the lucky break of having easily
persuadable people around him, or more generally on circumstances
being propitious. So I suggest that while it is part of what makes
God omnipotent to be able to do all those things which are his
contribution to a case of persuasion, it is not part of what makes
him omnipotent to be able to persuade. If omnipotence is intrinsic,
it is not even partly constituted by non-intrinsic powers.

But then what of power weakly to actualize? Flint and Freddoso

see persuasion as one instance of this, and so it is, if the relevant
CFs are true. Suppose that Zod o ers Smith the same blandishments
in two possible worlds featuring di erent true CFs, and due to the
di erences, Smith does as Zod asks in just one world. Again, it is
unintuitive that this shows Zod to have more power of any relevant
sort in that world. Our intuitions about persuasion shouldn’t change
if we learn that it sometimes takes advantage (as it were) of CFs.
Power to weakly actualize is power to accomplish certain things
only by means of extrinsic helps (the contingent presence of
propitious CFs). So it can’t help constitute an intrinsic attribute.
Omnipotence, then, is power simply to strongly actualize, or cause
to be the case.

(FF) avoids Wierenga’s McEar problem by speaking in (III) of

someone’s actualizing p rather than the omnipotent being’s
actualizing p. But (FF) has di culty with Lonely McEar. LM can
only scratch his left ear, but exists only in worlds in which he is the
only agent. In one such world, LM scratches his ear. So LM comes
out omnipotent on (FF).58 Further, LM suggests a series of
omnipotent beings one more powerful than the other: for there are
Lonely McEars. McEar-Nose-and-Throat, etc.59 Now there is a
dialectical problem here, in that LM and kin are possible just if God
is not a necessary being, and theists tend to believe the latter. But
God can exist only if someone can be omnipotent, and we’re
considering whether there is an adequate account of omnipotence, a
claim relevant to whether there is a property of omnipotence to be
had. So it’s not clearly legitimate to invoke the necessary existence
of God to block appeal to LM. Further, even if we do, we can treat
the Lonely McEar argument as trading on signi cant counterpossible
conditionals—something common in philosophy.

We can now draw a more general moral. The high-medieval

consensus view de ned omnipotence in terms of states of a airs
strongly actualizable at t. This naturally raises the question ‘by what
or whom?’ Wierenga’s answer ‘by the omnipotent being’ and (FF)’s
answer ‘by something’—which is to say by anything, or on a
reasonable interpretation even any group—face McEar problems.
The only other reasonable answer would be ‘by something(s) other
than the omnipotent being’. But this would fail for reasons Aquinas
saw: if nothing else at t in any possible world sharing W’s history
can do a miracle out of its own natural powers, it would follow that
one could be omnipotent without being able to do miracles out of
one’s own natural powers.60 We thus have yet to nd a parsing of
the high-medieval consensus view that can handle McEar.

A way out is not hard to spot. We must answer the question ‘by
what or whom?’ as (FF) does. But we must also jettison Wierenga
and (FF)’s claim that to assess omnipotence in W at t, only worlds
which share W’s history at t matter. This is what makes Lonely
McEar di cult. The thought behind the claim is that it is now
impossible to bring about anything incompatible with the past’s
having been as it was: if it was the case that P, now or in the future
nothing can occur that entails that it was the case that ¬P, nor can
the contents of the past be shifted. As this is so, Wierenga and (FF)
think, it does not count against someone’s claim to omnipotence
that he/she is unable to do so.61 They infer that the only power that
matters for being omnipotent at t in W is power having which is
compatible with t’s past in W. But there are other ways to deal with
the xity of the past. One is simply to note that power to bring
about (say) that the Germans lost World War II (or anything
entailing this) is not intrinsic: to have this power, one must be sited
either before World War II or outside time, and so whether one has
it depends on one’s circumstances. As we’ve seen, intrinsic power
alone counts for omnipotence; extrinsic powers, or their lack, are
irrelevant. This being so, there is no need to jigger one’s de nition
of omnipotence to allow for the past’s xity. What matters is simply
that the omnipotent being have all such intrinsic powers as would
allow it to bring about a German loss were it appropriately sited. To
parse the consensus view, then, let us simply speak of states of
a airs with the modal property of being strongly actualized in some
possible world. If the states of a airs relevant for counting as
omnipotent are those actualized in some possible world, not in some
possible world sharing history with W, this keeps McEar at bay.

STRENGTH VS. RANGE

Still, the consensus view on its own won’t do, however parsed. The
most basic problem with (FF) is one it shares with almost all
de nitions of omnipotence since Aquinas: it speaks only of the
omnipotent being’s range of action, not its strength or power. We
can agree in advance of any detailed account of omnipotence that
an omnipotent being is as powerful as it is possible to be: no one
could be more powerful than someone omnipotent. But now
consider two deities, Schmod and God. Schmod ‘can do all things’,
but nds some of them hard (though not ‘too hard’). God can do all
Schmod can, but nds none of it hard. One legitimate explanation of
this is that God is stronger. If someone is stronger than Schmod,
Schmod is not omnipotent, even if Schmod ‘can do all things’. If this
is a coherent example, a de nition of omnipotence needs to assure
maximal strength; talk of range alone is inadequate.

Two things in this argument need explicating: what ‘strength’ or

power can mean in talk of God, and what it consists in for Schmod
to nd something hard. Neither is a physical notion in this case, but
not all power is on its surface physical, and not everything we nd
hard seems hard to us due to what seems to us a physical lack. Some
people are smarter than others. Perhaps this has a purely brain-
based explanation, but we don’t need to know this to make sense of
talk of the smarter having more powerful minds. As long as we have
a concept of power that does not conceptually involve a physical
basis, we can ascribe power (as vs. powers) to someone we consider
non-physical. Again, if Jones nds doing calculus hard, this may be
because Jones’s brain does not make certain electrical connections
easily, but making sense of ‘this is hard for Jones and easier for
others’ doesn’t depend on such notions.
 Many powers can act with varying intensity: a light can shine
more brightly, a trumpet sound more loudly, a muscle contract or a
mind concentrate harder. These all vary the power of the powers
involved. How much strength/power is associated with a power is a
matter of how intensely one can exercise it, how long one can
sustain that intensity and how easily one can summon it. The last
involve inter alia whatever feelings of pain or e ort may accompany
the e ort and psychologically or physically discourage eliciting or
continuing that intensity. This capsule account of strength suggests
what hardness might be for Schmod. Perhaps Schmod simply has to
concentrate harder to make a decision, or exercise any of his
powers. Perhaps, unless he concentrates very hard, he occasionally
wills slightly the wrong thing, or his power does not accomplish
quite what he wants it to. Perhaps when Schmod wills to do certain
things, he feels just what we feel phenomenologically when we exert
e ort and accomplish something with di culty. Perhaps Schmod
tires: gradually, as he wills rst one and then another e ect, it takes
more e ort to concentrate, painful fatigue sensations increase, his
attempts become gradually more inaccurate, etc. We can understand
this in many ways we would with ourselves that do not a priori
imply the presence of a physical substrate for the mental activity or
e ort. But perhaps too Schmod has a body, and can do certain
things only by exerting its muscles,62 these muscles tire unless he
consciously wills to keep them fresh, and he does not always
remember to do this. I suggest, then, that there is no insuperable
objection to talk of strength as well as range here.
 A defender of a pure range account might reply that Schmod
counts as non-omnipotent on a pure range basis, and so we needn’t
invoke a separate strength consideration to explain this. Consider a
state of a airs S God can bring about without e ort, which Schmod
nds hard. There is in addition to S the possible state of a airs S’s
being brought about easily. God can actualize this. Schmod can’t—
unless by easily persuading God to bring about S.63 But then there is
still a related state of a airs only God can manage, S’s being brought
about easily without persuading someone else to bring about S. So
Schmod is not omnipotent, it would seem, having a less than
maximal range of action. (If only range counts for omnipotence, and
it is not possible to be more powerful than someone omnipotent,
being omnipotent entails having the maximal range.) But this sort of
counter creates problems for God too. There is also the possible
state of a airs S’s being brought about with di culty. Schmod can
strongly actualize this. God can do so only if he can make this
di cult for himself or else cause someone else to bring S about with
di culty. The claim that an omnipotent being can make something
di cult for himself courts paradox: can God make a stone so heavy
that he can lift it only with di culty? If he can’t, the paradox would
say, he is not omnipotent, unless the reason for this lies in some
limitation it is acceptable for an omnipotent being to face (e.g. that
this would entail a contradiction). But if he can, he can have
di culty lifting something, and that (so the paradox would say)
doesn’t be t omnipotence either. Be that as it may, rst making
something di cult and then having di culty dealing with it is not
the same thing as just having di culty ab initio, and so there still
seems to be a possible state of a airs Schmod can and God can’t
actualize, S’s being brought about with non-self-imposed di culty. God
could get this one done by causing or persuading Schmod to bring
about S, as Schmod could co-opt God above. But then as above there
would be another problem state of a airs, S’s being brought about
with non-self-imposed di culty and without persuading or causing
someone else to bring S about. So if the objection we’re considering is
good reason to consider Schmod non-omnipotent, it might also be
good reason to consider God so. On the other hand, if Schmod and
God thus tie, we might just as easily say they are both omnipotent,
as both (we may suppose) have the maximal possible range of
action. For neither can bring about all possible states of a airs, and
for each one Schmod can’t manage because it is too hard, there is
one God can’t, of some such form as we’ve seen, and every other
possible being (we can suppose) has a lesser range of action. Either
way, we lose the contrast between Schmod and God we wanted to
preserve and explain—that just one is omnipotent. If their ranges of
action are equal, talk of range of action doesn’t capture their
di erence in strength or power. So states of a airs like S’s being
brought about easily do not rescue range accounts.

WIELENBERG

If the note of strength is needed, it is natural to wonder whether talk

of an omnipotent being’s strength su ces on its own for a
de nition. Erik Wielenberg suggests that ‘x is omnipotent if and
only if it is not the case that there is some state of a airs… x is
unable to bring about…at least partially because of a lack of
power…’64 This is a pure strength de nition; range isn’t mentioned.
A range of action is not even implicit, because Wielenberg allows
that many sorts of things might prevent x’s using this power in
various ways. Wielenberg has it expressly in view that a
metaphysical impossibility of using one’s strength in a particular
way does not preclude being omnipotent, as long as the
metaphysical impossibility in no way involves a lack of power.65
Even if it is metaphysically impossible that Hercules lift a particular
stone, Wielenberg insists, it makes perfectly good sense to say that
he is strong enough to do so. (Hercules would then lack the ability
for some non-power reason.) McEar is thus a problem for
Wielenberg: we can imagine that McEar satis es Wielenberg’s
de nition but is metaphysically incapable of using his great power
to do anything but scratch his ear.66 However strong McEar is, we
jib at calling him omnipotent. This suggests that strength alone is
insu cient for omnipotence. An adequate de nition must contain
both a strength- and a range-clause. Below, I take Wielenberg’s
de niens as one conjunct in my nal account. This entails that
whatever an omnipotent being does, is done easily. It is not
implausible that the ‘ability’ to nd things hard is a mark not of
power but of weakness, and that an omnipotent being needn’t have
it.
 If there is a prima facie case for both sorts of clause in an
adequate de nition of omnipotence, a natural question is whether
one is basic, the other derived, and whether one can therefore
reduce in the end to the other. It’s plausible that if God can do all
things, this is atleast partly because he is so strong. One might
suggest to the contrary that God’s strength consists in his ability to
do all things. But that God can do all things is a fact about how God
can use his strength rather than something which constitutes his
having it. Ability to do a thing arises from at least three things more
basic, having power enough, having knowledge enough, and
possibly having the desire to do it.67 Further, since it takes more
than power to generate an ability, and so to include various states of
a airs in one’s range of action, facts about range don’t in the end
reduce to facts about strength. This is why a good de nition of
omnipotence must mention both.

RANGE

As a rst pass at what a maximal range of action is, we might

suggest

At t, S’s range of action is maximal = df. at t, (P)(◊(something(s)

cause(s) it to be the case that P) ⊃ (S is all-things-considered able to
bring it about that P)).
This ascribes to S not ability to bring about whatever is
metaphysically possible at t, but ability to bring about whatever it is
metaphysically possible to bring about at t. It leaves space for the
possibility that some states of a airs cannot be brought about at t. If
they cannot, of course, this doesn’t entail that S has a less than
maximal range of action. If a state of a airs can’t be brought about,
it is not in any possible range of action. Nor does it follow that S
lacks some ability: there is no such thing as ability to do what
metaphysically cannot be done (though there may be such a thing as
being strong enough—God cannot sin but is strong enough to do
so). Nor does it entail that S is less than maximally strong—nothing
about one’s strength follows from one’s being all-things-considered
unable to bring about what cannot be brought about. However, a
problem with this account is that it requires being all-things-
considered able. This is a sort of ability that includes having the
opportunity to act. At times later than t, no one later than t is all-
things-considered able to bring about states of a airs at t, unless
backward causation is possible. So all-things-considered ability
varies over time, and is extrinsic. Thus it is the wrong sort of thing
to gure in an account of omnipotence. As a second pass, one might
suggest

At t, S’s range of action is maximal = df. at t, (P)(◊(something(s) at

some time(s) or timelessly cause(s) it to be the case that P) ⊃ (S is
intrinsically such as to bring it about that P)).
Omnipotence is a matter of S’s intrinsic power. If S has intrinsically
a large enough range of action to count as omnipotent, then if at
some time S cannot bring about some state of a airs that at other
times someone could bring about, this is due to a lack of
opportunity which does not entail a lack of (intrinsic) power.

The second-pass condition won’t do either. It is at least possible

that some agents are incompatibilistically free. Someone can bring it
about that someone other than an omnipotent being does an
incompatibilistically free action—that free agent brings it about. But
an omnipotent being can’t bring this about. No one else can causally
determine me to do an act I do with incompatibilist freedom. This
would amount to bringing it about that someone is causally
determined to do an act he is not causally determined to do. If it
would imply a contradiction to causally determine me to do an act
with incompatibilist freedom, an omnipotent being can’t do this—
but it is an acceptable limitation on an omnipotent being’s power
not to be able to do it (if the principle of non-contradiction itself is
an acceptable limit). My second-pass range condition is in terms of
states of a airs someone can bring about. So I introduce a
quali cation for this in my nal account of omnipotence.

We are not yet nished, though. Even a range clause and a

strength clause, I now suggest, do not su ce for a full account of
omnipotence.

MODALITY AND GOD’S POWER

Intuitively, if God is omnipotent, his power is unlimited.
Philosophers add at once, though, that God can do only the
absolutely possible, i.e. bring about or help establish only absolutely
possible facts.68 This is well motivated. Were God able to bring
about the impossible, it would not be impossible. It would be
something God possibly brings about. Yet when philosophers
introduce such claims to (say) rst-year university students, the
frosh tend to reply, ‘doesn’t that “only” limit God’s power? If God is
omnipotent, how can there be things he can’t do?’

Some try to meet the students’ concern by arguing that there are
no ‘things God can’t do’, because if a state of a airs is absolutely
impossible, there is no act or task of bringing it about. Actions or
tasks (they say) are by de nition things someone can do. If a state of
a airs is absolutely impossible, nobody can bring it about. So there
is no act or task of doing this69—and so, the argument concludes, no
act or task God cannot do. But students tend not to feel comfortable
with this. And perhaps there is something to their feeling. For one
thing, it’s now more usual to de ne omnipotence in terms of the
range of states of a airs God can bring about.70 Discomfort with
acts or tasks God can’t do may well transmit itself to talk of states of
a airs he can’t bring about, and there is no way to deal with this
analogous to the act/task move. Moreover, perhaps the modal part
of our description of an act or task (‘can do’) smuggles something in
illicitly. We have the sentence

(RS) It is not the case that something is both round and square.
(RS) says something true. The contradictory of a truth is not
nonsense but a falsehood. So (RS)’s contradictory,

(¬RS) something is both round and square,

says something false—and so it says something. It is not nonsense.

We understand what it says, though not in the fullest possible way.
(We can’t picture what would make it true.) If we do, there is
something it says—a proposition it expresses, whatever our theory
of propositions be. (¬RS) expresses a proposition. If (¬RS) expresses
a proposition, so does

(¬RS*) something is round and square and made by someone.

For we understand this in whatever way we understand (¬RS).

(¬RS*) describes the result of performing a task. In whatever sense
we can conceive of this, we can conceive of the task being
performed. So prima facie,

(SM) someone makes it the case that (¬RS)

expresses a proposition, one closely related to (¬RS*). If (SM)

expresses a proposition, it describes the doing of an act or task, if
not one whose results we can picture, or one someone possibly does.
So to the extent that we understand (SM), we may be conceiving (in
some sense) an act or task God cannot do. To allow for this, we
could just say that an act (task) is something an act-
(task-)description would describe, and an act- (task-)description is a
sentence-frame whose main verb is a verb of agency, which yields a
comprehensible sentence if one lls its gap with a term for an agent
(e.g. ‘—is feeding a bird’). ‘Would describe’ brings merely possible
actions under this rubric, and the notion of an action can be
understood without a modal element: an act is something of a sort
to be done, or such as to be done, and so agency-verbs express
things of a sort for someone to do, etc.71 If we say only this, the
bare form of an act- (task-)description does not rule out impossible
actions (tasks). ‘—is making a round square’ yields comprehensible
sentences given the right terms in its gap, though not possibly true
sentences, or sentences whose truth we can picture, or sentences we
can understand in the fullest way. So it’s not as if there is no
alternative to saying that an act (task) is a thing someone can do. As
there is an alternative, one can’t just suppose without further ado
that the notion of an action (task) must have a modal element. And
so it is worth asking why the notion of an action (task) must have a
modal element. Might there be something question-begging in that?

More basically, though, the idea of acts an omnipotent being

can’t do may not be what disturbs the frosh; many are after all
content to say both that God is omnipotent and that God cannot do
evil. The act/task move supposes that some states of a airs are
metaphysically impossible even if an omnipotent being exists. This
may well be precisely what bothers the students. They may wonder
how anything can truly be impossible if an omnipotent being exists.
If we are charitable, we will not take them to be supposing that
nothing really is impossible, or that no omnipotent being can exist if
anything is impossible: rather, they wonder how the two t
together, how an unlimited power can face limits set by the bounds
of metaphysical possibility.

One source of the feeling that it limits God’s power if he can do

only what is absolutely possible, I think, may be that talk of
absolute possibility has no obvious connection with God. The limits
of the possible seem wholly independent of God. So if they are the
boundaries of his power, these boundaries seem imposed on him
from without. So it seems that what God can and cannot do is
determined from without—and this, I submit, may be the real root
of the frosh unease.72 For what determines this from without places
an extrinsic limit on God’s power: it is something outside him to
which even an omnipotent power must bow. This does not seem to
comport with omnipotence. The Frosh Intuition, then, is that

(FI) (x)(if x is omnipotent, nothing independent of x determines

what x can do).

The limits of absolute possibility are in a sense the weakest

constraint an omnipotent being could operate under: anything else
(having to respect natural law, say) would seem even less
appropriate. So (FI) denies not just this sort of constraint but any
other. Again, it would not be appropriate to omnipotence if
something independent of an omnipotent being, while not
determining what it could do, made some of its alternatives a good
deal harder than others. For an omnipotent being, nothing should be
at all hard. If this is intuitive, then intuitively
 (FI*) (x)(if x is omnipotent, nothing independent of x to any
degree constrains x’s actions).

One way to respect (FI) is to adopt the claim that

(G) (x)(if x is omnipotent, every possible state of a airs is

possible and every impossible state of a airs is impossible
due to x, x’s being in some intrinsic state or items x brings
to be).

On (G), what states of a airs are possible is not independent of an

omnipotent God. Instead, God as it were stretches out the realm of
possibility, and its having a certain extent is just his stretching it so
far and no further. The extent of the possible just expresses God’s
own nature, power, or activity. So if God can bring about only the
absolutely possible, this does not fence God in from without. Given
(G), God can do only what is possible because God can do just what
he makes it possible that he do. It is not the possible that places
limits on God, but God who delimits the possible.

(G) gives us (I claim) the best account of the relation between

God’s omnipotence and absolute possibility. We have two intuitions,
(FI) and that God can do only what is absolutely possible. These
seem to clash. (G) lets them coexist harmoniously. For God’s nature,
power, and action are not matters independent of God. If only these
bound God’s power, God meets (FI)’s condition on omnipotence.
Thus we can preserve both intuitions if we adopt (G). The harder it
is to nd other ways to do so, the more of these two intuitions’ force
(G) inherits.73
 Now some could object that far from helping an account of
omnipotence, (G) trivializes omnipotence or could expel it from
God’s nature altogether. Suppose that God is omnipotent. Suppose
(G) too, and that it is in God to have the entire range of the possible
consist of (say) his existing, being as he is, setting his power, and
there being one apple. Then God would count as omnipotent even if
all he could do is set the range of his power and make an apple. But
this trivializes omnipotence: an omnipotent being must be able to do
more. Suppose on the other hand that omnipotence precludes a
range of power this limited, but on (G) it is in God to have the range
of the possible come out this small. Then (G) entails that it is in God
to bring it about that he is not omnipotent. If so, being omnipotent
is not part of his nature. Further, it’s in him to bring it about either
that omnipotence is not part of deity or that he is not divine. As it’s
not clear that properties can have inessential parts—that deity could
contain omnipotence in some possible worlds but not all—it seems
to follow that it is in him not to be divine.

The shortest answer here is that (G) says nothing at all about
whether it is in God to have the range of the possible other than it
actually is. So (G) carries none of the suggested consequences.
Further, if (G) is true and it is in God to make the range of the
possible other than it actually is, it does not follow that it is in God
to make the range of the possible so small as to threaten his
omnipotence.74 Even if it is in God to have the range of the possible
be other than it is, his nature could constrain the way he could vary
its range. Further, if there is a problem here, it arises as we consider
restricted ranges for the possible, whether or not God is the reason
for them. If the real problem is making God’s power hostage to the
range of the possible, one could be pardoned for seeing (G) precisely
as an antidote to this. Thus I take (G) as one conjunct in my nal
account of omnipotence.

AN ACCOUNT OF OMNIPOTENCE

At this point I am ready to hazard an account. I propose

x is omnipotent at t if and only if (i) it is not the case at t that there

is some state of a airs x is unable to bring about at least partially
due to lack of power, (ii) all truthmakers of modal truths are either
x, x’s being in some intrinsic state or items x brings to be,75 (iii) at t,
(P)(◊(something(s) at some time(s) or timelessly cause(s) it to be the
case that P and P is not the doing of an action with incompatibilist
freedom by someone distinct from x) ⊃ (x is intrinsically such as to
strongly actualize it that P)).

All three clauses are necessary. As I’ve argued, de nitions purely in

terms of strength or range are inadequate just because they are
such, and (ii) is needed for the best account of an omnipotent
being’s relation to modal truth. This de nition can’t be called too
weak to capture an ‘intuitive’ sense of omnipotence (supposing for
the nonce that we have one). (i) guarantees that an omnipotent
being has all power—that its strength is without defect. If (i) is true,
there is no state of a airs, even an impossible one, such that x
cannot bring it about even partly because x is not strong enough.
This implies that an omnipotent being has enough raw power to do
even what is in fact absolutely impossible. If states of a airs are
impossible, it does not wind up possibly bringing them about—but
per (ii), the reason for this will be some constraint either entirely
internal to itself, rooted in its nature, or rooted in its action. A
greater degree of power over the states of a airs there are is
inconceivable.76 (ii) secures the de nition against the most general
external constraint problem. (iii) is the late-medieval consensus
account, with a needed quali cation. The quali cation is consistent
with (i): it is not due to lack of power in anyone that nothing but its
agent can cause an incompatibilistically free action.

NOTES

1. Hill often links power and strength, as at 2005: 149, but he’s
not consistent: ibid. 134.

2. Later we see that things are not quite this straightforward.

3. Cyril (1994: 48), Catechetical Lecture 8. 3.

4. Origen (1994: 289), De Principiis 2. 9. 1. I owe the reference to

Mark Edwards.

5. Ibid. 249–50, De Princ. 1. 2. 10. I owe the reference to Mark

Edwards.
 6. De Symbolo 2, my translation.

7. Swinburne (1994: 129). This account intends to allow there to

be multiple omnipotent beings, with restricted ranges of
choice.

8. Augustine, Sermon 213, PL 38, 1061. My translation.

9. Augustine (1950: 156–7) 5. 10.

10. And Pike’s (1969: 210) suggestion that Anselm is speaking of

mere ‘moral weakness’ is (I submit) just o the mark.

11. I Sent. d. 42 c. 3.

12. Ibid. d. 42 c. 3, my translation.

13. William of Auxerre (1986), 4. 215.

14. William likely has Abelard’s denial of this primarily in view.

15. ST Ia q. 25 a. 3; SCG lib. 2d. 25. For full discussion see my

‘Aquinas on Omnipotence’, in Stump and Davies (eds.)
(forthcoming).

16. e.g. ST Ia q. 25 a. 4.

17. QD de Potentia 1. 6.

18. So e.g. Aquinas, ST Ia q. 8.

19. More precisely, this is su cient for having the all-things-

considered power, power in the fullest sense. There are lesser
degrees or sorts of power. A baby has the power to walk
before it has learned, since it can learn, but this is not all-
things-considered power.
20. I owe the ‘can’t’ part of this disjunctive approach to Davis
1983: 81. Note incidentally that taking on a body isn’t
something we can do—since it would require that we rst
exist bodiless, and we do not. Nor is it even something that
can happen to us: we are not rst spirits who then fall into
bodies despite ourselves; if we and our bodies come to be at
once, we do not take on bodies, but rather simply come to be
in them; if rst our bodies exist and then we do, we do not
take our bodies on, but rather they acquire us.

21. See Duns Scotus 1950-, xvii. 524–5, Lectura I d. 42 q. 1 n. 11;

ibid. ii. 343–4, Ordinatio, d. 42 q. 1 n. 9.

22. See e.g. ST Ia q. 105.

23. no. 6 q. 1. See also Ordinatio I d. 20 q. 1, in Opera Theologiae,

iv. 36.

24. no. 4 q. 22, Opera Theologiae, ix. 404.

25. Thus when Patrick Grim writes that the ‘genuinely traditional
and unlimited notion of omnipotence’ is that of being able to
perform any task speci able without contradiction and that
‘the task of defending’ this ‘seems to have been abandoned’
(2007: 204, 200, 201), he makes at least two mistakes. If
anything quali es as a traditional concept of omnipotence, it is
not what Grim says. I’m tempted to say that Grim’s concept
does not go back further than twentieth-century misreadings
of Aquinas. And the task of defending the genuine traditional
 account, or an only slightly cleaned-up version of it, is still
ongoing.

26. Wierenga (1989: 25). I’ve rearranged this a bit.

27. Ibid.19–21, 27.

28. Discussions of freedom and foreknowledge frequently try to

make out a distinction between ‘hard’ and ‘soft’ facts. (See e.g.
many pieces collected in Fischer 1989.) Wierenga might mean
that initial segments up to t consist of all ‘hard facts’ up to t—
but he doesn’t say so.

29. Wierenga (1989: 25). I’ve rearranged this a bit.

30. Tom Flint pointed out the third reading, and the problem that
ensues on it.

31. Mackie (in Urban and Walton 1978: 77–8) raised the issue
McEar raises without making up an example involving a
particular individual. Plantinga (1967: 170) introduced McEar
without naming him. LaCroix (1977: 187) added the name.
Mavrodes (1977: 280) pointed out the importance of McEar’s
having his limits essentially. Flint and Freddoso (1983: 110 n.
4) discovered McEar’s medieval ancestry.

32. Wierenga 1989: 29.

33. For the ‘part of’ part of this see also Hill (2005: 136 n. 19).

34. Wierenga (1989: 29).

35. Swinburne (1998: 251 n. 17).

36. On the de nition I o er below, if there are things other than
oneself, it would not be possible to be made omnipotent unless
one had already been the creator of everything other than
oneself: no being who had not been creator could satisfy (ii).
But those failing this condition would not for any such reason
be disquali ed from satisfying (i) and (iii), and this would be
enough to count as omnipotent in many eyes.

37. Or else the power and its base are really just two aspects of a
single attribute. As this quali cation does not a ect the rest of
my argument, I henceforth ignore it.

38. Flint and Freddoso (1983: 99).

39. Ibid.96.

40. Ibid.97.

41. Ho man and Rosenkrantz (1988: 297 ).

42. Oppy (2005: 72–3).

43. If S brings it about that p, and it is the case that p . q, but S did
not bring it about that q, q has no cause, a full cause other
than S, a partial cause not including S, or a partial cause
including S. If no cause, one can infer that S brings it about
that p . q only via something relevantly like (HR). If q has a
full cause other than S, then as we’ve seen, it is not the case
that S brought it about that p . q. If q has only a partial cause
not including S, what to say will be a function of what we say
in cases where there is no cause and cases in which there is a
 full cause other than S, and so whether S brought it about that
p . q again will hinge partly on the fate of (HR) and relevantly
similar principles: if (HR) is false, and relevantly similar
principles are false, then S did not bring it about that p . q,
given that where there is a full cause other than S, S did not
bring it about that p . q. If q has only a partial cause including
S, the possibilities multiply dizzyingly:

(I) q has no components and has only a partial cause as a

whole.
(II) q has components and has only a partial cause because of
the causal stories behind those components’ obtaining.
(II-a) the causal story is that some bits of q have full causes and
other bits have only partial causes or no cause.
(II-b) the causal story is that all bits have only partial causes or
no cause.
For simplicity’s sake, let’s discuss a case where q = (r . t).
(II-a1) r has a full cause, which is S.
(II-a2) r has a full cause, which is a composite (S + u). This
alternative yields an in nity of others, since every u
may itself be a composite, and composites may be
involved because r may be conjunctive. The way to deal
with further iterations should be clear from what I say
about simpler cases.
(II-a3) r has a full cause which does not include S.
(II-a1a1) t’s partial cause is S.
(II-a1a2) t’s partial cause is a composite, (S + v).
(II-a1a3) t’s partial cause does not include S.
(II-a1a4) t has no cause.
(II-a2a1) t’s partial cause is S.
(II-a2a2) t’s partial cause is a composite, (S + v).
(II-a2a3) t’s partial cause does not include S.
(II-a2a4) t has no cause.
(II-a3a1) t’s partial cause is S.
(II-a3a2) t’s partial cause is a composite, (S + v).
(II-a3a3) t’s partial cause does not include S.
(II-a3a4) t has no cause.

Obviously alternatives multiply similarly under the

assumptions that r has no cause or only a partial cause which
does or doesn’t include S, but the ways to deal with these will
be displayed in discussing the other alternatives, and so there
will be no need (thankfully) to take them up.

As to (I), if q has no components and only a partial cause,

S, then the overall scenario is that S fully causes p, S partially
causes q and nothing else contributes causally to (p . q). My
intuition here is that S counts as bringing about (p . q), on
analogy to the case where S fully causes it that p and fully
 causes it that q. Under (II), there are these terminal
possibilities to consider:

(II-a1a1) S fully causes p, q = (r . t). S fully causes r and

partially causes t; nothing else causally contributes to t.
Here it seems clear that S brings it about that (p . q), on
analogy to the case where S fully causes it that p and
fully causes it that q.
(II-a1a2) S fully causes p, q = (r . t). S fully causes r. t’s partial
cause is a composite, S + u. There is not just one sort of
case here, because partial causes can divide up causal
responsibility in many ways. If S is mostly (i.e. way
more than half) responsible for t, the case is enough like
one in which S fully causes both conjuncts to warrant
the claim that S brings it about that (p . q). If most
responsibility lies with u, one would need something
like (HR), and so can’t conclude that S brings it about
that (p . q). And there will be a gray zone in which
intuition just is not clear. In the gray zone, though, it
will at least not clearly be warranted to say that S brings
it about that (p . q).
(II-a1a3) S fully causes p, q = (r . t). S fully causes r. t has a
partial cause which does not include S. This seems to
parallel the case in which S causes p and something else
causes q. So my intuition is that in this case S does not
bring it about that (p . q).
 (II-a1a4) S fully causes p, q = (r . t). S fully causes r, t has no
cause. One could have it that S brings it about that (p .
q) in this case only via something like (HR).
(II-a2a1) S fully causes p, q = (r . t). r has a full cause, (s + u).
t has a partial cause, s. Here again all will depend on
the degree of S’s causal responsibility for r and t.
(II-a2a2) S fully causes p. q = (r . t), r has a full cause, (S + u).
t has a partial cause, S +v.

Again we face partial-cause unclarity.

(II-a2a3) S fully causes p, q = (r . t). r has a full cause, (S + u).

t has a partial cause which does not include S. One
could have it that S brings it about that (p . q) in this
case only via something like (HR).
(II-a3a1) S fully causes p, q = (r . t). r has a full cause which
does not include S. S partially causes t; nothing else
causally contributes to t. Here and in a3a2–4 it seems
plain that S does not bring it about that (p . q).

44. Flint and Freddoso (1983: 97, 94). Note 22 could be read as
expressing an intention to be neutral on this issue, but the text
does not seem neutral. If the text does let us make some CFs
true, Flint’s later writing on Molinism retracts this.

45. Ibid.95.

46. Ibid.86.
47. This also vitiates one of Oppy’s examples. Accepting the Flint-
Freddoso claim that only Jones can a ect (2)’s truth-value,
Oppy suggests that the state of a airs (everyone in the room is
free with respect to giving to Oxfam at t) > (everyone in the
room freely chooses to give to Oxfam at t) can only be brought
about jointly by everyone in the room. But one su ciently
persuasive person can bring it about. The persuader need
merely cause the emotional dispositions of those in the room
so to change that were they asked, they would give—or else
give reasons they all freely adopt.

48. Oppy 2005: 74.

49. Flint and Freddoso (1983: 85).

50. Personal correspondence.

51. Gellman (1989: 333).

52. Flint and Freddoso (1983: 86).

53. Oppy (2005: 75–6) makes a closely related point,

independently.

54. If there were some non-Whiteheadian possible worlds, they too

would count for (FF)’s purposes, and the result might well be
that Whitehead’s God was not omnipotent. But
Whiteheadianism is hardly the sort of thing to come out
contingently true, so assuming even one Whitehead-world is
tantamount to assuming for illustration’s sake that all worlds
are Whiteheadian. If we assume this, it is not germane to point
 out that Whitehead’s God would have the maximal
Whiteheadian power in any possible world. My point is rather
that we, from our non-Whiteheadian vantage, see this as not
power enough.

55. One might ask here why just one has this ability if both satisfy
(FF). One answer might be that Jones strongly likes one of
them and strongly dislikes another. Of course both, being
omnipotent, might easily e ace Jones’s likes and dislikes, but
this would be relevantly like brainwashing and so would
defeat the claims that Jones freely accepts reasons given and
that the case is purely one of persuasion.

56. This does not disqualify the persuading-Jones example just

given against (FF), because Flint and Freddoso accept weak
actualization as partly constitutive of omnipotence and see
persuasion as a sort of weak actualization.

57. If there are true CFs, and an omnipotent being does not wholly
control their truth-value, these are contingent and can impede
its will too. I suggest that the right conclusion here is that a
Molinist God cannot be omnipotent.

58. This is essentially what Oppy (2005: 76) calls a ‘tempting

objection’ to (FF), framed so as to get around his problem with
it, but I hit on it independently.

59. I owe the thought of a series of godlets to Wielenberg (2000:

36–7).
60. ST Ia q. 25 a. 3.

61. Flint and Freddoso (1983: 87–9); Wierenga (1989: 16–17, 25–
6).

62. I owe this suggestion to an audience-member at the British

Society for Philosophy of Religion.

63. Here I’m indebted to a related suggestion by Tom Flint, in

correspondence.

64. Wielenberg (2000: 42).

65. Ibid.37–8, 39–40, 43–4.

66. There is also thus a puzzle in Wielenberg’s use of his ‘series of

deities’ objection against Wierenga and Flint-Freddoso. The
objection turns on the intuition that if a deity is less powerful
than something else, it is not omnipotent (surely correct), but
takes it as su cient for being less powerful that the deity have
a more restricted range of action than the other thing (ibid.
30, 36–7). On Wielenberg’s own showing, this is not su cient.

67. There seems intuitively to be a di erence between the modal

status of the three. Is Hugo able to lift 300lbs? Intuition
replies: not if he is not strong enough. Perhaps he is able to be
able, since he can develop the strength over time, but he is not
able now: able to be able doesn’t entail able. Again, not if he
does not know how to lift. If he is able to learn to lift, he is
still no more than able to be able to lift. But if he has strength
and knowledge enough to lift it, we do not require in addition
 that he should actually want to lift 300lbs. It seems enough
that he might want to—that he can ‘trigger’ the action simply
by having a desire. If Hugo literally cannot be made to desire
to lift 300lbs., though, is he able to lift it? If he cannot, in no
possible world does he lift 300lbs., even involuntarily.
(Perhaps his body does, acting under someone else’s control.)
In no possible world is an act of lifting ascribable to him: one
requisite for a desire-belief action-explanation is not there, and
so we can’t treat his body-movements as an action. So it seems
to me that if Hugo can’t want to lift 300lbs., he can’t lift it,
though perhaps his body can.

68. ‘Help’ allows for facts God cannot establish on his own: for it to
be a fact that I feed some sh, I must do the feeding.

69. So Swinburne (1993: 153).

70. So e.g. Ho man and Rosenkrantz (1988); Flint and Freddoso

(1983); Wierenga (1989).

71. Some might say they can’t understand this save by tacitly
introducing the modal ‘can do’ again. If this is your plight, we
might approach this in terms of conditional modalities. All
possible tasks are tasks someone could do just if (RS) were
true. Impossible tasks are those someone could do just if (RS)
were false.

72. At any rate, it’s the most cogent root I can ascribe to it. Worries
based on biblical texts such as ‘with God all things are
 possible’ (Matt. 19: 26) seem to me misplaced. For nothing in
such texts suggests a particular reading of ‘all things’.

73. (FI) is not the only intuition we have in this vicinity. We might
also think that

(O) if God is omnipotent, his power is unlimited, and

(U) if God’s power is unlimited, he does not limit it himself.

(O), (U), and (G) are co-tenable, for it is co-tenable with (G)
that it not be up to God what states of a airs there are and
that God have placed every state of a airs in his power’s
range. Had he done so, he would have set no limits to his
power, and everything would be possible. Con ict emerges if
we conjoin (O), (U), (G), and the claim that some states of
a airs are impossible. But in this con ict, if there are any
impossibilities, (O) or (U) should lose. If we reject (G), God’s
power is limited from without by facts of impossibility—and
intuitively limits imposed from without, over which one has
no control, are worse than self-imposed limits whose extent
one in some way controls. One might object that there being a
stock of states of a airs whose content is not up to God is itself
a non-self-imposed limit. But if this is right, then equally any
stock whose content God set would constitute a self-imposed
limit on his power. As one or another must be the case, it
would seem to follow that any extent of divine power involves
some limitation—in which case either (O) is false or
 omnipotence is impossible. I incline either to reject (O) or to
question the objection.

74. If it were necessary (due to God’s nature and that of

omnipotence, say) that the possible have a range of a certain
size, it would not follow that any particular state of a airs had
to be included in this. It could still be the case that any
candidate state of a airs (perhaps any candidate meeting some
conditions, e.g. not being an explicit contradiction) could wind
up possible.

75. My talk of truthmakers is meant in a very weak sense. There

are truthmakers of some sort on any non-de ationary theory of
truth—be they corresponding facts, cohering bodies of
propositions, convention-setting decisions, or what have you.
Equally, there are truthmaking relations on each such theory—
correspondence, coherence, etc. What I mean by saying that
truths have truthmakers is that for each truth, there is at least
one truthmaking relation which links it to the sort of
truthmaker allotted it in some non-de ationary theory of
truth. In e ect, then, my talk of truthmakers commits me to no
more than some disjunction of non-de ationary theories of
truth being true. My truthmaking relations are those speci ed
in theories of truth, rather than being any sort of entailment.
This matters. Any fact entails that 2 + 2 = 4. But not every
fact is linked to this truth by all truthmaking relations. So I am
 not committed to the claim that every fact (etc.) whatever
truthmakes every necessary truth.

76. One could hold that a greater degree of power still would
consist in being able to add to the stock of states of a airs
available to bear modal status. I would happily redo the
de nition so as to include this note.

REFERENCES

AQUINAS, THOMAS, QD de Potentia.

——— Summa Theologiae.

——— Summa Contra Gentiles.

AUGUSTINE, De Symbolo.

——— (1950). The City of God, trans. Marcus Dods. New York:
Random House.

——— Sermon 213, PL 38.

CYRIL OF JERUSALEM (1994). Catechetical Lecture 8, trans. E. Gi ord, in

P. Scha and H. Wace (eds.), Nicene and Ante-Nicene Fathers, 2nd
ser. Peabody: Hendrickson, v. 7.

DAVIS, STEPHEN (1983). Logic and the Nature of God. Grand Rapids:
Eerdmans.

DUNS SCOTUS, JOHN (1950-). Lectura, in Opera Omnia, ed. C. Balic et

al. Vatican City: Typis Polyglottis Vaticana, xvii.
——— (1950–). Ordinatio, in Opera Omnia, ed. C. Balic et al. Vatican
City: Typis Polyglottis Vaticana, ii.

FISCHER, JOHN M. (ed.) (1989). God, Foreknowledge and Freedom.

Stanford: Stanford University Press.

FLINT, THOMAS, and FREDDOSO, ALFRED (1983). ‘Maximal Power’, in

Alfred Freddoso (ed.), The Existence and Nature of God. Notre
Dame: University of Notre Dame Press, 81–113.

GELLMAN, JEROME (1989). ‘The Limits of Maximal Power’,

Philosophical Studies 55: 329–36.

GRIM, PATRICK (2007). ‘Impossibility Arguments’, in Michael Martin

(ed.), The Cambridge Companion to Atheism. Cambridge:
Cambridge University Press, 199–214.

HILL, DANIEL (2005). Divinity and Maximal Greatness. London:

Routledge.

HOFFMAN, JOSHUA, and ROSENKRANTZ, GARY (1988). ‘Omnipotence

Redux’, Philosophy and Phenomenological Research 49: 283–301.

LACROIX, RICHARD (1977). ‘The Impossibility of De ning

“Omnipotence” Philosophical Studies 32: 181–90.

MAVRODES, GEORGE (1977). ‘De ning Omnipotence’, Philosophical

Studies 32:191–202.

OPPY, GRAHAM (2005). ‘Omnipotence’, Philosophy and

Phenomenological Research 71: 58–84.
ORIGEN (1994). De Principiis, trans. F. Crombie, in P. Scha and H.
Wace (eds.), Nicene and Ante-Nicene Fathers, 2nd series. Peabody:
Hendrickson, v. 7.

PIKE, NELSON (1969). ‘Omnipotence and God’s Ability to Sin’,

American Philosophical Quarterly 6: 208–16.

PLANTINGA, ALVIN (1967). God and Other Minds. Ithaca: Cornell

University Press.

STUMP, E., and DAVIES, B. (eds.) (forthcoming). The Oxford Handbook

of Aquinas. Oxford: Oxford University Press.

SWINBURNE, RICHARD (1994). The Christian God. Oxford: Oxford

University Press.

——— (1993). The Coherence of Theism, rev. edn. Oxford: Oxford

University Press.

——— (1968). Space and Time. London: Macmillan and Co., Ltd.

——— (1998). Providence and the Problem of Evil. Oxford: Oxford

University Press.

URBAN, LINWOOD, and WALTON, DOUGLAS (eds.) (1978). The Power of

God. Oxford: Oxford University Press.

WIELENBERG, ERIK J. (2000). ‘Omnipotence Again’, Faith and

Philosophy 17: 26–47.

WIERENGA, EDWARD (1989). The Nature of God. Ithaca: Cornell

University Press.
WILLIAM OF AUXERRE (1986). Summa Aurea, ed. J. Ribailler. Paris:
CNRS.

WILLIAM OF OCKHAM, Ordinatio, Opera Theologiae v. IV.

——— Quodlibeta, Opera Theologiae v. IX.

 CHAPTER 9

OMNIPRESENCE

HUD HUDSON

Do not I ll heaven and earth? saith the Lord.

Jeremiah 23: 24.1

I. PRELIMINARIES

ACCORDING to the tradition of western theism, God is said to enjoy

the attribute of being everywhere present. But what is it, exactly, for
God to manifest ubiquitous presence? Well, presumably, it is for
God to bear a certain relation—the ‘being present at’ relation—to
every place.

Quarreling about the alleged relata of this preliminary response

can generate a pair of dismissive answers to a rather natural follow-
up question—‘And in just what manner does this “being present at”
relation relate God to places?’

‘In no way at all,’ says the atheist, ‘for no divine being exists.’ ‘In
no way at all,’ says the relationalist, ‘for regions do not exist.’2 Let
us put on hold these dismissive answers in the ensuing discussion
and openly acknowledge a working assumption of theism and
substantivalism for the remainder of the chapter. Thus, for present
purposes, we shall count ourselves among the realists about God
and regions.3

Most of us who subscribe to some kind of region realism do so

because we countenance a substantivalist spacetime. Historically,
however, it is more accurate to construe omnipresence as a relation
between God and space, rather than as a relation between God and
spacetime. Similarly (and for obvious reasons) eternality has been
historically conceived as a relation between God and time, rather
than as a relation between God and spacetime. In the discussion to
follow I will attempt to respect this historical antecedent where I
can, asking after a variety of relations that might be thought to bind
God to spaces (on the supposition that there are such things).

In this chapter, then, I intend primarily to focus my discussion

on the ‘being present at’ relation that gures so prominently in the
divine attribute of omnipresence, on both fundamental and
derivative readings of that relation, and on a host of philosophical
problems that arise for each reading. The chapter will be divided
between a discussion of the historical positions of Anselm and
Aquinas, a note on the controversy stirred up by the modern
contributions of Hartshorne, Swinburne, Taliaferro, and Wierenga, a
brief glance at two curious and underexplored approaches, an
investigation of the promising prospects for further inquiry a orded
by recent work on the metaphysics of location, and some concluding
comments on special problems of occupation for the Christian theist.

II. TWO HISTORICAL VIEWS AND A RECENT CONTROYERSY

One would think the ‘being present at’ relation is a thoroughly

straightforward one, in current idiom—a perfectly natural and
fundamental, external relation of occupation between objects and
regions. At each moment they are present, for example, this die
occupies a roughly cubical region, that tower a roughly cylindrical
region, and the Earth a roughly spherical region. Such a simple
occupation reading of omnipresence, however, has often been
avoided for fear that it would con ict with other divine attributes.

In his Monologium and in his Proslogium, St Anselm explores

some of these con icts.4 Anselm begins in ch. 20 of the Monologium
—’[God] exists in every place at every time’—by recognizing a need
to assert some kind of presence relation or other on the grounds that
nothing can be good or even exist (not even the regions themselves)
where God is not. Thus, he endorses the claim that God exists
everywhere. But in ch. 21—‘[God] exists in no place or time’—he
also recognizes a need to deny the most familiar kind of presence
relation on the grounds that it either violates the doctrine of divine
simplicity with God’s being partly here partly there or else leads to
the impossibility of God’s being wholly present at two di erent
places, a trick that might be pulled o by a universal but not by a
non-repeatable substance. Thus, he endorses the claim that God
exists nowhere. Reconciliation in ch. 22—‘How [God] exists in
every place and time, and in none’—is achieved by noting two
di erent senses of ‘being present at a place’. The rst sense, enjoyed
by the die, tower, and planet in the example above, is just the
fundamental and familiar occupation relation. Anselm dismisses this
reading of God’s omnipresence, adding to the di culties of
simplicity and impossibility just mentioned a third problem of
unacceptably con ning the divine being by the constraints that
seem wedded to being literally located, restrictions that are
appropriate for creaturely items but not for their creator. (An
association of certain kinds of predication with unacceptable
con nement was once more common than perhaps it is today and
was responsible for much medieval mischief in motivating attempts
to show that God is extra-categorial.) The second sense, the
acceptable and derivative one, is then endorsed but largely left in
mystery.

Edward Wierenga has cast some light on this mystery.5 Taking as

clues certain passages in the Proslogium, Wierenga nds in Anselm
the double view that souls, like God, exhibit the acceptable form of
being wholly present in more places than one and that this kind of
multi-presence is wedded to sensations that are themselves related
to di erent places. Wierenga then oats the promising hypothesis
that Anselm’s positive and preferred account of God’s being at every
place and time amounts to God’s sensing or perceiving at each place
and time. To avoid the issue of embodiment, a precondition of
certain kinds of perception, Wierenga redescribes the relation as a
kind of inner sense or immediate knowledge of the goings-on at the
relevant location. Omnipresence for Anselm, then, is ultimately
reducible to a kind of knowledge, immediate and localized for every
region.

In addition to proposing an analysis of omnipresence in terms of

knowledge, Anselm has identi ed three puzzles for anyone who
wishes to champion the alternative, literal occupation account: The
problem of simplicity—how can something that is not mereologically
composite occupy more than one region? The problem of
multilocation—how can something occupy (in the ‘wholly present’
sense) two numerically distinct regions? The problem of containment
—if to occupy a region is to be contained by it, how can something
that is essentially free of the constraints that bind all creaturely
things occupy a region?

In his Summa contra Gentiles, St Thomas Aquinas o ers an

interpretation of omnipresence that analyzes divine presence by
appeal to power rather than knowledge.6 Like Anselm’s, this account
should be distinguished from the fundamental occupation relation,
and unless it is so distinguished it may well give rise to a nest of
further problems clashing with the Thomistic reading of eternity as
atemporality.7 Echoing Anselm, Aquinas argues rst, that God must
be present in all things insofar as he must sustain them in existence
(which in a sparsely populated world falls short of omnipresence,
unless the regions themselves are similarly sustained), and second,
that insofar as the presence of an incorporeal being is a function of
its power, God’s in nite power ensures that he is everywhere.8
Aquinas both clari es and supplements this emphasis on power in
the later Summa Theologiae.9 Not just one kind of power, but two are
at issue: the preserving or sustaining of a creaturely thing and an
absolute and immediate control over all such things. Thus God is in
all things, ‘giving them being, power, and operation’, which for
Aquinas necessitates a relation of direct contact. Again, though, the
kind of presence a orded through the contact of power is di erent
from that of the contact of ‘dimensive quantity’ or simple location
for corporeal things, for unless power can take the place of a
fundamental occupation relation, Aquinas seems to concede that
God’s incorporeality and the ban on the literal co-location of bodies
would both be in jeopardy. So long as God is present by contact of
power and not by contact of dimensive quantity, a necessary
condition of God’s causality is insured while the genuine problems
of co-location are kept at bay. Moreover, in addition to these two
sorts of power, a kind of accessibility makes the list which Aquinas
glosses with the scriptural metaphor, ‘all things are bare and open to
His eyes’ (Hebrews 4:13), suggesting an incorporation of the
unmediated knowledge that characterized the Anselmian proposal.
It is unclear whether the vision metaphor introduces a part of the
analysis of omnipresence or merely observes a consequence of it
with respect to knowledge, and since the primary emphasis is
clearly on power, it seems best to regard Aquinas as putting a new
interpretation on the table; whereas for Anselm omnipresence is
primarily a kind of knowledge, for Aquinas it is primarily a kind of
power.

Re ecting on this brief rehearsal of the Thomistic position can

help us add three more entries on our list of puzzles to be addressed
by anyone who wishes to champion the alternative, literal
occupation account: The problem of timelessness—how can something
occupy a region and be atemporal? The problem of incorporeality—
how can something occupy a region and fail to have a body? The
problem of colocation—how can two numerically distinct things each
occupy the same region?

As Wierenga has noted, both Anselm and Aquinas develop

accounts of omnipresence and its characteristic ‘being present at’
relation that are parasitic on our understanding of the
straightforward, non-mysterious occupation relation with which we
are all familiar. In other words, the special way in which God is
credited with being everywhere is tantamount to God’s having
knowledge or control over and sustaining in existence those items
that stand in fundamental occupation relations to places. Owing to
their derivative nature, I will henceforth refer to these historical
conceptions as ‘non-occupation accounts’ of omnipresence.10

The non-occupation accounts of omnipresence have been

in uential, and new variants have been formulated, expanded, and
advocated by recent writers in the philosophy of religion. Four
prominent contributors on this theme over the last century include
Charles Hartshorne, Richard Swinburne, Charles Taliaferro, and
Edward Wierenga.11 Despite the advantages of the historical and
derivative conceptions of presence, a version of the problem of
incorporeality resurfaces to divide these authors. Swinburne, for
example, joined the debate as a non-occupation theorist of a
Thomistic stripe who wished to illuminate the consequences of
acknowledging the sort of universal power that makes for divine
omnipresence. Following (to a degree) the lead of Hartshorne, this
investigation led him to the question of God’s embodiment.
Hartshorne had earlier endorsed the rather surprising thesis that the
immediacy that characterized God’s knowledge of and power over
everything found in the world doubled as a guarantee that the world
thus conceived is God’s body.12 Accordingly, to the extent that
omnipresence is read as knowledge and power, God’s omnipresence
determines God’s embodiment. Hartshorne’s conclusion rests on the
dubitable foundation that a mind has as its own whichever body it
both knows and controls in a non-mediated way, a premise that
seems at once too weak and too strong: too weak for it fails to
secure my exclusive relation to my body (since both God and I
presumably meet the relevant criteria) and too strong for, plausibly,
there exist parts of my body that I have no non-mediated power
over and no knowledge of at all—mediated or otherwise.
Swinburne’s discussion is more modest and more nuanced than
Hartshorne’s, but nevertheless accords to God a restricted form of
embodiment, allegedly compatible with ‘the traditional theistic view
that God has no body’.13 On Swinburne’s view, restricted
embodiment means both that God is able to move directly any
object (capable of motion) without the bene t of causal
intermediaries and that he knows directly (again without causal
intermediaries) the qualities exempli ed in any region at any time.
Thus, restricted embodiment is a consequence of a combination of
Anselmian and Thomistic themes. According to Swinburne, the
embodiment is limited insofar as God has no particular orientation
or restricted point of view on the world (as we do) and since God is
not pained by disturbances in material bodies nor a ected in
thought by the states of those objects (as we are).

But for some, limited embodiment is problem enough. Taliaferro,

for instance, vigorously resists the suggestion to see the world as
God’s body, citing the sort of immediacy essential to the non-
occupancy accounts of omnipresence as altogether di erent in kind
from the relations that we bear to our own bodies.14 Wierenga has
also taken to task the embodiment theses found in Hartshorne and
Swinburne on the grounds that whereas Aquinas stressed God’s
presence as a presence in things that stand in fundamental
occupation relations to regions, these later theorists extend the
relevant doctrines of knowledge and power to the regions
themselves, whether occupied or not. But, complains Wierenga, it
would be unmotivated to take God’s knowledge and power of and
over a region that happens to be occupied by some material object
as a reason to count that object among the parts of God’s body,
when the knowledge and power in question are indi erent to
whether the region is occupied at all. The object isn’t doing any
work, so to speak, in providing access to the region, and thus fails to
have even a diminished claim on being an instrument or body
through which God knows or can manifest power.15

On the strength of this four-way discussion between Hartshorne,

Swinburne, Taliaferro, and Wierenga, it is rather interesting to note
that one of the most compelling motivations backing non-
occupation accounts of omnipresence—namely, that they permit an
adherence to the incorporeality of God untroubled by quali cations
and partial concessions—is much less secure than it is often taken to
be.

III. A BRIEF GLANCE AT TWO UNDER-EXPLORED NON-OCCUPATION

RELATIONS

Before renewing an investigation into the prospects of reviving an

account of omnipresence featuring a fundamental (rather than
derived) occupation relation, perhaps it would be worthwhile to
note two other rather underexplored non-occupation relations that
(unlike the accounts emphasizing knowledge and power) do not as
easily raise the question of regarding the world as God’s body.
Curiously, each proposal has some genuine purchase on the notion
of being everywhere present. Both suggestions are admittedly
bizarre, but the detour through them won’t take much time and they
are fun and interesting to entertain.

The rst takes the relation between God and region to be

mereological. Not, as some pantheism might have it, with God
claiming the regions (as well as everything else) as proper parts, but
the other way around. That is to say, each region (as well as
everything else) would have God as a proper part. By occupying the
unique position of being a locus of universal overlap all the while
remaining a mereological simple, God would thus turn out to
manifest exactly the right quali cations to function as the null
individual—that elusive counterpart to the null set in set theory
whose most salient characteristic is being a proper part of anything
distinct from it. Omnipresence would thus amount to being a proper
part of each region.16

The second takes the relation between God and spacetime to be

that of numerical identity. Locations have locations, even if they
cannot be properly said to occupy or ll those locations; spacetime
itself, then, has an arguable claim on omnipresence. On this view,
our substantivalist spacetime would (despite current opinion) be
without beginning or end and a necessarily existing entity, as well
as being the subject of omnipotence, omniscience, perfect goodness,
and the lot. Spinoza’s monism (under at least one plausible reading)
presents itself as a historical candidate for such an interpretation,
and perhaps there is also something of a stoic precedent in all of
this, as well. Compare the many-sided imagery of Acts 17: 27–8:
‘God is not far from each one of us. For in him we live and move
and have our being.’ Moreover, it is worth noting that there is
somewhat less pressure to deny divine attributes of the receptacle
than to deny them of the fusion of its material contents (as is more
common in the literature on embodiment and the corporeality of
God). Finally, such an identi cation would ground one
straightforward interpretation of the immanence of God and would
e ectively sidestep the increasingly common complaint of the
incoherence of taking God to be outside spacetime (but at a
considerable price!). Omnipresence would thus amount to being
everywhere not in the sense of occupying a region but in the sense
of being identical to the most inclusive region.17

IV. AN OLD-TIME REYIYAL

What of reviving the literal occupation account of omnipresence?

Again, insofar as the derivative readings pro ered by Anselm and
Aquinas are parasitic on this fundamental relation, it is not as if a
proponent of one of those historical camps or of their descendants
lacks the conceptual resources to entertain the view. Moreover,
many of the advantages of the historical views need not be lost, and
they may be improved upon. Not only may omnipresence be
signi cantly related to omnipotence and omniscience, literal
occupation of every place may go some considerable distance
toward resolving some of those perplexing ‘how-is-that-possible’
questions that threaten such incredible claims of direct and
unmediated knowledge and power. To put it another way: rather
than having knowledge and power as components of its analysis,
omnipresence may receive an explanation quite independent of
omniscience and omnipotence, but then be available as a tool to
explore the possibility (and mechanics) of these other divine
attributes.

Still, we must not forget our list of puzzles, which have certainly
discouraged any appeal to the perfectly natural and fundamental,
external relation of occupation as the key to omnipresence. A
reminder:

The problem of simplicity—how can something that is not

mereologically composite occupy more than one region?
The problem of multilocation—how can something occupy (in the
‘wholly present’ sense) two numerically distinct regions?
The problem of containment—if to occupy a region is to be contained
by it, how can something that is essentially free of the constraints
that bind all creaturely things occupy a region?
The problem of timelessness—how can something occupy a region and
be atemporal?
The problem of incorporeality—how can something occupy a region
and fail to have a body?
The problem of co-location—how can two numerically distinct things
each occupy the same region?

Fortunately, recent work in contemporary analytic metaphysics

o ers some hope of intelligibly addressing these worries and of
reestablishing the fundamental occupation relation as one of the
candidates for providing the proper analysis of omnipresence.18
 V. OCCUPATION RELATIONS

Ordinary objects stand in a variety of location relations. In this

section, I propose to explore di erent candidate-descriptions of
these kinds of occupation that have emerged in the recent literature,
descriptions that presuppose a primitive and fundamental
occupation relation. Although occasioned more or less exclusively
by work on the metaphysics of material objects, the attention that
has recently been devoted to the metaphysics of location relations
may provide some genuinely valuable insights into the nature of
omnipresence.19

There are many ways into our topic. Let us begin by posing a
pair of questions about locations:

(Q1) When an object, x, is located at a non-point-sized region, r,

is x thereby located at each of the subregions of r, as well?

(Q2) When an object, x, is located at each of two regions, r and r

*, is x thereby located at the fusion of r and r *, as well?

A rmative answers to (Q1) can be rooted in very di erent kinds of

theory. To get a sense of the debate in question it will help to have
some machinery before us. What follows is certainly not exhaustive
but is nevertheless representative of some of the ways relations
between objects and regions have been recently conceived. Consider
the following ve de nitions deriving from work by Josh Parsons.20
‘x is entirely located at r’ =df x is located at r and there is no region
of spacetime disjoint from r at which x is located.
‘x is wholly located at r’ =df x is located at r and there is no proper
part of x not located at r.
‘x is partly located at r’ =df x has a proper part entirely located at r.
‘x pertends’ =df x is an object that is entirely located at a non-point-
sized region, r, and for each proper subregion of r, r *, x has a
proper part entirely located at r*.21
‘x entends =df x is an object that is wholly and entirely located at a
non-point-sized region, r, and for each proper subregion of r, r *, x
is wholly located at r*.22

Take someone who thinks that all non-point-sized objects are

composite and pertend. Given the de nitions of ‘partly located’ and
‘pertending’ this theorist holds that strictly speaking the answer to
(Q1) is negative; an object always has exactly one location. But he
can explain why we might tend to regard it as a rmative, since
when an object, x, is located at a non-point-sized region, r, x is
partly located at each of the subregions of r as well (even if x is
neither wholly nor entirely located at those regions). The pertension
theorist, however, should be careful to add that ‘being partly located
at region r does not entail ‘being located at region r—for despite
what is suggested by its name, being partly located is not a species
of location.
 A theorist who would unquali edly answer (Q1) in the
a rmative is one who thinks that some non-point-sized object’s are
composite and pertend while others are simple and entend. What
one may have thought was exclusively an a priori battle eld has
recently been an arena in which a posteriori arguments from
contemporary physics have provided unexpected support favoring
recognition of some entending objects.23 One thing that emerges,
then, is that these two theorists disagree about whether an object’s
occupying a non-point-sized region guarantees that it sports proper
parts, with our entension theorist leaving open the exotic possibility
of extended simples.

On the other hand, a kind of theorist who would answer (Q1)

clearly in the negative is one who thinks that some non-point-sized
objects are spanners.

‘x spans’ = df x is an object that is wholly and entirely located at

exactly one non-point-sized region, r, and there is no proper
subregion of r, r *, such that any part of x is located at r *.24

Although the proponents of spanning object’s accept (as do the

friends of entension) the possibility of non-point-sized mereological
simples, they deny (against both the pertension and entension
theorists) that an object’s occupying a non-point-sized region
guarantees that it either occupies or partly occupies each of that
region’s proper subregions. Spanners do not enjoy any variety of
multiple location.

Recall our second question:

(Q2) When an object, x, is located at each of two regions, r and r

*, is x thereby located at the fusion of r and r*, as well?

A rmative answers to (Q2) might initially seem automatic and

inescapable. One who thinks that all objects pertend is likely to
imagine cases in which (Q2)’s corresponding conditional is
vacuously satis ed on the grounds that pertending objects are never
located at more than one region (at best being partly located at
more than one region). One who thinks that some objects entend
may consider cases in which the non-point-sized simple itself makes
both the antecedent and consequent of (Q2)’s corresponding
conditional true. Finally, one who thinks that some objects span will
not thereby see any threat to an a rmative answer, for a spanner
also satis es (Q2)’s corresponding conditional vacuously.

Nevertheless, I think that the same sorts of consideration that

lead some to take entension seriously may also lead to uncovering a
neglected notion of occupation and to a negative answer to (Q2). If
we begin thinking of ‘being located at’ as a one-one relation, we are
left with a choice between pertension and spanning for non-point-
sized objects (and with some minor explaining to do involving
partial-occupying, if we opt for pertension). Indeed, thinking that
location is one–one and accepting the possibility of non-point-sized
simples would be one straightforward motivation for accepting the
possibility of spanners. But ‘being located at’ is a perfectly natural
external relation, and without some argument to the contrary,
perhaps one should take as a default position that a single object
can bear this relation to more than one region. Accordingly, the
possibility of entension appears to gain some plausibility.25 But once
one is willing to grant that ‘being located at’ can hold in a one-many
pattern, one should not restrict that pattern without good reason,
and entension embodies a restriction.

According to our account of entension above, when an object

entends it is wholly located at each of the regions where it is located
at all. But if we are willing to claim that occupation is a one-many
relation, we might brie y consider a maximally liberal proposal
according to which any set of regions is such that there could be a
single object that occupies all and only the members of that set.
Consider, for example, an object that bears this perfectly natural
external relation of occupation to a cubical region, S, and also to
another cubical region, S* (where’s and S* do not overlap), and to
no other regions. Moreover, let us add that our object is a simple
and thus fails to be partly located at any region. By hypothesis, this
object is neither located nor partly located at proper subregions of s
and S* and it also fails to be located at the fusion of s and S*. Such
an object would ensure a negative answer to (Q2).

Objects of this kind would nevertheless enjoy multiple location.

The simples of this species would in one respect be like entending
objects (since they would be wholly located at more than one
region) but could in another respect be like spanners (for they could
in fact be located at a non-point-sized region without also having
themselves or their parts located at any of its proper subregions).
The composites of this species would in one respect be like
entending objects (since they would be located at more than one
region) but could in another respect be like pertending objects
(since they could be partly located in some regions). Let us then add
one nal de nition.

‘x multiply locates’ =df (i) x is an object that is located at more than

one region, and (ii) x is not located at the fusion of the regions at
which x is located.

Entending, spanning, and multiply located objects would be (or in

the case of multiple locaters would occasionally be) non-point-sized
or even extended mereological simples. There are some long-
standing friends of the possibility (and perhaps of the actuality) of
extended simples, and the number of their supporters is ever
increasing. Philosophers who endorse the possibility of extended
simples (on non-theistic grounds) include John Bigelow, Ned
Markosian, Fraser MacBride, Kris McDaniel, Josh Parsons, Ted
Sider, and Peter Simons.26 On the working assumption, then, that
we have a number of conceptual possibilities alive in the
contemporary literature now clearly displayed, how can we make
use of any of this machinery to revive a literal occupation account
of omnipresence?
 VI. OCCUPATION ACCOUNTS OF OMNIPRESENCE AND OUR SIX PUZZLES

Pertension, entension, spanning, and multiple location provide us

with four clearly di erent ways to conceive of God’s relation to
regions.

Recall our rst puzzle: the problem of simplicity—how can

something that is not mereologically composite occupy more than
one region? Assuming divine simplicity, this is exactly the sort of
question that historically has proved worrisome for a literal
occupation account of omnipresence, for pertension is the natural
default understanding of literal occupation. So reasoned Anselm, for
example, as he took the occupation of an extended region to require
proper parts. Let us soften Anselm’s rejection, though, to disqualify
pertension only and leave untouched the other three conceptions
that accommodate simplicity.

Spanning, like pertension, is a one-one relation, and although it

does not require proper parts, this conception would require a
unique region at which God would be located, all others (whether
sub-or-super-regions of the privileged region) being thereby
rendered ineligible to host the divine presence. On the grounds that
this conception is too restrictive, let us also disqualify a spanning
conception of omnipresence.

Multiple location appears rather better o in this regard, at least

at rst blush. As it says right on the label, multiple location is not
one-one, and a single object enjoying this feature can be found at
several regions. But not at all of them. A de nitional constraint on
multiple location as it appears above ensures that such an object is
not located at the fusion of the regions at which it is located.
Consequently, whereas this conception would be in nitely more
liberal than spanning, it would ban the most inclusive region from
being among those that can host God. On the grounds that this
conception is likewise too restrictive, let us disqualify a multiple
location conception of omnipresence.

Accordingly, let us understand our literal occupation account of

omnipresence as ubiquitous entension. Once again, to entend is to
be wholly and entirely located at some non-point-sized region (in
the case of omnipresence, at the maximally inclusive region) and to
be wholly located at each of that region’s proper subregions (in the
case of omnipresence, at every other region there is).27 Whereas our
earlier de nition of being ‘entirely located’ would then require that
there be no region disjoint from the maximally inclusive region at
which God is also located, it should be obvious that this condition is
automatically satis ed. Moreover, our earlier de nition of being
‘wholly located’ would then require that for every region there is,
God does not have any part that fails to be at that region, but again
assuming the mereological simplicity of God it should be obvious
that this condition is automatically satis ed as well. Recall, then,
our second puzzle: the problem of multilocation— how can something
occupy (in the ‘wholly present’ sense) two numerically distinct
regions? In exactly the same way God can occupy more than one
region without forfeiting mereological simplicity—by entending.

Two of Anselm’s three worries thus receive answers, but what of

the third puzzle we extracted from our discussion of Anselm: the
problem of containment—if to occupy a region is to be contained by
it, how can something that is essentially free of the constraints that
bind all creaturely things occupy a region? I think the best the
entension theorist of omnipresence can hope for here is to insist that
freedom from the constraint of location consists in God’s bearing
occupation relations accidentally rather than essentially. That is to
say, whereas God is wholly present at every region there is—a type
of presence that embeds the perfectly natural and fundamental
location relation—God would have existed even if there had been
no regions at all. In this respect, God enjoys a freedom from
occupation that many of God’s creatures do not; for creatures, but
not their creator, occupation proves to be an ontological condition.

A similar sort of concession seems equally appropriate in

response to our fourth di culty: the problem of timelessness—how
can something occupy a region and be atemporal? Despite the
inventive and ingenious literature on atemporality, perhaps nothing
can.28 It would seem that occupying a region of spacetime is
su cient for having some temporal location or other and to that
extent also ensures some literal temporal predication, but once
again, this may well be among God’s accidental rather than his
essential features. Had there been no spacetime, God would still
have existed. Mystery that it is, this verdict is no worse than the
declaration that God is atemporal to begin with—it only serves to
show that those who do not care for an essential tie between God
and time do not have to accept one on the entension reading of
omnipresence. And if an accidental tie is deemed unacceptable on
the grounds that God would thereby once again forfeit simplicity,
we may simply note that this complaint is better directed at the
temporal-parts commitments of pertension, not the mereologically
neutral stance of entension.

Our fth problem strikes me as considerably more di cult: the

problem of incorporeality—how can something occupy a region and
fail to have a body? My own view of the matter is that anything that
occupies a region is a material object, and that the occupier inherits
the shape, size, dimensionality, topology, and boundaries of the
region in which it is entirely located.29 Anyone similarly attracted to
the simple occupancy analysis of ‘material object’ and these related
theses has a bullet to bite if he wants to endorse an entension-based
reading of omnipresence, for God will then exemplify the shape,
size, dimensionality, topology, and boundaries of whatever is the
most inclusive region. We can again tender our previous reassurance
and declare that these are merely accidental and extrinsic
characteristics of God, but undoubtedly the concession will be
declined by many who think that these properties cannot be
accommodated in any form. It would seem that some kind of
embodiment will turn out to be an unavoidable cost of the present
hypothesis, but as we observed in the debate between Hartshorne,
Swinburne, Taliaferro, and Wierenga (explored in sect. II above), it
may be a cost borne by adherents of traditional non-occupancy
accounts as well. To be fair, however, the simple occupancy analysis
of ‘material object’ is certainly controversial (and negotiable for the
traditional theist). Moreover, it would be a reasonable rejoinder that
these are simply not the same kind of ‘body problem’ in any event
and that there is precious little support to be garnered on this
matter from Hartshorne and Swinburne’s earlier and hedged
embodiment theses.

Finally, recall our sixth puzzle: the problem of co-location—how

can two numerically distinct things each occupy the same region? A
common answer is that they can do so provided that they are of
fundamentally di erent kinds. The ‘fundamentally’ seems required,
for ‘being a co-located thing’ is a kind, and unsurprisingly all
coincident entities (if there are any) are at least of that sort. What,
though, are the fundamental kinds? Do objects divide at the most
basic level into the concrete and the abstract? Or perhaps into
substances, properties/relations, and facts? Or maybe into the
material and the immaterial? Or into objects, stu , and spacetime?
Or into the divine and the non-divine? Of course, one could dodge
the heavy burden of giving a full-blown answer to the question at
hand and maintain that whatever the correct response turns out to
be—God is of a fundamental kind all his own. Accordingly, whereas
God’s omnipresence would ensure God’s being co-located with every
other thing, such coincidence would always be of the harmless and
acceptable variety. It is worth noting, though, that even with such
widespread space-sharing, appeals to location might still be able to
serve their historical individuating function. Since the entension
theorist recognizes occupation as a one-many relation, it should
come as no surprise to learn that she may be attracted to an
individuation principle for located objects that says—necessarily, for
any located objects, x and y, x is located at all and only the same
regions as y i x = y.30

VII. SOME CONCLUDING REMARKS on OMNIPRESENCE AND CHRISTIANITY

What additional puzzles might confront the Christian theorist

tempted by entension and its apparently satisfying reading of
omnipresence? It depends.

Recall the nal suggestion of the preceding section: namely, that

the entension theorist may be attracted to an individuation principle
for located objects that says—necessarily, for any located objects, x
and y, x is located at all and only the same regions as y i x = y. If
our Christian theist is a Trinitarian recognizing three numerically
distinct persons each of whom is omnipresent, perhaps she will
hesitate over that individuation principle. Or what if our Christian
theorist accepts the Incarnation and holds that there is a special
sense in which the Son but not the Father was hanging on the cross,
despite the fact that both were and are omnipresent? Or what if our
Christian theorist maintains that the Son is literally present in the
consecrated bread and wine, while the similarly omnipresent Father
and the Holy Spirit are not?

These are excellent and di cult questions, and I have neither the
space nor expertise to enter here the controversies with which they
are associated. I will remark in closing, however, that once again
the relation of entension may serve as a promising source of
resolution for some of those debates. Even if one took the
Anselmian/Thomistic accounts to su ce for omnipresence,
entension remains available to provide partial readings of additional
special claims of location such as Christ’s presence in the Eucharist
or the speci c comings and goings of the Holy Spirit. Accordingly, if
the ‘located at’ in the principle noted above ranges over these very
di erent types of occupation, even the Trinitarian may embrace the
individuation principle without apology.

At the outset of this chapter, I noted that according to the

tradition of western theism God is said to enjoy the attribute of
being everywhere present, and I asked—‘what is it, exactly, for God
to manifest ubiquitous presence?’ As we have seen, for Anselm it is a
kind of knowledge, for Aquinas a kind of power, and for many of
their intellectual descendants a kind of mixture of knowledge and
power. On the strength of recent work in the metaphysics of
location relations, however, some western theists (and especially
some Christian theists) may wish to entertain with full seriousness
that God’s omnipresence involves a non-derivative and literal
location relation—the relation of entension.
NOTES

For comments and criticism on earlier versions of this paper I

thank Michael Rea, Thomas Flint, Andrew Arlig, Joseph
Jedwab, Robert Pasnau, Jonathan Scha er, Edward Wierenga,
Dean Zimmerman, the participants at the 2006 Mereology,
Topology, and Location conference at Rutgers University and
the philosophy departments at the University of Colorado at
Boulder, Baylor University, and West Virginia University.

1. Compare Psalm 139: 7–8: ‘Whither shall I go from thy spirit?

Or whither shall I ee from thy presence? If I ascend up into
heaven, thou art there: if I make my bed in hell, behold, thou
art there.’

2. Theorists other than relationalists can give the second

dismissive answer, too, so long as they do not include regions
in their ontology (i.e. so long as like the relationalists they do
not take spacetime to be an independent thing partially lled
or occupied by other entities).

3. With the hope of remaining as neutral as possible on the

debate about the nature of the divine being as conceived in
western theism, I will say very little on the term ‘God’;
nevertheless, in the subsequent discussion I will have occasion
to comment on the relation between omnipresence and other
commonly ascribed divine attributes such as simplicity,
eternality, incorporeality, omnipotence, and omniscience. On
 relationalism and substantivalism see Earman 1989 and
Nerlich 1994. Of course I’m not suggesting that the
relationalist need turn atheist, but I will not be investigating
relationalist reconstruals of occupation relations in what
follows.

4. Especially, Monologium chs. 20–2 and Proslogium ch. 13, in

Anselm 1948.

5. Wierenga 1988. I owe much of my understanding of Anselm

and Aquinas on omnipresence to Wierenga from the days
when I was among his students. I am delighted to
acknowledge that here.

6. Summa contra Gentiles, lib. 3 d. 68 in Aquinas 1975.

7. For a discussion of the tension between Aquinas’s doctrines of

omnipresence and eternity see La Croix 1982.

8. Perhaps there is a questionable move in sliding from ‘in nite’

to ‘all’ (e.g. one can think about in nitely many natural
numbers without thinking about them all and presumably, one
could bear in nitely many such presence-as-power relations
without bearing them to all of the regions). But let that go.

9. Summa Theologiae I. q. 8 in Aquinas 1975.

10. Wierenga 2006 and 1997. For the record, scholars are divided
on whether Anselm and Aquinas really advocate the non-
occupation accounts as a substitution for (as opposed to an
addition to) a literal reading of omnipresence, and Robert
 Pasnau (in private communication) has made an interesting
case for the combined view.

11. Hartshorne1941; Swinburne 1977; Taliaferro 1994 and 1997;

and Wierenga 1988, 1997, and 2006.

12. Hartshorne 1941.

13. Swinburne 1977: 102–4.

14. Taliaferro 1994 and 1997.

15. Wierenga 1988, 1997, and 2006.

16. For a more comprehensive discussion of the null individual, its

friends and foes, its philosophical pro le, and the prospects of
and obstacles to identifying it with God, see Hudson
(forthcoming a).

17. Note, however, that mereological simplicity seems

irredeemably forfeit on this proposal, and unless divine
simplicity is reinterpreted as something like there being no
di erence between subject and attribute or between essence
and existence in God, that undesirable consequence may be
enough to do it in.

18. The material in the next section ‘Occupation Relations’, which

introduces and discusses a number of alleged kinds of literal
occupation, is drawn largely from ch. 4 of my 2006. In that
work, I focus on material objects, but the present section
di ers from its predecessor by omitting the adjective
‘material’.
19. For accessible and interesting examples of this literature see
Casati and Varzi 1999; Gilmore 2003 and forthcoming;
McDaniel 2003; and Parsons forthcoming.

20. Parsons (unpublished). The rst two de nitions (i.e. of ‘entirely

located’ and ‘wholly located’) and the fourth and fth
de nitions (i.e. of ‘pertending’ and ‘entending’), while inspired
by Parsons, use a di erent primitive and have di erent
content than the de nitions given to those phrases by him.
Note that the de nition of ‘entirely located’ involves a claim
about the non-existence of a certain kind of region, while that
of ‘wholly located’ involves a claim about the non-existence of
a certain kind of object. Parsons is to be commended for
noting and correctly emphasizing the importance of this
crucial distinction.

21. In this and the de nitions to follow, I use ‘non-point-sized’

rather than ‘extended’ in order to be neutral (i.e. in order to
leave open the possibility of a receptacle that is the fusion of
at-least-two-yet-no-more-than-countably-many point-sized
regions—a region which would then be both non-point-sized
and non-extended).

22. Why the fanciness? Why not just say ‘x entends’ means ‘x is
located at a non-point-sized region and is a mereological
simple’? This won’t do, for the proposed de niens would then
apply to three of the four di erent ways an object may be
thought to be related to regions (to be discussed below), and
 one of the main aims of this section is to clearly distinguish
those di erent ways.

23. See the discussion of non-locality and quantum mechanics in

Parsons (unpublished). See also the support for entension
contributed by the null individual in Hudson (forthcoming a).

24. See McDaniel (forthcoming a) from which I borrow the term

‘spanners’. My characterization, however, di ers from his in
using ‘entirely located at exactly one’ and in replacing
‘continuous region’ with ‘non-point-sized region’ so as not to
prejudge the possibility of spatially or temporally disconnected
simples.

25. See Sider (forthcoming) who argues in this fashion not only for
the possibility of a single object occupying more than one
region but also for the possibility of a single region hosting
more than one object. Again, though, perhaps this establishes
at best a presumption in favor of the relevant thesis which
may be trumped by good arguments against multiple
occupancy or co-location.

26. Bigelow 1995; Markosian 1998; MacBride 1998; McDaniel

(forthcoming b); Parsons 2004 and (unpublished); Sider
(forthcoming); Simons 2004.

27. One quali cation: should it turn out that there is no maximally
inclusive region (i.e. if every space is a proper subregion of a
larger space), then entension is disquali ed owing to its
 de nitional link to being entirely located somewhere or other,
and omnipresence would simply amount to being wholly
present at each of the in nitely many contained regions.

28. An exception to this rule: Perhaps something can occupy a

region without being temporal if the world at which the region
is to be found is a space-only world. For a brief discussion of
this issue see Hudson (forthcoming b).

29. See the Introduction to Hudson 2006 and Markosian 2000.

Note that I did not say ‘of the region in which it is wholly
located’, for the entension theorist doesn’t think there is any
such unique region. For a discussion of what to say about
shape properties for the pertension, entension, spanning, and
multiple location theorists, see ‘the problem of shapes’ in ch. 4
of my 2006.

30. Of course, the pertension theorist attracted to location-as-

individuation principles could say this too, and ignore what
from his point of view is idle complexity in talk of regions in
the plural. For more on location and individuation principles,
see ‘the problem of diachoric identity’ in ch. 4 of my 2006.

REFERENCES

ANSELM (1948). St. Anselm: Proslogium; Monologium, trans. Sidney

Norton Deane. La Salle: Open Court.
AQUINAS, ST THOMAS (1945). Summa Theologiae, in Basic Writings of
Saint Thomas Aquinas, ed. Anton Pegis, 2 vols. New York:
Random House.

——— (1975). Summa contra Gentiles, trans. James F. Anderson.

Notre Dame: University of Notre Dame Press.

BIGELOW, JOHN (1995). The Reality of Numbers. Oxford: Oxford

University Press.

CASATI, ROBERTO, and VARZI, Achille (1999). Parts and Places: The
Structures of Spatial Representation. Cambridge, Mass.: Bradford
Books.

EARMAN, JOHN (1989). World Enough and Space-time: Absolute Versus

Relational Theories of Space and Time. Cambridge: MIT.

GILMORE, CODY (2003). ‘In Defense of Spatially Related Universals’,

Australasian Journal of Philosophy 81: 420–8.

——— (forthcoming). ‘Time Travel, Coinciding Objects, and

Persistence, in Dean Zimmermen (ed.), Oxford Studies in
Metaphysics. Oxford: Oxford University Press.

HARTSHORNE, CHARLES (1941). Man’s Vision of God and the Logic of

Theism. Chicago: Willett, Clark.

HUDSON, HUD (forthcoming a). ‘Con ning Composition, The Journal

of Philosophy.

——— (forthcoming b). ‘Lesser Kinds Quartet, The Monist.

——— (2006). The Metaphysics of Hyperspace. Oxford: Oxford
University Press.

LA CROIX, RICHARD R. (1982). ‘Aquinas on God’s Omnipresence and

Timelessness, Philosophy and Phenomenological Research 42: 391–
9.

MACBRIDE, FRASER (1998). ‘Where are Particulars and Universals?’

Dialectica 52: 20327.

MCDANIEL, KRIS (2003). ‘No Paradox of Multi-Location’, Analysis 63:

309–11.

——— (forthcoming a). ‘Brutal Simples’, in Zimmerman (ed.),

Oxford Studies in Metaphysics.

——— (forthcoming b). ‘Extended Simples’ Philosophical Studies.

MARKOSIAN, NED (1998). ‘Simples Australasian Journal of Philosophy

76: 213–26.

——— (2000). ‘What Are Physical Objects?’ Philosophy and

Phenomenological Research 61: 375–95.

NERLICH, GRAHAM (1994). What Spacetime Explains. Cambridge:

Cambridge University Press.

PARSONS, JOSH (2004). ‘Distributional Properties’, in Frank Jackson

and Graham Priest (eds.). Lewisian Themes. Oxford: Oxford
University Press, 173–80.

——— ‘Entension, Unpublished manuscript.

——— (forthcoming). ‘Theories of Location’, in Zimmerman (ed.),
Oxford Studies in Metaphysics.

SIDER, THEODORE (forthcoming). ‘Parthood’ Philosophical Review.

SIMONS, PETER (2004). ‘Extended Simples: A Third Way Between

Atoms and Gunk’, The Monist 87: 371–84.

SWINBURNE, RICHARD (1977).

The Coherence of Theism. Oxford: Oxford University Press. TALIAFERRO,

CHARLES (1994). Consciousness and the Mind of God. Cambridge:
Cambridge University Press.

——— (1997). ‘Incorporeality’, in Philip L. Quinn and Charles

Taliaferro (eds.), A Companion to the Philosophy of Religion.
Oxford: Blackwell, 271–8.

WIERENGA, EDWARD (1988). ‘Anselm on Omnipresence’, New

Scholasticism 52: 30–41.

——— (1997). ‘Omnipresence’ in Quinn and Taliaferro (eds.), A

Companion to the Philosophy of Religion.

——— (2006). ‘Omnipresence’ Edward N. Zalta (ed.), The Stanford

Encyclopedia of Philosophy (Spring ed.).
<http://plato.stanford.edu/archives/spr2006/entries/omnipres
ence/>, accessed 16 May 2008.
 CHAPTER 10

MORAL PERFECTION

LAURA GARCIA

PERFECT goodness is one of those attributes included in the

conception of God as the greatest conceivable being. Some refer to
this as the Anselmian concept of God, in honor of the eleventh-
century monk St Anselm who claimed to have discovered a proof for
God’s existence based simply on the de nition of God as a being
than which none greater can be conceived. Anselm’s de nition can
be rephrased in modal terms, as the claim that God is the greatest
possible being. The version of theism that operates from this
de nition of God we might call ‘perfect-being theology’. Traditional
proofs for the existence of God seek to provide strong rational
support to the claim that there is such a perfect being, and familiar
objections to God’s existence (e.g. the problem of evil) aim to
demonstrate that no such being exists.

In the 1970s Alvin Plantinga made use of the Anselmian concept

of God to develop a modal version of Anselm’s ontological argument
for God’s existence. His de nition describes the God of perfect-being
theology as one that exists necessarily and is essentially omnipotent,
omniscient, and morally perfect, and this de nition has become
standard in discussions about the nature and existence of the God of
western theism. Hence these discussions operate with a relatively
thin conception of God, since many of the key terms in the
de nition, including essential moral perfection, remain unde ned.
Philosophers nd this attractive in some ways, since it permits (even
invites) an a priori approach to explicating the divine perfections. In
fact, proposed de nitions or analyses of each divine attribute
abound in the philosophical literature. Striving for compossibility
among the perfections steers these discussions toward ever leaner
de nitions that will present fewer potential con icts with the
de nitions of other attributes.

One drawback for the minimalist approach, however, is that it

can impede the e ort to connect philosophical theology with
religious faith. In his treatise on the Anselmian concept of God,
philosopher Thomas Morris defends his claim that the God of the
philosophers (the God of perfect-being theology) is the God of
Scripture: ‘If the object of worship in the western tradition of
theology is intended to be the ultimate reality, and if the Anselmian
conception of God is coherent, if maximal perfection is possibly
exempli ed, then the God of religious devotion is the God of the
philosophers.’1 The number of conditions here is daunting to say the
least. But many philosophers believe that there is enough shared
content in the western (or at least in the Judaeo-Christian) concept
of God to provide at least a starting point for philosophical analysis
of that concept. In what follows, we will discuss three major ways of
modeling divine moral perfection and consider some of the major
objections to the claim that God is necessarily morally perfect.

THE POSSIBILITY OF MORAL PERFECTION

For most theists, moral perfection is the attribute most essential to a

divine being. If it turns out that moral perfection is incompatible
with omnipotence or omniscience, theists are more likely to
surrender the latter than to question divine goodness. One reason is
that only a perfectly good being is worthy of worship, while a being
who (for some plausible reason) has limited power or knowledge
might still be worthy of worship, as long as that being has as much
power or knowledge as it is possible for a being to have. Peter
Geach defends a non-Anselmian conception of God owing to his
belief that future contingent propositions about what humans will
freely choose are (at present) neither true nor false, and so cannot
be known (now) with certainty even by an omniscient being.2 It also
follows from Geach’s view that there are some things God cannot
prevent (free choices), though he can of course prevent the typical
e ects of those choices from occurring. Geach contends that since
God has perfect goodness and can carry out his plan for the world
no matter which choices free creatures make, he remains worthy of
worship. Eschewing the traditional doctrine of omnipotence, Geach
proposes that God is almighty, having as much power as a being can
possibly have in a world containing other genuinely free creatures.
 Perfect-being advocates criticize Geach’s proposal as inadequate
to the traditional theistic understanding of God, though objections
focus primarily on Geach’s view about future contingent
propositions. That view requires that some propositions are neither
true nor false, which violates a basic assumption of logic, the law of
bivalence, which states that every genuine proposition has a truth-
value. While many theists agree that propositions about future free
choices are genuinely contingent and that nite creatures cannot
know their truth-value with certainty, they insist nonetheless that
God knows with certainty for each such proposition whether it is
true or false. If Geach is correct in asserting that knowledge of
future contingent propositions is logically impossible, this will
require revisions in de nitions of divine omniscience and
omnipotence. Divine moral perfection remains untouched, however,
since presumably a morally perfect being will always choose the
good, whatever the extent of his knowledge and power.

But the moral attributes of God have come under attack as well.
William Rowe argues that perfect being theology is logically
incompatible with divine freedom.3 His case for this claim requires
the following two principles:

(a) Where possible world x is better than possible world y, God’s

act of creating x will be better than his act of creating y.

(b) If one action is better than another, then God cannot choose
the less perfect action over the more perfect action.4
Principle (a) assumes that there is some way of assessing the overall
value or desirability of possible worlds. Rowe admits that it may be
logically impossible for there to be a best of all possible worlds, but
he concludes from this that, given principles (a) and (b), God is not
free to create a world at all. In such a situation, God cannot choose
something less than the best, yet necessarily any world he creates
will be less good than some other one he could create. If it turns out
that, contrary to what we’ve assumed so far, there is a best possible
world, then God is morally required to create that world. He will
not be free to refrain from creating, nor will he be free to choose
among a variety of logically possible alternative worlds.

Rowe anticipates the objection that if God is a necessary being,

every possible world includes God, so it is never the case that one
possible world is better than another in terms of overall value—
every one already has in nite value. In reply, Rowe makes a
distinction between quantitative value and qualitative value,
claiming that even if every possible world contains the same amount
of quantitative value (an in nite amount), worlds with creatures or
with a variety of creatures are qualitatively more valuable than a
world with only God. That is, such worlds contain more kinds of
value. Against this it might be said that a world’s quality can be
assessed in more than one way. Some philosophers, in the tradition
of the Greek philosopher Parmenides, considered a world with only
one perfect being to be the summit of perfection and saw the
existence of lesser beings as undermining the overall quality of a
world. But Rowe could grant the Parmenidean conception of quality
without surrendering his argument for divine necessitation, since by
(a) and (b) a Parmenidean God would be morally obliged to refrain
from creating.

Most of Rowe’s theistic interlocutors grant principles (a) and (b)

and so nd themselves hard pressed to maintain both God’s moral
perfection and his perfect freedom. It may be useful to consider a
parallel dilemma that arises within perfect being theology. Paul
Helm and some others in the Reformed theological tradition propose
that among the criteria for rationality is the following principle: a
rational agent must always have a reason for acting, a reason that
justi es choosing this speci c action over every alternative action.5
Given this principle and the assumptions that God is perfectly
rational and that he created the actual world, God must have a
reason to create rather than not create and a reason to prefer
creating the actual world to creating any alternative possible world.
It follows in turn that the actual world is the only possible world
(understanding possibility here as metaphysical possibility rather
than as mere logical possibility or as conceivability). Since creating
the actual world was the best action God could perform, he is not
free to choose any alternative course of action. While such a view is
logically coherent, it is disturbingly counterintuitive. In my opinion,
this is a reason to reject the rationality requirements proposed
above and to allow that a rational agent may choose any satisfactory
means to accomplish his or her ends. If several satisfactory means
are available, a rational agent needs no additional reason for
choosing one of them rather than the others. If this leaves some free
choices partially unexplained (other than by saying they were
chosen for the desired end), then so be it.

For similar reasons, theists may wish to reject Rowe’s principle

(a), at least as he understands it. Theists in the Thomistic tradition
conceive of God as containing all the perfections found in creatures
(but in a much higher way). This means that a world with God in it
is already maximally excellent, not just quantitatively but
qualitatively as well. Creatures re ect the perfections of God and
participate in them to a greater or lesser degree, but the addition of
creatures neither increases nor decreases the qualitative goodness of
the world in Thomas’s perspective. Whatever God’s reasons for
creating, this theological tradition emphatically denies that they
include making the world a better place in some way.

Theists in the Anselmian tradition should be equally suspicious

about the claim that a moral agent has a duty to perform the best
action he can, as Rowe’s principle (b) requires. One troubling aspect
of this claim is that the ‘best action’ is understood in the context of
principle (a) as the action that achieves the best outcome. This
simply assumes that the moral evaluation of actions is based
primarily on their results, an assumption that is highly questionable
and that many theists reject. Without (a) and (b), the claim that
divine moral perfection is logically incompatible with divine
freedom does not succeed. Further, as we shall see below, the virtue
theory of morality provides a di erent model for understanding
God’s motives in creating a world, a model that preserves a place for
divine freedom.

MODELS OF MORAL PERFECTION

Conceptions of divine goodness are generally modeled on

conceptions of human goodness. Di erent moral theories, then, give
rise to di erent analyses of moral perfection for humans and so, by
extension, for God. Historically speaking, the Big Three moral
theories are the virtue theory, developed by Aristotle and St Thomas
Aquinas, the duty theory, with Immanuel Kant as its main
champion, and the utility theory, advocated by Jeremy Bentham
and John Stuart Mill, and sometimes defended these days under the
more inclusive banner of consequentialism. Recent discussions of
divine goodness rely almost exclusively on the latter two theories,
de ning moral perfection either as doing the best thing one can do
(either all told or for each person a ected) or ful lling all one’s
duties. As we shall see, neither of these theories works well as an
account of divine goodness. Virtue-based theories show more
promise, depending on what role is assigned to the virtues in
assessing the moral quality of actions and persons.

Maximizing the Good

Utilitarian and consequentialist moral theories treat certain

desirable states of a airs as the primary locus of value. Generally
the valuable states of a airs involve persons having positive
experiences of some kind, such as feeling pleasure or getting things
they want. Actions are good if they produce more value or less
disvalue than any alternative action available to the agent. In order
to avoid egoism, most consequentialist theories require that a good
action produce more total value for everyone a ected, not just for
the agent. (How this value should be distributed among persons
a ected by the act is a matter of dispute among consequentialists.)
Given the di culties involved in calculating the probable e ects of
each action, some consequentialists modify the theory to allow that
an action is good if it is of a type generally known to produce the
most value in situations similar to the present one. For the most
part, morally good persons will be those who consistently choose
morally good actions. Virtues may be treated as morally relevant as
well, but only if (and because) they increase an agent’s tendency to
produce good actions.

Consequentialism is open to many lines of criticism. First, there

is something odd about treating moral value as simply a function of
non-moral value (the value of various states of a airs). While that
move has the advantage of providing an objective and quasi-
empirical basis for moral judgments, it overlooks the importance we
ascribe to motivations in assessing moral agents and their actions.
Second, it is di cult if not impossible for a human agent to
calculate the total e ects of her actions even in the short term,
much less over the course of her lifetime and beyond. Third, this
theory con icts with one of our rmly rooted moral intuitions, that
in our actions we should be motivated by concern for others and
their true welfare rather than by an attempt to calculate the
optimi c action (the one that optimizes the proportion of value over
disvalue). In fact, we want others to care about us and our welfare
even if they can do nothing to help us or are relatively inept in their
e orts to help.

Some of these criticisms lose their relevance when we turn to

consider divine moral perfection. Presumably an omniscient moral
agent can anticipate all of the e ects of each of his actions6 and an
omnipotent agent can see to it that his actions have only the e ects
he wants them to have. Taking valuable states of a airs as primary
then, a divine agent should be able both to maximize value in the
universe and to maximize value for each created human being.
Unfortunately, the task of maximizing non-moral value is not
logically possible and so is not one that even God can perform. For
any amount of such value, i.e. for any number of creatures and their
pleasures or satisfactions, there could always be more—more
creatures, more pleasures, or more varieties of pleasure. Similar
problems have plagued the project of de ning the best of all
possible worlds. For any world God creates, surely there could be a
world with one more creature in it, one extra day, and so on.7 If we
omit the requirement that God maximize value overall, perhaps we
could retain the lesser requirement that God must maximize value
for each of his creatures (or each of his rational creatures). But the
same paradox of maximization applies in the individual case as well,
since there could always be more and di erent positive experiences.

One way to get around the di culties involved in maximizing

bene ts to individuals would be to understand bene ts primarily in
terms of moral value. Perhaps the focus of a morally perfect being
should be on maximizing the amount of moral good in the universe,
where moral good is a result of the morally right actions of free
agents. On this view, the truly valuable states of a airs (creatures’
morally good actions) necessarily incorporate states of a airs (free
choices) that are not up to God. God cannot directly cause creatures
to choose what is morally right without removing their freedom.8 In
his free will defense against the problem of evil, Alvin Plantinga
makes use of this conception of moral good as consisting in the
number of morally good choices made by creatures. He contends
that it is logically possible that God cannot create a world with as
much moral good as that found in the actual world without
permitting an amount of moral evil at least as great as what is
present in the actual world.9 If Plantinga is correct, then God’s
moral perfection is not compromised by the existence of moral evil
in the amount we nd in our world. But we are still left with a
maximization problem, since it seems possible for God to have
created additional moral good in the world simply by creating more
creatures with moral freedom, even if this also means permitting
additional moral evil. Beyond this, some will contend that a
perfectly good being should maximize both moral and non-moral
good for his creatures, which will again raise the paradoxes of
maximization that William Rowe has pointed out.

Plantinga may well be onto something in focusing on moral

good, however. It does seem to most theists that maximizing the
non-moral value in a person’s life is not necessarily the best thing to
do for him. As John Henry Newman explains:

The Church … regards this world, and all that is in it, as a mere
shadow, as dust and ashes, compared with the value of one single
soul. She holds that, unless she can, in her own way, do good to
souls, it is no use her doing anything; she holds that it were better
for sun and moon to drop from heaven, for the earth to fail, and for
all the many millions who are upon it to die of starvation in
extremest agony, so far as temporal a iction goes, than that one
soul, I will not say, should be lost, but should commit one single
venial sin, should tell one willful untruth, though it harmed no one,
or steal one poor farthing without excuse. She considers the action
of this world and the action of the soul simply incommensurate,
viewed in their respective spheres.10

This passage suggests that the most important goods for human
persons are ones that cannot be achieved without their free
participation. St Augustine claims that ‘God, who created you
without you, will not save you without you.’11 On this view,
maximizing the good for a given person requires that person’s free
cooperation and continues to depend upon it throughout his or her
earthly life. Further, one person’s hardships may be an occasion for
another person’s moral growth or conversion, so that a divine being
must choose among many di erent kinds of goods—maximizing
some may be incompatible with maximizing others, and there is no
guarantee that di erent kinds of goods will be commensurable with
one another. Given these di culties within the consequentialist
model, then, there is reason to search for a more satisfactory
explication of divine moral perfection.

Doing One’s Duty

Duty-based theories of morality treat actions as the primary locus of

moral value, evaluating actions as good or bad, right or wrong,
depending on whether they accord with what the moral law
prescribes. Duties themselves are variously grounded. Immanuel
Kant claims that the fundamental moral law, the Categorical
Imperative, draws its authority from reason. Acting against this law
leads a person into contradictions and incoherence. Other duty
theorists claim that some basic moral laws expressing absolute
moral duties are known immediately through intuition. Among
intuition theorists, some hold that intuition yields particular
judgments about which actions are right or wrong, while others
claim it only tells us which kinds of action are right or wrong. A
third group claims that we intuit only which things are intrinsically
good, so that principles of right and wrong action must be derived
subsequently.12
 Divine command theories of morality hold that the moral law
rests on God’s commands, so that following God’s will (generally as
revealed in Scripture) is what duty requires. Similarly, for God to do
his duty is for him to act according to his will. The result, of course,
is that since God always and necessarily acts according to his own
will, he is necessarily morally perfect. On the other hand, since
there is no independent source to give content to the concept of
moral goodness, this theory cannot rule out the possibility of divine
actions that we would consider cruel or malevolent. Indeed, on this
theory such actions would actually be morally good. This outcome
has led many philosophers to reject the divine command theory of
morality, though some continue to support versions of it that
purport to escape the unpleasant implications just noted.13 In what
follows, we will assume that there is some objective basis for
ascertaining one’s moral duties independently of God’s commands,
and that humans do in fact know what duty demands of them in
most situations. The Ten Commandments might serve as a brief
summary of these duties.

Duty theories of morality do not require the maximizing

calculations involved in the various versions of consequentialism. As
long as one can apply a universal moral norm to a speci c instance,
one will know whether a given action is right or wrong. Lying is
wrong; to omit this income from my tax return would be lying;
hence, it is wrong for me to omit it. But the duty theory faces
di culties of its own. One of these stems from cases of con icting
duties, where it seems one has a strong duty to do two (or more)
actions that are incompatible with each other. Perhaps Smith is
kayaking with her daughter and the daughter’s friend when the
girls’ kayak overturns; she is obliged to try to rescue both, but an
approaching waterfall makes it possible to rescue only one person,
and the daughter’s friend is the one nearest to her. Some duty
theories provide rules for resolving con icts between duties, though
these can be controversial and di cult to defend, and they
introduce complications into the theory. Kant famously denied any
moral importance to so-called ‘special relations’, which seems to
imply that in the case of the kayak incident, Smith should try to
save the friend rather than her daughter. If one disagrees with Kant
and allows some relevance to special relations, this complicates the
moral theory and makes it more di cult to apply.

A separate criticism of duty theories is that they place

insu cient emphasis on the agent’s motives and intentions in
acting. Kant o ered a negative criterion here, arguing that an action
is morally good only if the agent acts out of respect for the moral
law alone, but this requirement seems both impossible and
misguided. Presumably Smith’s daughter would not be consoled to
nd that her mother had saved her life purely out of respect for the
moral law, because it was her duty to do so. She would rightly feel
that love should also have played a role, and not just a supporting
one. Finally, our moral intuitions suggest that ful lling one’s duties
is a kind of minimum in the moral life, and that a truly good person
would go far beyond what duty requires in her dealings with other
people. She would often do things that were beyond the call of duty,
that were ‘supererogatory’. But since every good deed that exceeds
what duty requires falls into this category, it is hard to say just how
much extra good one must do in order to qualify as a really good
person, much less a perfectly good person.

Applying the duty model to divine goodness results in something

like the following de nition: God’s moral perfection consists in his
perfectly ful lling all his moral duties. Thomas Morris treats this as
capturing at least part of the concept of divine goodness. However,
he anticipates a major obstacle to applying the duty theory to God,
since that model assumes that an agent has a duty to act in a certain
way only if he or she is free to do otherwise. According to perfect-
being theology, God necessarily acts in accordance with moral
principles. While humans can fail in their duties, it is logically
impossible for God to lie, to break his promises, and the like.
Morris’s solution is to claim that the duty model of moral goodness
applies to God only analogically. God always acts in the way that a
being with moral duties would act in his situation, even though
strictly speaking he has no moral duties.

Morris is reluctant to call this a de nition of divine moral

perfection however, since he endorses the standard duty theory as
the correct analysis of our concept of moral goodness, and he also
believes that Scripture ascribes duties to God in a fairly
straightforward way. ‘In the biblical tradition, God has been
experienced as one who makes and keeps promises, who enters into
covenant relations, and who does not lie.’14 (Of course, the biblical
tradition also presents God as a being who cannot lie and cannot do
evil, so it is not clear that the biblical tradition helps resolve the
philosophical question of whether God has duties.) Morris concludes
that while God’s ful llment of his duties is not morally good, it is
nonetheless good in a way that is not reducible to metaphysical (i.e.
non-moral) goodness. He introduces the term ‘volitional goodness’,
which is ‘goodness residing in or arising out of a being’s will or
character when that will or character is not itself causally
determined in that respect by anything independent of that being’.15
Volitional goodness covers both actions that ful ll duties God would
have were he subject to duties and supererogatory acts for which
other courses of action were metaphysically possible for God. This
preserves the duty model as the primary analysis of moral goodness
without surrendering the claim that God is perfectly good, since
volitional goodness includes both moral goodness and acts of
supererogation, and God has more volitional goodness than any
other being.

While this represents a valiant attempt to reconcile divine

goodness with the standard duty model of morality, it is not a
complete success. For on this view God’s acting in accordance with
duty is not morally good. It may be good in some metaphysical way
—what Morris calls ‘axiological goodness’—but it is not indicative
of God’s moral perfection.16 On Morris’s duty model, an act can be
morally good only if it derives from the kind of free choice among
alternatives that a perfect being lacks when that being is subject to
moral requirements. Volitionally good acts are not freely chosen,
but are necessitated by the divine nature. Even though divine acts of
supererogation do count as morally good (because freely chosen),
they are not morally good according to the standard duty-based
concept of moral goodness.

Morris attempts to bypass this problem by suggesting that ‘when

religious people claim that God is morally good, meaning that he
acts in accord with moral principles, they are merely using that
axiological conception with which they are most familiar, moral
goodness, to describe or model an aspect of divinity functionally
isomorphic with, though ontologically di erent from, human
goodness’.17 Religious people might well respond that they thought
they were using the ordinary (human) concept of moral goodness in
claiming that God is perfectly good. But it is possible to preserve the
standard (ordinary) duty model of moral goodness for God only if
one surrenders the claim that having moral duties requires being
able to act against them, and Morris believes that claim is essential
to the duty model.

Even if Morris’s modi ed duty model (the volitional

interpretation of divine goodness) succeeds, it generates a problem
of its own. Including acts of supererogation within moral perfection
introduces the maximizing principle once again; if some number of
gracious, supererogatory deeds is good, wouldn’t more of them be
even better? This point is pressed at some length by Earl Conee,
who claims that supererogatory acts are in fact essential to the
concept of moral perfection: ‘Nothing less than the best is perfect.
The agent of a morally imperfect act is not a morally perfect agent.
An act is morally imperfect if its agent thereby neglects one that is
morally better. And supererogatory acts are precisely those which
are morally better than others that are right.’18 Conee anticipates
the objection that supererogatory acts are not morally required and
so do not belong to the concept of moral perfection as such. Colin
McGinn, for example, contends that ‘surely if an agent always
conforms his actions to the moral norms that apply to him, there
can be no room left for moral imperfection to creep in’.19 Conee
replies that the anticipated consequences of one’s actions clearly
contribute in some way to the moral value of those actions, and that
the value of these consequences has no intrinsic maximum. If the
duty theorist accepts the moral relevance of an action’s accurately
foreseen e ects, then I believe Conee’s argument succeeds. He
concludes (along with Rowe) that it is impossible for any being to
exhibit moral perfection, since for any given value of an action’s
foreseen consequences, there could in principle be an action that
yields a higher value. Hence the duty-plus-supererogation model of
divine moral perfection succumbs to the same maximization
problems that beset the consequentialist model.

If the duty theorist retreats from the di culties raised by

supererogatory acts back into the duty model simpliciter, still further
obstacles lie in the path of explicating divine goodness by way of
the duty model. First among these is Conee’s contention that a
perfectly good being would not be content simply to ful ll his
duties, since even human moral goodness often goes beyond this to
include acts of heroic generosity and sel essness. Further, some
believers deny that God has any duties at all, since he is the
sovereign lord of the universe and every creature owes its existence
and all of its powers to him. While there is much that creatures owe
to God, what could the sovereign creator owe to creatures? Part of
the strangeness in the idea that God has duties stems from this vast
asymmetry in the relationship between God and humans, and part
of it is due to the felt absurdity of saying that there are things God
should do, as though there were a chance that he might not do them.
These considerations make it di cult (though perhaps not
impossible) to formulate an adequate de nition of divine moral
perfection within the con nes of the standard duty model.

A Combination Theory

Some philosophers, aware of the di culties in both the duty and the
consequentialist models, attempt to combine the two in the hope of
counteracting the weaknesses of each with the strengths of the
other. One such attempt is sketched in some detail by T. J. Mawson:
‘God’s perfect goodness then is his perfectly ful lling his duties
toward his creatures and, furthermore, whenever there is a logically
possible best thing for him to do for them, his doing that too, his
perfectly loving them.’20 Mawson takes it for granted that God
ful lls all his duties, and that he ful lls them necessarily, so he
disagrees with Morris’s claim that having duties requires that one is
free to act against them. Both Morris and Mawson note that an
omnipotent and omniscient being necessarily ful lls his duties, then,
but Mawson claims this is not because God lacks an important kind
of freedom but because he has perfect freedom. ‘These properties
[omnipotence and omniscience] entail that there is nothing that
constrains God’s actions (no external power that can trump his will
and no ignorance that can misdirect it).’21 Hence a perfectly free
being necessarily ful lls all his duties. The e ectiveness of this
solution depends on whether one nds it essential to the concept of
a morally good action that the agent be free to fail to perform it,
since on Mawson’s view God acts dutifully as an automatic and
logically necessary e ect of his recognizing the right action to
perform and having the power to perform it.

One di culty for Mawson’s ‘best-action’ model of divine

goodness is that it requires God to choose whatever is best for each
creature. For now we may set aside the problems that beset
maximizing requirements in general, since Mawson requires that
God act for the best vis-à-vis each person only when a best action is
logically possible. Still, God’s plans for the universe or for the
human race might be such that the best action in a given situation
has little to do with bene ts to this speci c individual (except
insofar as having a part in an excellent overall plan counts as a
bene t). Must God bene t a particular person at the expense of
other important values, including bene ts for other persons?
Further, the best-action model requires that it is possible to assign a
relative value to bene ts and harms in order to determine, say,
whether a bene t that lasts for all eternity outweighs a temporal
harm. If bene ts of very diverse kinds are commensurable in this
way, no doubt an omniscient being can make the calculations,
perhaps even minimizing the harm-to-help ratio for each creature.
Nonetheless, Mawson’s theory confronts three further obstacles:
maximizing bene ts for one person may not be compatible with
maximizing bene ts for others; many important personal bene ts
cannot be e ected by God’s actions alone (moral growth, friendship,
etc.); and God on this theory would be acting immorally in
preferring some other valuable state of a airs to maximizing
bene ts for each human being.

Finally, both duty models and consequentialist models of

morality focus on right actions to the neglect of other important
features of moral agents such as their attitudes and intentions. On
both theories, reason determines what is the right action in a given
situation either by anticipating the outcome of the action or by
assessing its conformity with the moral law (or some combination of
both). The role of the will is simply to choose the action that reason
commends. On consequence-based theories, the assumption is often
that there is one best action, thus limiting God’s options to a single
alternative. Duty theories allow God a wider range of choices, as
long as none of his actions violate any duties. But if God has a duty
to do the best he can (as Conee suggests), then in the divine case the
duty model collapses into the consequentialist model. That is, if God
must choose the overall best action (the action producing the best
consequences), assuming that such an action (in God’s case) will
never be one that violates a duty, the duty criterion can simply be
omitted.22

One advantage to an intellect-centered approach to moral

evaluation is that it simpli es the argument for divine moral
perfection. As Richard Swinburne points out, since an omniscient
being necessarily knows what is the right thing to do and an
omnipotent being necessarily has the power to do what is right,
God’s moral perfection is entailed by his omniscience and
omnipotence.23 The disadvantage of this intellect-based approach is
that it seems to make moral motivation entirely a matter of
rationality. Indeed, Immanuel Kant claimed that reason should be
the only motive for right action, as other motives can con ict with
the directives of reason. But it seems that the human will is
naturally drawn toward the good (that is, by non-moral values such
as existence, beauty, pleasure, ful llment, and knowledge) and that
we make choices for the sake of such good ends. So perhaps moral
excellence consists at least partly in what kinds of goods a person
wills (or seeks) and whether he or she values them appropriately.
Such considerations lead in the direction of virtue theories of ethics
as potential models for conceiving of divine moral perfection.
Acting Virtuously

Virtue theories treat various internal features of moral agents as the

locus of moral value, especially traits of character, intentions,
motives, and the like. Morally good traits are those that dispose a
person to choose and act well. Acting well is generally de ned in
terms of seeking what furthers the genuine good of persons, either
that of the agent herself or, when others are the objects of an action,
their genuine good. Some read Aristotle’s ethical theory as a virtue-
based theory that attempts to show that seeking one’s own genuine
good (or ourishing) requires that one also seek the genuine good of
others for their own sake. Whether or not that claim is true, we will
take it for granted in this discussion that to act virtuously is to act
solely from virtuous intentions, desiring, favoring, and hoping to
attain what is truly good for the person (or persons) who is the
object of one’s actions. A virtuous person also takes into account the
e ect of her act on any persons she foresees will be a ected by it,
aware unintended but foreseen harm to others may render an
otherwise virtuous action vicious in a particular instance. This is the
subject of the much-maligned and widely misunderstood doctrine of
double e ect.

Well-being or ourishing is of course a type of non-moral good,

but it is not the locus of moral value for virtue theories. It can be
given content by re ecting on what Aristotle called the nal cause
of a thing—its goal or full development, ourishing, wellness, etc.
Aristotle de nes the virtues as settled dispositions or states of
character that incline a person to feel, respond, and act in ways that
serve the ourishing of persons. More recently, Jorge Garcia
proposed a di erent way to give content to the goal of virtuous
actions, de ning the virtues in connection with the roles we ll in
the lives of persons (ourselves and others). On Garcia’s view, virtues
are the states of character (dispositions to feel, respond, intend, act,
etc.) that make us good in our morally signi cant roles, as parent,
friend, colleague, and so on. In order to know what traits make an
agent good in a role, we can re ect on what we (humans) naturally
expect from those who ll that role in our lives. Garcia argues that
what we most want and expect from other persons is that they be
motivated by goodwill toward us, that they be concerned for our
genuine well-being for its own sake. When we assess their character
and their actions from a moral point of view, their motives and
intentions are the key to our assessments. Persons of good will may
fail to actually bene t us, owing to ignorance, ineptitude, lack of
opportunity, or other factors outside their control, but we do not
condemn them morally for this kind of failure.24

Lists of moral virtues (and corresponding vices) go back at least

as far as Socrates, and there is a surprising amount of agreement
among them. Such personal virtues as temperance and courage (or
fortitude) typically appear along with interpersonal virtues such as
fairness, generosity, and friendliness. The in uence of Christianity
on western thought has added such qualities as compassion and
humility to the list. While this presents a rather daunting moral
program, Garcia’s virtue theory claims that the whole set can be
reduced to one overarching virtue—genuine concern for the well-
being or good of persons, especially of all the persons a ected by or
in the scope of one’s actions. An action performed from this
intention (and from no harmful or neglectful intentions) is a morally
good act regardless of whether it is successful (that is, whether it
actually bene ts the person or persons involved). On this view, love
of persons is both the root and key component in all the virtues.

One common criticism of virtue theories is that they cannot

accommodate the kind of moral absolutes that one nds within duty
theories of morality, and especially within the moral teaching of the
major theistic religions. Striving to acquire the virtues is a gradual
process, one that evokes the image of a sliding scale rather than of a
line that cannot be crossed. Garcia replies that virtue theories do
support moral absolutes, since some actions are such that it is
always wrong to intend them—taking an innocent human life,
treating a person with ridicule or contempt, and so on. Deliberately
acting in this way is acting from a motive directly opposed to the
virtue of love or goodwill, and such actions are always wrong. The
extent to which one’s motives depart from care and goodwill
determines the degree of their viciousness and so the degree to
which they are morally blameworthy. While it is wrong to remain
indi erent to another’s su ering, it is worse to purposely in ict
su ering on an innocent person. To in ict su ering on one’s own
child is more blameworthy still, given the greater love and care that
is properly owed to a son or daughter.

A virtue theory along these lines can be extended to provide a

conception of divine moral goodness, though the transition is not
exactly seamless. Perhaps the most obvious de nition of divine
moral perfection in terms of the virtues is to say that a morally
perfect being perfectly exempli es all the (standard) moral virtues.
But this will not do. Some human virtues, such as temperance or
courage, have no obvious divine equivalent,25 and others, such as
humility, seem inappropriate for God. Further, unlike human agents,
a divine agent cannot fail to act from virtuous motives or intentions.
This last point raises Morris’s worry about whether a trait
exempli ed necessarily by an agent can be properly included in an
account of that agent’s moral perfection. Human virtues must be
acquired gradually by acting well in various situations and thereby
developing the right habits. Nothing like this process could
characterize divine virtues.

On the positive side, virtue theories locate moral perfection in

the will rather than in the intellect, treating right actions as those
that are properly motivated and directed toward the right ends. This
model allows room for freedom to operate even for a perfectly good
being, since there are many di erent ends that can be chosen for the
right reasons. Without the maximizing requirement of
consequentialist theories or the rationality requirement of duty
theories, God is left free to create or not to create, and to create any
number of worlds that re ect his glory out of his gracious
generosity. This in turn suggests a more promising de nition of
divine moral perfection, understood not in terms of perfectly
exemplifying the whole list of virtues but as exemplifying perfect
love. If love is at the root of the human virtues, this de nition has
the further advantage of providing a clear analogy between human
and divine moral goodness.

St Augustine and St Thomas Aquinas both endorse a concept of

moral perfection along these lines. Thomas argues that God wills the
existence of creatures out of his goodness, but that he does not
create out of necessity:

Accordingly as to things willed by God, we must observe that He

wills something of absolute necessity: but this is not true of all that
He wills. For the divine will has a necessary relation to the divine
goodness, since that is its proper object. Hence God wills His own
goodness necessarily, even as we will our own happiness
necessarily.…[But] since the goodness of God is perfect, and can
exist without other things inasmuch as no perfection can accrue to
Him from them, it follows that His willing things apart from Himself
is not absolutely necessary.26

While God has no need to create, Thomas does say that, having
created human beings, God necessarily wills their good. He cannot
hate anything that he has made.27 Since God is the highest good, the
greatest good for human persons is to know and love God and to
live in his presence forever. Hence, this is the destiny that God wills
for each person.

We saw earlier that the duty model of divine goodness hit a snag
over the question of whether a perfect being could be free to act
against a requirement of duty. Perfect-being theologians are inclined
to answer this question in the negative, though this puts a strain on
the idea that God has duties in the ordinary sense. There are at least
four possible solutions to that problem. First, retain the standard
duty model and surrender perfect-being theology, so that it is
possible for God to act against a moral duty (adding perhaps that he
never actually does so). While some theistic philosophers might nd
this acceptable, those committed to perfect-being theology would
vigorously oppose it. Second, retain the essential moral perfection of
God and surrender the freedom requirement of the duty model, so
that even a being who necessarily ful lls his duties quali es as
morally perfect. While this move is attractive in some ways, it sits
uneasily with important components of the duty model, e.g., that
ful lling one’s duties is worthy of moral praise, or that duties are
inherently prescriptive (since for a perfect being they will simply be
descriptive). A third solution, that of Thomas Morris, retains both
perfect-being theology and the freedom requirement of the duty
model but argues that a being who necessarily acts according to
duty can still be called perfectly good by extension, as long as his
actions are not coerced by anything outside his nature. This solution
has promise, but it runs counter to the spirit of the duty model. Any
necessitation of one’s actions, whether from without or within,
undermines the prescriptive element essential to the duty model.

Fourth, one could retain perfect-being theology and jettison the

duty model of moral perfection. Morris objects to this move in part
on the grounds that the Bible describes God as having duties that he
can be counted on to ful ll. But virtue-based moral theories can also
accept the language of moral duties, de ning these in terms of the
virtues. For instance, if all one’s actions should be motivated by
genuine concern for the good of others, then one has a duty not to
intend harm to them. Admittedly, there is a di culty in ascribing
duties to God in this way. Does God have a duty not to harm the
persons that he has created? Putting the question like this, I believe,
shows both what is right and what is wrong about the duty model of
divine goodness. On the one hand, traditional theists want to say
that God must not harm his creatures, and on the other hand they
want to say that of course God could never do such a thing anyway.
Why? I believe the answer is not that God is necessarily dutiful but
that he is necessarily virtuous, necessarily loving. Indeed, love is
one of the names of God. Creatures are not virtuous by nature and
can be tempted toward vicious actions of one kind or another, so the
language of duties (especially as prohibiting certain kinds of
actions) is appropriate for them. But God as perfectly virtuous
cannot act viciously, so to say that he has a duty not to act badly
makes little sense.
 As we saw earlier, Morris acknowledges that there is a clear
tension between divine perfection and the language of duties, and
he sides with perfect-being theology in claiming that moral
goodness is essential to God. Putting these two together, Morris
arrives at a picture of a God who rst knows what his duties are vis-
à-vis other beings in the universe and then necessarily wills in
accordance with that knowledge. Of course this vastly
underdetermines an agent’s choices even for nite agents. It tells us
not to do anything wrong, but little about the good that we should
do. As such, it hardly seems a recipe for exemplary moral
achievement. ‘Doing one’s duty’ can even carry negative
connotations, since conforming one’s actions to a rule may be done
for a wide variety of motives. This is one reason for supplementing
the duty model, giving moral extra credit for supererogatory actions
that go beyond the call of duty. But if supererogatory actions are the
real key to moral excellence, it appears that duty is not the most
fundamental moral concept after all. A virtue theory, on the other
hand, can explain what is good about supererogatory acts, since
they go beyond the demands of justice to include heroic and self-
sacri cing actions motivated by love and compassion.

EVALUATING MORAL AGENTS

Consequence-based and duty theories of morality share a common

limitation in focusing on actions rather than on agents. While
actions are more accessible for moral evaluation, they present only
part of the picture. In evaluating the moral caliber of a person,
emphasis falls on his or her character—desires, concerns, a ections,
intentions, humility, perseverance, and the like. Part of the work of
reason in the moral life is to determine what is the right action in a
speci c situation. Aristotle calls this the work of practical reason,
applying general principles to the particular circumstances one
confronts. Among the in uences on practical reason are a person’s
formed attitudes, habitual responses, and guiding concerns, and
these are the province of the virtues. Hence, even if the bene ts
model or the duty model of morality can provide an acceptable
theory of right action, both would fall short of providing a full
moral assessment of the agent. But when we are attempting to
explicate divine moral perfection, this is precisely the kind of
assessment that is needed.

Focusing on the intentions of moral agents rather than on their

actions also makes it possible to de ne God’s moral perfection
without compromising his freedom. While God necessarily wills his
own being and goodness, he already possesses these, so nothing
compels him to create. Also, since any universe would be created
out of love, there are few limits on which worlds God can create.
There may be limitations beyond simple logical ones, perhaps moral
or aesthetic considerations, but in nitely many worlds will be
compatible with these requirements. With respect to individual
creatures, Thomas says that non-rational creatures exist for the sake
of the perfect ordering of the universe as a whole, and also for the
sake of human beings who are entrusted with the care, cultivation,
and proper use of the natural world. Humans, on the other hand, are
willed for their own sake; that is, they are not a proper object of use
by anyone (including God), since they are intrinsically good and not
merely instrumentally good. God desires the true good of persons,
which is that they should live forever in communion with him and
with one another, free from pain and sorrow and death.

It may seem that there are easy counterexamples to this last

claim, since human beings are not free from pain and sorrow in this
life and may (because of age or some other impediment) fail to seek
communion with God. The reply to this objection is complex. First,
recall that the good for persons is one that cannot be realized apart
from their free cooperation, since there is no genuine love or
communion that is coerced. Second, while God does not will evil to
any person, he may allow pain and su ering that result from
humans’ misuse of freedom, from disturbances in the natural order
(perhaps as an e ect of original sin), from the ordinary operation of
physical laws, and so on. The compatibility of these forms of pain
and su ering with God’s perfect goodness is defended in theistic
responses to the problem of evil. Third, as William Alston has
argued persuasively, it is di cult for nite minds to assess all the
kinds of positive goods or values there are (especially if some take
place in the next life) or to understand the many connections there
may be between the pains of this life and highly signi cant goods,
known and unknown.28 The Christian Scriptures claim that God can
make every situation in a person’s life conducive to this end, though
perhaps not without that person’s free cooperation (or at least
consent).29

The virtue theory of divine moral perfection entails that God’s

love for each person is unconditional and unimpeded by
incompatible motives. But what is it that love requires with respect
to each person? Presumably, God desires that each one should come
to enjoy communion with God and with other persons. Can this end
be reached by a person who has not freely chosen it? If the answer
is yes, then God is not required to provide for each person an
opportunity to make such a choice. If the answer is no, then
presumably God does provide such an opportunity for each person,
either before or after death.

On this point, Thomas Flint raises the following objection: ‘Once

God creates [a person], can he have perfect concern for [her] if he
chooses to put [her] in situations less conducive to [her] happiness
than he might have? Maximizing concerns may be less here than
they are for the consequentialist, but it’s not obvious that they have
totally evaporated.’30 A virtue theorist might reply that what is
morally signi cant is that God wills the true good of each person
and that he uses every available means toward that end. But it is
also important to note that the end in view is not exactly the same
for each person. Each soul is unique, and each is given gifts and
graces, challenges and trials, that are known to that person and to
God alone. Perhaps persons in heaven are not experiencing some
kind of generic human bliss, but are aware of God’s beauty, truth,
joy, and mercy as it is re ected in (the whole of) her life and in
other unique ways in the lives of countless others. At least it seems
safe to assume that what might be called God’s courtship of each
person involves a complex and delicate interplay among things in
that person’s life that are chosen and things that are not. But it is
not obvious that there is one situation or set of situations that would
be maximally conducive to bringing a given person into loving union
with God. A scenario truly maximally conducive toward this end
would be one that removes the person’s freedom to resist, but then
the goal of genuine loving communion is lost. Surely a person who
is the recipient of in nite and unconditional love should be more
free, not less so. If God aims to win a person’s heart and will, his
consent to what God wills, perhaps any situation that is an occasion
for this consent will serve as well as any other. One of the strengths
of Aristotle’s virtue theory of morality is that it allows for many
ways of living a virtuous life, and I believe it is one of the strengths
of the virtue analysis of moral perfection that it allows for many
ways of arriving at eternal happiness.

CONCLUSION

Treating God’s goodness as a feature primarily of his will, as virtue

theories suggest, rather than as a feature primarily of his intellect,
sustains a more positive understanding of divine freedom. Richard
Swinburne and others de ne perfect freedom negatively, as freedom
from external coercion and from forces that could interfere with or
prevent what one wills. This is surely an important aspect of
freedom, but it neglects the Thomistic insight that the will is
appetitive, that is oriented toward the good. In perfect freedom,
then, the will seeks the highest good and perfectly possesses what it
wills.31 While Thomas does not use this language, it is surely
consistent with his view to say that the rst principle of morality is
the law of love. God loves his own being and goodness, and wills
everything else ‘for the sake of his goodness’ as Thomas puts it. But
nite persons (angels and humans) are created as ends in
themselves (as intrinsically good), not as a means to divine
goodness, since God already possesses that end (his own goodness)
and cannot possess it any more perfectly. Rather, creation is ordered
to God’s goodness in the sense that it re ects and reveals his
goodness, beauty, and wisdom, and has its end in God.

What the apostle Paul says to the early Christians is also the
purpose of all created things, that they ‘exist for the praise of his
glory’.32 St Thomas adds that human persons arrive at this end only
by imitating the moral perfection of God, loving him above all,
loving ourselves as we ought, and loving others as God does, for
their own sake. Wishing others well, in this perspective, means
especially wanting them to obtain the highest of all goods,
friendship with God. Some criticize Christians for a perceived lack
of attention to the material and economic needs of those around
them, and certainly these are essential to a person’s welfare. But one
cannot blame Christians for working even harder to bring others
closer to God, to a good that is unsurpassable, imperishable, and can
never be taken away. Using St Paul’s language, C. S. Lewis spoke
eloquently about the ‘weight of glory’, the in nite value and eternal
destiny of each human being that requires us to treat others with
respect, and even with a form of reverence.33 Since divine goodness
is, among other things, a model for human goodness, our attitudes
toward others should be more like those of God, ready to o er
mercy and forgiveness, to share the burdens and sorrows of those
around us, to reach out to those in need, and to defend the weak
and despised. While our e orts may not always succeed, believers
hope in a God who uses all their e orts to accomplish his will and
whose love for each human being is beyond our imagining.

NOTES

I am grateful to Jorge Garcia and to the editors of this volume,

Michael Rea and Thomas Flint, for their many valuable
comments and suggestions on earlier versions of this chapter.
Eileen Sweeney and Oliva Blanchette provided helpful
discussions of the Thomistic understanding of creation. Any
aws or shortcomings are, of course, my own responsibility.

1. Thomas V. Morris, Anselmian Explorations: Essays in

Philosophical Theology (Notre Dame: University of Notre Dame
Press, 1987), 20.
2. See Peter Geach, Providence and Evil: The Stanton Lectures
1971–2 (London: Cambridge University Press, 1977), 58. ‘God,
like some grand master of chess, can carry out his plan even if
he has announced it beforehand…. No line of play that nite
players may think of can force God to improvise: his
knowledge of the game already embraces all the possible
variant lines of play, theirs does not.’ Geach also points out
that God can have strongly justi ed beliefs about future
contingents because he knows each person’s character
thoroughly and knows the way things are presently tending.

3. See William Rowe, Can God Be Free? (Oxford: Clarendon,

2004) and his critique of the standard theistic replies in his
article ‘Freedom, Divine’ in E. Craig (ed.), Routledge
Encyclopedia of Philosophy (London: Routledge, 1998);
<http://www.rep.routledge.com/article/K025SECT3>,
accessed 15 January 2008.

4. Rowe, Can God Be Free?, 128.

5. This rationality requirement is more complicated in the case of

nite persons, since presumably the alternatives must be ones
that the agent knows about (or perhaps should know about)
and that he or she has considered (or should have considered).
For God, the requirement can be stated more simply.

6. Thomas Flint points out that this is true only if God knows the
truth-values of counterfactuals of freedom, i.e. propositions of
the form: In circumstances C, agent S freely chooses action A. As
 we have seen, Peter Geach (and some others) denies that God
can know with certainty the truth-values of such propositions.

7. For an introduction to this issue see Robert M. Adams, ‘Must

God Create the Best?’, Philosophical Review 81 (1972), 317–32,
and Alvin Plantinga, God, Freedom and Evil (New York: Harper
& Row, 1974), 89–91. In considering Gaunilo’s objection to
Anselm’s ontological argument for God, Plantinga notes that
some descriptions of the form ‘the greatest X’ are logically
incoherent, as when X denotes a being with properties that
have no intrinsic maximum. His examples are ‘the greatest
number’ and ‘the greatest island’. The phrase ‘the greatest
being’ escapes this problem, since the qualities such a being
would have—knowledge, power, and goodness—do have
intrinsic maxima; e.g. to know (immediately and infallibly) the
truth-value of every proposition is to have maximal
knowledge, and to be able to bring about any logically
possible state of a airs (or something along these lines) is to
have maximal power.

8. Here I assume that compatibilism is false. Compatibilists claim

that a person can be morally responsible for one of her actions
even if that action is causally necessitated by a chain of prior
events that originates outside her.

9. Plantinga, God Freedom and Evil, 7–57. The key claim is on p.

56: ‘It’s possible that every world containing as much moral
 good as the actual world, but less moral evil, [is such that]
God could not have created it’.

10. John Henry Cardinal Newman, Lecture 8, in Certain Di culties

Felt by Anglicans in Catholic Teaching, i. 239–40;
<www.newmanreader.org/works/anglicans/volume1/lecture
8.html>, accessed 9 October 2007.

11. St Augustine, Sermon 169, 13 (PL 38. 923).

12. A clear and succinct summary of these varieties of moral

intuitionism can be found in Robert L. Frazier, ‘Intuitionism in
Ethics’, in E. Craig (ed.), Routledge Encyclopedia of Philosophy
(London: Routledge, 1998);
<http://www.rep.routledge.com/article/L041>, accessed 14
January 2008. According to Frazier, H. A. Prichard, Joseph
Butler, and W. D. Ross are intuitionists of the rst sort,
Richard Price, Thomas Reid, and Henry Sidgwick examples of
the second type of intuitionist, and G. E. Moore a
representative of the third. Further, while Price and Moore
think of intuition as an intellectual capacity, Prichard and
Thomas Reid describe it as closer to a type of perception.

13. See Janine Idziak (ed.), Divine Command Morality: Historical and
Contemporary Readings (New York: Edwin Mellen, 1979),
especially the essay by Philip Quinn. Also see Robert M.
Adams, ‘Divine Command Metaethics Modi ed Again’, Journal
of Religious Ethics 7 (1979), 71–9, and William P. Alston, ‘What
Euthyphro Should Have Said’, in William Craig (ed.),
 Philosophy of Religion: A Reader and Guide (New Brunswick:
Rutgers University Press, 2002), 283–98.

14. Thomas V. Morris, ‘Duty and Divine Goodness’, American

Philosophical Quarterly 21 (1984), 264.

15. Thomas V. Morris, Our Idea of God: An Introduction to

Philosophical Theology (Notre Dame: University of Notre Dame
Press, 1991), 62.

16. Axiology is the branch of philosophy that deals with questions

about value, and includes ethics (the good for persons),
politics (the good for communities or nations), and aesthetics
(the nature of beauty). Morris uses the term here to indicate a
kind of goodness of being that is more than just moral
goodness and (it seems) more than just duty-ful llment.

17. Ibid. 62–3.

18. Earl Conee, ‘The Nature and Impossibility of Moral Perfection’,

Philosophy and Phenomenological Research 54 (1994), 815–25.

19. Colin McGinn, ‘Must I Be Morally Perfect?’, Analysis 52 (1992),

32–4. McGinn’s answer is ‘yes’ since he interprets moral
perfection as a matter of having no moral defects (never
violating one’s moral duties), which is a goal that is possible to
achieve (at least in principle).

20. T. M. Mawson, Belief in God: An Introduction to Philosophy of

Religion (Oxford: Clarendon, 2005), 59.

21. Ibid.67.
22. On the other hand, if God’s duties to each individual might
prevent him from maximizing overall value (or from achieving
the relevant consequentialist objective), the duty model could
result in a di erent outcome than the consequentialist model.

23. Richard Swinburne, The Coherence of Theism (Oxford:

Clarendon, 1977), 202.

24. For a fuller explication and defense of this virtue theory see J.
L. A. Garcia, ‘Interpersonal Virtues: Whose Interest Do They
Serve?’, American Catholic Philosophical Quarterly 71 (1997),
31–60.

25. Christian theologians may of course appeal to the doctrine of

the Incarnation as a counterexample to this claim. But
presumably even though Jesus as man experienced emotions
and desires and so could exemplify the virtues of courage and
temperance, it would be odd to say that he exempli ed these
virtues as God. There is an extended sense of the term in
which one might describe God as exhibiting humility in
deigning to come among us as a man (in the Incarnation), but
this strikes me as an analogical use of the term.

26. St Thomas Aquinas, Summa Theologiae I q. 19 a. 3.

27. Unless having duties is compatible with necessarily ful lling

them, God does not have a duty to will the good of each
person. On Thomas’s view, God’s perfect goodness entails that
 it is logically impossible for him to create a person without
willing that person’s good.

28. William P. Alston, ‘Some Temporarily Final Thoughts on

Evidential Arguments From Evil’, in Daniel Howard-Snyder
(ed.), The Evidential Argument from Evil (Bloomington: Indiana
University Press, 1996), 311–32.

29. ‘We know that all things work together for those who love God,
who are called according to his purpose’. Romans 8: 28.

30. Thomas Flint, in editorial comments on an earlier version of

this chapter, February 2008.

31. Such freedom wills the highest good (namely, God) and
necessarily possesses this good, as being one and the same
with it.

32. Ephesians 1: 12.

33. C. S. Lewis, ‘The Weight of Glory’, in The Weight of Glory and

Other Essays (New York: HarperCollins, 1949, 1976, 1980),
25–46.
 PART III

GOD AND CREATION

 CHAPTER 11

DIVINE ACTION AND EVOLUTION

ROBIN COLLINS

I. INTRODUCTION

ALTHOUGH many issues relate to divine action—such as the doctrine

of creation, God’s relation to human beings, God’s relation to time,
the intrinsic nature of God, and the like—the issue of divine action
as it speci cally relates to evolution can be divided into two major
issues: (1) that of understanding the ways in which God might have
in uenced the evolutionary process, if at all, and (2) that of
attempting to gain insight into God’s ultimate purposes in creating
the universe and life by means of an evolutionary process. This
chapter will primarily focus on (2), an issue that has largely been
neglected by philosophers writing on divine action. This issue is
particularly important given that the broadly evolutionary picture—
namely that the universe, life, and human culture have gone
through a long developmental process—is now beyond reasonable
doubt and yet at the same time constitutes a drastic shift in our
understanding of the universe.
 The literature on issue (1) has largely focused on whether it is
possible to provide a satisfactory ‘non-interventionist’ account of
God’s interaction with the world.1 The typical requirement for a
non-interventionist account is that God’s interaction should not
‘break’ the laws of nature, though what exactly this means is
typically not spelled out. One widely discussed proposal along these
lines is that God in uences the course of the world by determining
the outcome of some, if not all, quantum events. Purportedly, this
does not break the laws of nature since the laws of physics do not
determine which quantum events will occur.

One major problem with the literature is the lack of well-

developed, convincing arguments for rejecting an ‘interventionist’
account of the laws of nature. Robert Russell (2006: 584), for
instance, claims that a major problem with interventionism is that
‘it suggests that God is normally absent from the web of natural
processes, acting only in the gaps that God causes’. The force of this
objection is highly questionable, however, since it is unclear why
God could not act both through natural processes and by means of
occasional divine interventions, something Christians have
traditionally held. Typically, little further argument is o ered in
support of this, and related, objections to interventionism. In any
case, philosopher William Alston (1994) has carefully examined the
major objections commonly o ered to interventionism and
concludes that he can nd no reasons to reject this view. At least for
those who hold a traditional view of God in which God has the
ability to intervene, I agree with Alston that the reasons typically
o ered against interventionism are unconvincing, unless one
already makes highly controversial assumptions about God’s
purposes in creation.

What often fails to be explicitly recognized is that whether we

should seek for non-interventionist accounts of divine action, and
what constraints we should put on such accounts, largely depend on
the view we have of God’s purposes in creation. If, for example, one
of God’s primary purposes was to display his power over nature,
then a natural way for God to do this would be for God to suspend
the normal operation of nature. An advocate of such a view,
therefore, would likely have no problem with the claim that God has
intervened in cosmic history on a regular basis. On the other hand,
if one of God’s primary purposes in creating the world was to allow
creation to ‘make itself’, a position taken by most of the current
major writers on divine action and evolution (see below, sect. II),
then one would at most allow very limited divine intervention. Or,
if one thinks that God creates the world by an evolutionary process
to secure an epistemic distance between God and humans, as some
have proposed (see below, sect. II), then one might be happy with
divine intervention, as long as it could not be detected. Accordingly,
once we have some understanding of what God’s purposes for
creation might be, then we can ask whether a particular mode of
divine action will help ful ll that purpose or not; without such an
understanding, the only constraints for Christians seem to be
Scripture, Church tradition, and logical coherence, constraints the
traditional interventionist accounts seem to meet. As Keith Ward
(1990: 269) notes, God’s ultimate purpose for creation ‘provides the
rationale for all particular Divine actions which continually shape
and direct the temporal process of the universe’. Thus, this chapter
will be devoted to addressing this prior and key question of what
God’s ultimate purposes might be for creating the world,
particularly focusing on what God’s purpose might have been in
creating the world via a seemingly partly chance-driven
evolutionary process.

Throughout this chapter, I take it as a well-established theory

that life on earth arose through the process of biological evolution,
by which I mean the process of descent with modi cation from the
rst cell. I will leave open, however, the question of whether this
process was guided by God.

The Issue

To many, evolution provides evidence against the existence of God,

and for others, at least a theological perplexity. As theologian John
Haught (2006: 702–3) has eloquently summarized the problem:

It is the absence of purposive design and the presence of accident

that seem to rule out the existence of God…. Evolution requires
enough time for a su ciently large number of minute random
variations to supply the mindless process of natural selection with
adaptable outcomes. The fact that so much time is required for this
unwieldy epic to transpire, and so much death occurs along the
way, and so many mistakes and monstrosities appear, only to be
discarded renders evolution all the more devoid of purpose
apparently. Any creator who would ‘fool around’ so ine ciently for
billions of years in order to produce living and thinking beings
seems much less competent than the most mediocre human
engineers.

This perplexity can be broken down into two questions: (1) Why did
God create a world that had to undergo a very long developmental
process to give rise to life, and then conscious, moral agents, instead
of creating a world that was fully formed from the beginning? And
(2) Why did this developmental process, whether or not it was
guided by God, involve so many apparent accidents and chance
events, which has led to so much su ering and death over millions
of years? Doesn’t this seem contrary to the character of an all good,
all loving God who has a providential purpose for creation?

Although one could respond to these questions by leaving it a

‘mystery’ as to why God created the universe in this way (similar to
the strategy of what is often labeled a ‘defense’ in the case of the
evidential problem of evil), if overused such a strategy undermines
the claim that the theistic hypothesis provides a fruitful basis for
understanding the existence and nature of the universe and human
experience. In any case, I will pursue another strategy in this
chapter, that of attempting to o er a plausible speculation as to why
God might have created the world through an evolutionary process.
Apart from apologetic purposes, engaging in such speculation, I
believe, is important in its own right for gaining a clearer
theological and philosophical understanding of what God’s purposes
might be for the universe and ourselves.

II. MAJOR CURRENT ACCOUNTS OF GOD’S PURPOSES

The Autonomy of Creation Explanation

In the last twenty years, the most common answer given by those
writing in the area of science and religion as to why God used a
partly chance-driven evolutionary process to create the world is that
such a process is required for creation to be truly independent of
God. Scientist-theologian John Polkinghorne, for example, says that
he believes the only possible solution to the wastefulness of
evolution and other sorts of natural evil ‘lies in a variation of the
free-will defence, applied to the whole created order. In his great act
of creation I believe that God allows the physical world to be itself,
not in Manichaean opposition to him, but in that independence
which is Love’s gift of freedom to the one beloved…. The cosmos is
given the opportunity to be itself (1989: 67). On the next page,
Polkinghorne goes on to state that ‘God accords to the processes of
the world the same respect he accords to the actions of humanity’
(ibid. 68). Elsewhere Polkinghorne (1998: 14) summarizes his free
process defence as the claim that ‘a world allowed to make itself is
better than the puppet theatre of a Cosmic Tyrant’. Polkinghorne’s
free-process defence, therefore, is an attempt to extend the free will
theodicy—the idea that God has given his creatures free will and
thus must allow them to do evil—to creation as a whole.2

John Haught, currently the leading theologian writing on

evolution and Christian theology, presents a similar explanation.
According to Haught (2003:168),

an instantaneously nished universe, one from which our present

condition of historical becoming and existential ambiguity could be
envisaged as a subsequent estrangement, would in principle have
been only an emanation or appendage of deity and not something
truly other than God. A world that is not clearly distinct from God
could not be the recipient of divine love. And an instantaneously
completed world could never have established an independent
existence vis-à-vis its creator.

Michael Murray criticizes Haught’s claim that the universe could not
be truly independent of God if God created it in an instant. As
Murray (forthcoming, ch. 6) rightly points out, if an artist creates a
fully-formed painting, the painting is still a truly distinct entity from
the artist. Although Haught is not explicit about what exactly he
means by autonomy, perhaps a more charitable interpretation of
Haught’s idea of independence is that the universe be allowed to
‘make itself’. As Haught (2000: 41) says elsewhere, ‘A world given a
lease to become more and more autonomous, even to help create
itself and eventually attain the status of human consciousness, has
much more integrity and value than any conceivable world
determined in every respect by a “divine designer”.’ Under this
interpretation, Haught could be seen as extending both the free will
theodicy and the soul-making theodicy to the universe itself (as
Polkinghorne also seems to do). In the soul-making theodicy, God
creates a universe in which natural and moral evil can occur so that
human beings can develop a freely formed virtuous character by
acting virtuously in response to evil. If God simply created our
characters fully formed, then we would not have been allowed to
‘make ourselves’. Indeed, the soul-making theodicy can be seen as
an extension of the free will theodicy: it could be argued that one
major reason that free will is valuable is that it gives us the
opportunity partly to choose our own character.

One outstanding problem with Polkinghorne’s and Haught’s

explanation is that of understanding what is meant by creation
‘making itself’ or being ‘autonomous’. The whole appeal of this idea,
it seems, arises because the metaphors of ‘making itself’ and
‘autonomy’ tempt us to endow nature with the sort of will and
choice we nd in ourselves. Assuming that non-human creation does
not have a will to decide its own destiny, it is hard to see what these
concepts could mean when applied to the non-human world, other
than that creation simply unfolds in accordance with the
deterministic and statistical laws with which God endowed it.
Clearly, merely to follow a statistical law is not the same as free
choice: presumably, a radioactive atom that decays does not ‘decide’
to decay but merely follows the statistical rules of quantum
mechanics.

The worry here is that both Polkinghorne and Haught are overly
anthropomorphizing nature: eliminate the semi-anthropomorphic
metaphors, and the idea of creation ‘making itself’ seems to lose
much, if not all, of its appeal. The reason is that it is unclear what is
supposed to be good about creation making itself, once we de-
anthropomorphize nature. In the case of human freedom, we have a
strong intuition that moral responsibility requires free will and that
moral agency is a great good in and of itself. Further, arguably,
beings without free will could not authentically love God (or even
one another) if God determined all their choices. Presumably,
however, non-human creation neither has moral agency nor does it
love God (except, perhaps, some higher non-human animals), or at
least not unless one adopts a radically di erent view of nature than
delivered by modern science: namely, a view in which non-human,
non-higher-animal creation does have a will and can make choices.
Neither Haught nor Polkinghorne, however, advocates such a view.3

One response to the above criticisms is to argue that God’s

creating the universe and life through an evolutionary process is
more consistent with the kenotic, self-emptying, non-compelling
character of God’s love as revealed in Jesus, even if it is not required
by it. As Lutheran theologian George Murphy (2003: 372) states, ‘if
the character of the true God is shown to us in the saving work of
the cross, then we may expect that all of the divine activity will be
consistent with this character. We should not be surprised if this
God in his creative and providential work “makes himself
nothing”…and “humbles himself” rather than overwhelm the world
with arbitrarily exercised power.’ The problem with this response is
that it is once again di cult to see how, without a highly
anthropomorphic view of nature, it would be in any way
inconsistent with God’s character as revealed in Jesus to create the
universe in an instant with its future development determined by
natural law. For instance, it is not in any way inconsistent for one to
treat other human beings in a loving way that respects their
autonomy, while being a ‘tyrant’ with one’s car or computer. At
best, this ‘kenosis’ response seems to provide only a weak reason for
God’s creating by means of an evolutionary process—namely, that
everything else being equal, we would expect creation to re ect the
creator’s character, much as we might expect an artwork to re ect
the character of the artist.

Divine Hiddenness as an Explanation

Another explanation for God’s using an evolutionary process is that

this is necessary to keep God’s existence from being too evident. If
God’s existence were obvious, that would take away from human
free choice, much as seeing a police o cer in the rear-view mirror
takes away—or at least severely reduces—the choice of obeying the
speed limit for its own sake. As Murray points out, one major
problem with this view is that until the last two centuries, the
theory of evolution did not exist, and people saw what they
considered powerful evidence of design all around them. Yet,
presumably they had the freedom to believe or disbelieve in God.
Further, Christians should at least be cautious concerning this
argument since it seems to be in tension, if not con ict, with various
statements in the New Testament. For example, the Apostle Paul
states that God’s existence is evident from the things that are made
(Romans 1: 20) and the book of James says that ‘even the demons
believe—and shudder’ (2:19, NRSV). Thus, even if one is convinced
that God exists, much room still seems to be left open with regard to
believing in God, in the sense of trusting, relying on, and
committing oneself to God.

The Chaos to Order Proposal

After reviewing and rejecting as seriously awed several other

proposals for why God created a universe via an evolutionary
process that results in so much animal su ering, Michael Murray
(forthcoming, chs. 5 and 6) proposes as a viable possibility the idea
that bringing about a universe from a state of chaos to order via
lawlike means is intrinsically valuable (ibid. ch. 6). One problem
with this idea is technical: from the perspective of the second law of
thermodynamics, the early universe was in a very low entropy state
and hence actually highly ordered, at least by the technical
de nition of order given by statistical mechanics. (Statistical
mechanics provides the standard physicist’s understanding of the
laws of thermodynamics.) One could circumvent this problem,
however, simply by claiming that what is intrinsically valuable is for
God to bring about a universe by a developmental process that
partly involves seemingly chance occurrences. The other problem is
justifying such a claim. Murray supports this claim by noting that
there is a signi cant thread of past Christian thinkers (such as St
Augustine) who claimed that it would be grander for God to create
the world in a developmental way instead of fully formed. This
claim seems correct as far as it goes: it does seem to be a greater
accomplishment to create a universe that ‘forms itself’ into stars and
galaxies from an initial reball than to create a fully formed
universe. Arguably, a universe that forms itself also would be
aesthetically richer since it would not only contain a greater variety
of states, but these states would coherently build on each other.

Although a combination of the value of expressing God’s

grandeur and aesthetic values could plausibly be thought to be
su cient reason for God to create a universe that required billions
of years to develop, one might doubt (1) whether it could
adequately explain why seemingly chance processes have played
such a major role in the universe’s evolution and (2) whether it is
su cient to o set the seeming disvalue of the su ering of non-
human sentient life during the evolutionary process. These doubts, I
believe, should be su cient to motivate us to search for additional
reasons, whether or not such reasons ultimately can be found.
 III. MY OWN PROPOSAL

My own highly speculative proposal begins with the claim that the
evolutionary process allows for certain types of interconnection
between humans and non-human creation that are potentially of
signi cant value. I do not claim that these interconnections provide
the sole reason for God’s creating by means of a seemingly chance-
driven evolutionary process, only that they provide one reason. We
will rst explicate what these interconnections are, and then present
some reasons for thinking that they are of value. The three types of
interconnections that we will consider are what I will call emergent,
ancestral, and redemptive interconnections.

Before proceeding, however, a brief explanation should be given

about what is meant by ‘interconnection’. To begin, an
interconnection is a special sort of relation between persons or
between persons and non-personal aspects of creation. Why
hypothesize the existence of such a special relationship? The basis
for this hypothesis is that people commonly claim to feel deeply
connected to other human beings, such as their parents, their
spouses, or someone who has greatly helped them in times of
su ering and hardship. Further, many people, particularly people in
indigenous societies, claim to feel a profound connection to the
surrounding earth and animals. This common human experience is
the basis for claiming that there is this sort of special and signi cant
relationship which in common discourse is often called a
‘connection’.
 People commonly experience the persons/entities that they are
connected to, and the connections themselves, as being ‘part’ of
their selves. Thus, it could plausibly be argued that this idea of
interconnection presupposes that the self is partly constituted by its
relations to other entities. Since one can be more or less connected
with another person or entity, this in turn presupposes that the self
does not have de nite boundaries: some relations can be a deeper
part of the self than others. Why make this presupposition? Once
again we can appeal to experience. People commonly experience
their relations with other persons, or even parts of non-human
creation, as both profoundly important and as being in some way
deeply signi cant to what they are as human beings. Further, the
loss of one of these relations is often experienced as a loss of an
aspect of one’s self. For example, the loss of a loved one—
particularly a spouse—is often experienced as a loss of some part of
one’s own self. Once again, indigenous peoples often claim this sort
of thing about the land. Even if this talk of another entity or our
relation to it as being part of one’s very selfhood is only
metaphorical, the metaphor is still picking out some special relation
that people nd of great signi cance, which is all that we require
for the argument in this chapter.

Emergent Interconnection

An emergent interconnection refers to a sort of interconnection that

occurs under the hypothesis that, at least in part, the human body
and soul emerge out of the basic materials of the universe. I de ne a
weak emergent connection as an emergent connection that occurs if
the human body is the result of an evolutionary process, but the
human soul is directly created by God. On the other hand, I de ne a
strong emergent connection as an emergent connection that occurs if
the human soul itself—including its libertarian agency—is formed in
some way out of what is already present, in nascent form, in the
universe itself. If this latter a view of the human person is correct,
then our consciousness and libertarian agency would be more
deeply interconnected with the universe than under a traditional
substance dualist account of the soul. Further, if libertarian agency
emerges out of the matter in some way, then it seems that
indeterminism would have to be built into the very fabric of the
cosmos. Given that a strong emergent connection results in greater
value than a weak emergent connection, this would provide one
reason for why God created the universe by means of an
indeterministic evolutionary process instead of a deterministic
evolutionary process. We will explore the value of emergent
connections more below.

Although I believe emergent interconnections are important, I do

not want to promote a physicalist view of the mind, since I believe
that there are powerful arguments for some form of dualism in
which the self is partly a separate entity from the brain, whether
that be a classical form of substance dualism, an emergentist form in
which the partially independent soul is generated from the brain, or
some sort of Thomistic view.4 Exactly how one should view the
relationship between the mind and the brain is a very deep issue
that we cannot explore here. Instead, I simply note that the
existence of even a strong sort of emergent connection seems
compatible with some forms of dualism in which the human person
is partly separable from the body.

Ancestral Interconnections

Emergent interconnections could be realized by God creating a

world with fully formed animal and plant life with human and
animal souls somehow arising out of matter. In such a world,
however, a certain sort of connection would be missing, what I will
call an ancestral connection. An ancestral connection occurs when
one being shares a common ancestor with another—e.g. according
to the theory of evolution, human and apes both arose by common
descent from the primate group called prosimians. Many thinkers
writing in the areas of evolutionary theory and ecology intuit a
special value in the fact that we are ancestrally connected with
other living organisms. For example, Stanford University
evolutionary biologist Joan Roughgarden (2006: 18) writes: ‘Our
material continuity with the rest of living creation is not a threat to
Christian beliefs. Just the opposite…Evolution’s discovery of a single
tree of life extends a Christian view of the body and family beyond
humans out to all of living creation.’
Redemptive Interconnections

By ‘redemptive interconnection’ I mean the sort of interconnection

that would occur between humans and creation if, through God’s
grace, human beings help ‘redeem’ the world by helping make God
present in creation in such a way that creation more fully
participates in the life of God.5 To explain this more, we will rst
need to sketch a Christian eschatology that I believe is suggested by
the New Testament.

Many New Testament scriptures speak of the ultimate ful llment

or redemption of creation. Romans 8: 21, for example, tells us that
‘creation itself will be set free from its bondage to decay and will
obtain the freedom of the glory of the children of God’ (NRSV).
Similarly, other scriptures speak of God’s ultimate purpose being
directed toward the redemption of all creation. In Ephesians 1: 10,
this ultimate purpose is to ‘gather all things in him [Christ], both in
heaven and earth’; in Ephesians 4:10 it is for Christ to ‘ ll all
things’; in Colossians 1: 20 it is to ‘reconcile to himself all things,
whether on earth or in heaven’; and nally, in 1 Corinthians 15: 28
it is for God to be ‘all in all’. These scriptures suggest, if not imply,
that God’s ultimate purpose for the material cosmos is that it
become a full participant in the divine life.

This idea of the entire created order ultimately participating in

the life of God is particularly emphasized in Eastern Orthodox.
According to standard Eastern Orthodox theology, this complete
participation of humans and creation in the divine life is understood
as participation in what the Orthodox call the ‘energies’ of God in
contrast to the essence of God (Lossky 1976: 74–5, 97–101, 133–4).
For the Orthodox, the energies of God refer to the life of God—that
is, ‘God in his activity and self-manifestation’ (Ware 1976: 22)—
whereas the essence of God refers to God’s innermost self, which is
forever inaccessible to us. Using this distinction, Orthodox
theologians claim to be able to a rm the eventual complete
participation of redeemed humanity and creation in the divine life
while at the same time excluding ‘any pantheistic identi cation
between God and creation’ (ibid. 23).6

Within the context of the New Testament, the passages cited

above suggest that this redemption will take place through the
actions of human beings. In the New Testament, Christians are
referred to as the ‘body of Christ’, which suggests that one primary
way in which Christ acts in the world is through human beings, just
as we act through our bodies. Hence if God’s purposes are to redeem
all creation—as is clear, I believe, from the scriptures cited above—
it makes sense that it will be at least in part through his ‘body’,
which consists of human beings. Our redemption comes to us rst,
and then spreads out to all creation through us.7 In any case, I agree
with philosopher and theologian Keith Ward (2001:165) when he
states that ‘I can think of nothing more important for the Christian
faith in our day than to recover the truly cosmic sense of
redemption that was characteristic both of the biblical writings and
of the Church Fathers. Redemption will not be seen as a saving of a
few human beings from destruction of one small planet. It will be
seen as a reconstituting of the whole cosmos in the presence of God,
in a more glorious form.’

How will this redemption work? One can only speculate. I

speculate that it will be through both the action of God through
human beings and the activation of subtle orders in nature of which
we are presently unaware, though this particular speculation is not
essential to my overall proposal in this chapter. Two analogies
might make this more plausible. First, consider the case of
metamorphosis, in which one organism undergoes a radical
transformation in its morphology, such as a tadpole becoming a frog
or a caterpillar becoming a butter y. If we had not seen this
happen, we would have never predicted it; clearly it involves subtle
processes that we do not understand very well yet. Whether
metamorphosis can be understood by current chemistry or requires
some new sort of science—such as the non-local ‘morphic elds’
suggested by biologist Rupert Sheldrake (1988)—is unclear. The
suggestion is that in analogy to cases of metamorphosis, the
universe itself will undergo a radical, currently unimaginable
transformation and yet still be the same universe, just as a
caterpillar could be thought to remain the same entity through its
radical transformation to a butter y; the caterpillar does not die and
is then replaced by a butter y, but instead transforms into a
butter y, implying a retention of its identity.
 The second analogy is the development of life and nally
conscious beings. If we imagine ourselves as disembodied physicists
existing in the rst minute of the big bang, we probably never
would have predicted the eventual emergence of life and
consciousness using only our current knowledge of physics. We
probably would have thought the idea was absurd. Finally, the
revolutions of quantum theory and general relativity have taught us
that nature is full of surprises and matter is much more mysterious
than we thought; consequently, one must be extremely careful when
extrapolating the future of the universe based on our current
understanding of physics.

Although not essential to my argument concerning redemptive

interconnections, I further speculate that the universe’s being
‘subject to decay’ is the result of the universe’s operation being
governed mostly by law, instead of God’s being more fully involved
in its operation beyond sustaining its laws. To elaborate on this
idea, consider the second law of thermodynamics. The second law
states that within a closed system, the entropy of a system always
increases. Within statistical mechanics, this notion of entropy is
given a precise de nition in terms of probabilities over what
physicists call ‘phase space’. In the standard explanation of the
second law, the claim that the entropy of a closed system must
always increase is reduced to the claim that the system proceeds
from a less probable macroscopic state to a more probable
macroscopic state. The example of perfume in a bottle nicely
illustrates this idea. If one opens a bottle of perfume, the molecules
eventually will evaporate and become distributed throughout the
room. In order for this to happen, no special initial distribution of
the positions and velocities of the molecules in the bottle is
required; virtually all initial distributions would yield the same
result. In contrast, unless the perfume molecules had just the right
enormously improbable con guration of positions and velocities in
the room, the perfume has a virtually zero chance of returning to
the bottle (within any reasonable period of time) apart from some
intervention such as the room being cooled causing the perfume to
condense. This leads to the following suggestion: without some sort
of external intervention or exactly the right initial conditions, the
laws of probability dictate that any su ciently complex closed
system that both obeys laws similar to those in our universe and
contains systems of enormous complexity, will tend, however
slowly, to move from order to disorder; or at least this occurrence is
highly probable. Of course, it might be possible for God to
circumvent this movement from order to disorder by creating a
universe with radically di erent laws or initial conditions, but we
cannot say what other goods might be lacking in such a universe.

Under this idea, therefore, the fact that the universe is ‘subject to
decay’—and hence su ering and death occur in it—is because it is
governed by a set of laws, whether those laws are deterministic or
indeterministic. Only if God becomes directly and constantly
involved with the universe (beyond simply sustaining it) could this
process of decay be stopped; even by God’s carefully choosing the
universe’s laws and initial conditions, avoiding some sort of
imperfection might be impossible.8 My suggestion, therefore, is that
God has left it up to human beings to help bring about this further
divine involvement with the universe that is necessary for it not to
be subject to decay.

My proposal is that by our rst sharing in the divine life through

Christ, this life is able to spread throughout all creation and hence
reverse the decay. This proposal ts well with Romans 8: 20–1,
since this passage suggests that God created nature to be ‘subject to
frustration’ so that it would participate in the new life of redeemed
humanity. According to the NRSV version of this passage (which
does not relevantly vary among translations), ‘the creation was
subjected to futility, not of its own will but by the will of the one
who subjected it, in hope that creation itself will be set free from its
bondage to decay and will obtain the freedom of the glory of the
children of God’. Notice that the passage suggests that God
subjected it to futility (‘by the will of the one who subjected it’) for
the purpose that it be set free, as indicated by the phrase ‘in hope
that’.

What will this sharing in the divine life be like? Beyond the
claim that the universe would no longer be subject to decay, one
can only speculate. Perhaps the fabric of the universe would be in
perfect harmony with the wills and thoughts of resurrected human
beings. Further, perhaps all su ciently sentient animals will
become fully conscious in some way, taking on new forms that are
in continuity with their current bodies, but which are also radically
di erent, with current forms of metamorphosis being only a
foretaste. It might even be the case that the universe itself gains a
‘soul’ of some sort, with which human beings could be in
communion.9

Why think this scenario is at all plausible? First, such a scenario

is suggested by the Apostle Paul’s statement that creation itself will
share ‘in the glorious liberty of the children of God’. More
importantly, however, given that this sort of scenario would result
in a richer and more valuable universe, then theists would have
good reason to believe that this is God’s ultimate destiny for the
universe, unless we had some reason to believe it was either
contrary to some well-established knowledge about the nature of the
universe or in some way impossible for God to bring about, neither
of which we have good reason to believe. In any case, it seems
plausible to suppose that in such a transformed creation, there
would be a deep intercommunion between non-human creation and
human beings. Further, the interconnections hypothesized above
would become particularly valuable, as now they would involve
interconnections between conscious beings. For example, these
redeemed non-human creatures would have a connection of
appreciation for being redeemed from decay. (See more below for
the value of this sort of interconnection.) Finally, such a scenario
would provide additional resources for understanding why an all
good, loving God would create a world that would involve so much
animal su ering and pain. This issue becomes particularly pressing
with regard to the redemptive status of highly evolved non-human
hominids such as Neanderthals and homo erectus. The existence of
such beings—which have a form of sentience between currently
existing non-human primates and humans—really presses the case, I
believe, for including all God’s creation in God’s redemptive plan.

At this point, one might ask: couldn’t we still play a major role in
creation’s ultimate ful llment even if God created it in a perfect
state? For example, through watering and tending an oak tree from
its being a seedling to a fully mature tree, one can contribute to its
nal perfection as a fully grown oak, even though at every stage of
the oak’s development it could be said to be perfect for that stage of
development. By analogy, then, it seems that God could have
created the universe in an immature, yet perfect state, and then
have given human beings the opportunity to become co-creators
with God to bring it to ful llment.

One response to this objection is that the depth of redemptive

interconnection would not be as great. It seems to involve a deeper
interconnection actually to be the instrument by which the divine
life becomes fully present in creation, which involves a much more
radical transformation. Bringing creation from a state of ‘bondage to
corruption’ to sharing fully in the divine life would involve
contributing to the ful llment of creation in a greater way than
simply helping it come to maturity from one state of perfection to
another. To help see this, imagine that creation (or perhaps just the
higher animals) became self-conscious: creation would appreciate
our work of bringing it to ful llment more if we brought it from a
state of bondage to decay to sharing in the divine life than if we
simply helped it come to maturity, since the former involves a much
greater transition.

Summary

It should be clear from the foregoing that a world created to evolve

and eventually produce human beings in a partially chance-driven
way allows for the realization of emergent, ancestral, and
redemptive interconnections. Further, it is at least di cult to
conceive of how God could have created the world by some
alternative means and at the same time for these interconnections to
be realized to as great an extent. Thus, given that these
interconnections are of signi cant value, they provide a reason for
an all good God to have created the world in this way.

IV. THE VALUE OF INTERCONNECTIONS

Why think these connections are valuable? To answer this question,

we will begin by considering the value of interconnections between
persons. First, an essential part of love seems to be interconnection,
at least insofar as it goes beyond mere benevolence: the lover
desires to be deeply connected with the beloved, to participate in
the life of the beloved in some way. Without this interconnection,
love is in some important sense incomplete. Second, we all value
interconnection. Few of us would think that a life lived alone,
without signi cant and deep interconnections with others, would be
ideal; rather, we would think it was missing something of great
value. Third, part of traditional theism, especially Christian theism,
is the idea of the great value of communion with God and with
others in the body of Christ. This is taken as one of the major
reasons that God created other beings with free will. Finally, further
hints regarding the signi cance we attach to certain sorts of
connections abound, from adopted children looking for their
biological parents and people expending great e ort to determine
their family tree, to claims of the importance of apostolic
succession, which involves a purported mystical connection in the
body of Christ.

Our next question is: why think that the three types of
interconnections with creation elaborated above would be of value?
One answer is that such interconnections pave the way for a deeper
intercommunion with creation, an intercommunion in which
creation in some sense becomes part of what we are, and ultimately
through us is taken into the divine life. To see how this might take
place, we will begin by looking at four analogies.

The rst analogy is that of the purported value of soul-making,

as one would nd in the soul-making theodicy. This analogy will
particularly illustrate the value of redemptive interconnections.
Unlike some appeals to the universe ‘making itself’, our use of the
soul-making theodicy will not involve any anthropomorphism of
nature, at least not of the current pre-redeemed nature. As
mentioned above, the idea behind the soul-making theodicy is that
there is some great good in human beings having a major part in
determining their future character, instead of God simply endowing
them with a perfect character from the beginning. This goes beyond
the mere value of free will, since God could have created us with a
perfect character and then simply given us the choice of good over
evil, something many think was the case with angelic beings.

I suggest that the core value in being partly responsible for our
own souls is that our character becomes more fully our own in a
way that would otherwise not be possible. The reason, I suggest, is
that our ‘deepest choosing self’—that is, our self considered as the
underlying agent that makes choices—gets interwoven and
connected with those aspects of our character that we help develop
through our actions.10 Finally, by co-creating our character with
God’s grace in Christ, there is an interweaving between Christ, our
agency, and our character, so that not only does our character
become deeply our own, but in some sense through Christ God’s
character and life also become deeply our own.

Applied to creation, our choosing to co-create the world with

God from a state of ‘frustration’ to one of ‘liberty’ means that we
become interwoven and interconnected with creation, and it
becomes a joint work both of God and of us. In analogy to our
character traits becoming our own, this interconnection helps
creation become part of who we are, and so leads to an expansion of
our selves. The claims here parallel part of Karl Marx’s insights
concerning the alienation of labor: through the e orts and toil we
put into the products of our labor, they become an extension
(objecti cation) of ourselves, and further by others enjoying the
fruits of our labors, a means of sharing ourselves with others. This
important connection between our self and our labor is something
Marx saw the capitalist system as undercutting, hence resulting in
the ‘alienation’ of labor.

A second analogy for seeing how the interconnections discussed

above form the basis for a particularly deep communion between
humans and creation is that of being deeply involved in the lives of
other persons. When we are deeply involved in the lives of others in
a truly loving way, their joys become our joys, and their sorrows
become our sorrows. Likewise, it is plausible to think that our deep
involvement with creation—especially our redemptive involvement
—paves the way for its joys, pains, and su erings to become, in
some signi cant sense, our joys, pains, and su erings, and thus
through us to be taken up into the very life of Christ.

A third analogy is that of the relation we have with our own

body. Traditional Christian theology holds that even though the soul
survives bodily death, the soul is incomplete, if not radically
incomplete, without the body. Why is this? One plausible reason
seems to be that the body is ‘part’ of what we are, at least in some
signi cant metaphorical sense. A further question is: what makes
the body part of what we are? One answer is that we are deeply
interconnected in this life with our bodies: it is through our bodies
that we become conscious and interact in the world. Indeed, the
interconnections with our bodies seem to be analogous or parallel to
the emergent and redemptive interconnections described above: our
self is at least in part formed from our bodies and we in turn shape
and transform the nature of our own bodies. In analogy to how
these interconnections with our body make it part of our very self, I
propose that the emergent, ancestral, and redemptive
interconnections hypothesized above could plausibly be thought to
pave the way for the non-human creation becoming ‘part’ of what
we are. If this is correct, then it is at least plausible to hold that the
full redemption of humans will require the full redemption of the
creation itself. This gains plausibility when we consider that our
bodies are intertwined with the rest of creation, and thus it makes
sense that the full redemption of our bodies in the resurrection will
involve—and hence be simultaneous with—the redemption of the
cosmos (e.g. see Romans 8: 23). As John Haught (2003:155) has
stated, ‘as long as the universe is un nished, so also is each one of
us. Because of the intricate way each organism is tied into the
cosmic story … no living being can attain a satisfying ful llment
independently of the cosmos. Our personal redemption awaits the
salvation of the whole.’

The last analogy is that of the connection we have to our

parents, and to our relatives. People commonly perceive that
familial connections are deeply important to their own identity. One
reason might be that these connections make those related in this
way part of what we are as persons. So, perhaps the emergent and
ancestral connections with the rest of creation, particularly the
higher-level animals, help make them in some signi cant way part
of what we are; we are in some sense one family with creation. This
in turn helps pave the way for an even deeper intercommunion that
results from establishing redemptive interconnections with creation.

Finally, I propose that these interconnections gain their full

value by being taken up into conscious experience, speci cally our
conscious experience, God’s conscious experience, and if non-human
creation becomes conscious, its conscious experience. For example,
the internal interconnections and interweavings that are a result of
the universe’s long evolutionary history would move from being an
externally realized aesthetic value to an internally, consciously
realized beauty and richness. Thus, the richer the universe is—both
synchronically and diachronically—the richer we will be. If,
moreover, the higher animals, or even possibly the creation itself,
become conscious, they will also share in this richness, and
moreover, there will be a bond of appreciation for being redeemed,
something I assume is of intrinsic value. (I discuss the intrinsic value
of this sort of bond more below when I brie y outline my
‘connection-building theodicy’.) Finally, I propose, this will in turn
enrich God. As Keith Ward (2001: 158–9) notes, through creation
‘God realizes, makes actual, aspects of the divine being that
otherwise would have remained potential. Entering into relationship
and communion, and cooperating in realizing new forms of nite
value, are realizations of great values in and for the divine nature
itself.’ Similarly, George Murphy (2003, p.385) surmises that ‘by
assuming our common humanity in the incarnation, God took on
these evolutionary relationships [that we have with the animals]
and became a participant in this history [of life on earth]’.

At this point, one might ask: couldn’t the creation be part of who
we are even if God made it instantaneously and fully formed? Yes,
to a certain extent, since it would be connected to us: our bodily
existence would be dependent on the material processes around us,
and in turn we would a ect the universe. It is plausible to suppose,
however, that in this circumstance its connection with us would not
be nearly as deep: it would lack the ancestral and redemptive
interconnection discussed above. This would, I suggest, take away
from the depth of intercommunion with creation and the depth to
which it would become part of our own selves, at least given the
assumption that intercommunion is built o the right sort of
previously established interconnections. Of course, as mentioned
previously, an incomplete but non-evolutionary world could still
have similar redemptive interconnections, but it would still lack the
ancestral interconnections discussed above.

Interconnection, Original Sin, and Theodicy

Not only can one make the sort of case in the last section for
interconnections being important, but this idea of certain kinds of
interconnection being intrinsically valuable is also a highly fruitful
one in other areas. As our rst example, the importance of
interconnections makes sense of why God created human beings to
go through a developmental process, and why humans in general
are deeply interconnected and interdependent on each other,
especially through time. Not only are we each dependent on our
parents (or some equivalent group of human beings) for our
existence, and our development into adulthood, but our entire
present human culture is dependent on the choices made by
previous generations.

Further, this idea can help make sense of why God allows evil
and the doctrine of original sin. Elsewhere, I have constructed both
a theodicy and an account of original sin based on this idea, both of
which I will brie y summarize here.11 I call my theodicy the
connection-building theodicy (CBT) and claim that it o ers at least a
partial explanation for God’s allowing evil. This theodicy begins
with the assumption that positive connections between individuals,
such as connections of appreciation for being helped in times of
su ering, of being forgiven of sin, of being helped out of spiritual
and moral darkness, and the like, are of intrinsic value, a value in
which all the parties involved in the connection share. Further, it
hypothesizes that these connections of appreciation and intimacy
have the potential of being an ongoing part of one’s life for all
eternity. For example, when all things are brought to light, we will
have an ongoing appreciation for those who self-sacri cially helped
us in times of su ering, since we will always remember what they
did. Accordingly, I claim, the goodness of this connection keeps
growing, becoming of very large, if not of in nite, value, thus
outweighing the nite evils that God must allow in order for the
types of connections in question to exist.

One way of understanding this theodicy is to see it as greatly

adding to the virtuous-response theodicy, in which evil allows for
certain sorts of virtuous responses such as self-sacri cial love. One
problem with this theodicy is that the hypothesized greater goods
are the virtuous responses themselves, which are of nite temporal
duration and thus arguably not of great enough value to compensate
for the evils in the world. In contrast, the CBT postulates that these
virtuous responses give rise to further positive special relations
between two persons that become part of each person’s ongoing
selfhood. Because of this, the value of these connections is not
restricted to the time at which they were formed, but can keep
growing. The CBT, therefore, can be seen as postulating something
of potentially great extrinsic value resulting from many of our
virtuous responses, thereby signi cantly strengthening the virtuous-
response theodicy and similar theodicies, such as that advocated by
Richard Swinburne (2004: ch. 11).

Finally, the above explanation for why God used an evolutionary

process can be seen as an extension of this theodicy to the natural
world, even though the purpose of this chapter is not to address the
problem of evil per se. As an extension of this theodicy, it can be
seen to provide at least one reason for thinking that the value of the
connections that evolution makes possible outweighs the su ering
that occurs in the process: since the connections are eternal, their
value could be thought to be very large or in nite, thus outweighing
the nite su ering that necessarily result from the evolutionary
process.

This idea of the importance of interconnections also helps makes

sense of the idea of original sin in an evolutionary context, as I
argue in more detail elsewhere (Collins 2003). If we think of
original sin as a sort of negative spiritual inheritance resulting from
previous sinful acts of human beings—such as the sort of spiritual
darkness that the Apostle Paul claims has fallen on the human race
because of human evil (see Romans 1:18–21)—then the idea of
original sin makes good sense even in the context of evolutionary
theory; the existence of original sin can be seen as a logical
consequence of the spiritual interconnection of all human beings
and of their having morally signi cant libertarian free will.

V. CONCLUSION

I have argued that God’s creating human beings and other living
organisms through an evolutionary process allows for richer and
deeper sorts of interconnections between humans and non-human
creation than would otherwise be possible. I have presented reasons
for thinking that these interconnections are of signi cant value, the
main reason being that they allow for creation to become more
deeply united with ourselves, in fact so united that there exists a
deep communion between us and the rest of creation. This
communion is not only an intrinsic good, but it enriches us, since
part of this communion is creation becoming part of our very self,
and thus we consciously share in the richness of creation. As a nal
comment, it is important to note that this idea that communion with
nature is a great good and an ideal to be sought after has made
intuitive sense to many people in the past and across cultures,
although of course they did not view it from within an evolutionary
framework. For instance, consider the Western Inscription. Written by
Chang Tsai (1020–79), this was one of the most in uential writings
in China, especially for the Neo-Confucian philosophy that
dominated Chinese thought from around AD 1000 until the early
twentieth century. This text proclaims that ‘Heaven is my father and
Earth is my mother, and even such a small creature as I nds an
intimate place in their midst. Therefore that which lls the universe
I regard as my body and that which directs the universe I consider
as my nature. All people are my brothers and sisters, and all things
are my companions.’ (Quoted in Chan 1963: 497.) The modern
academic West has largely lost touch with this intuition since being
in the grip of an overly mechanical, reductive view of nature, a view
that I argue elsewhere is severely called into question by quantum
mechanics and general relativity (Collins 2006).12
NOTES

I would like to thank Thomas Flint, Michael Rea, Keith Ward,

David Schenk, John Polkinghorne, and John Haught for
comments on an earlier draft of this chapter.

1. For a good overview of this literature on divine action as it

relates to issue (1), see Tracy 2006.

2. Polkinghorne has retained this explanation in later writings.

See e.g. his 2001: 94–6.

3. In his book Faith of a Physicist, Polkinghorne (1994: 85) brie y

addresses this criticism. His reply is to suggest another reason
why God might have created a world by means of an
evolutionary process, namely that ‘only a world endowed with
its own spontaneity and its own reliability could have given
rise to beings able to exercise choice’. As he notes, this idea
bears a ‘cousinly relation’ to the idea of emergent
interconnection that I develop below (private communication).
An outstanding issue is why such an ‘emergent
interconnection’ would be of value, something I address
below.

Finally, to be fair to Haught, he does o er other reasons,

such as aesthetic reasons, for why God created a world that
develops by an evolutionary process. Nonetheless, a core
reason running throughout his writings is that a loving God
would allow the universe to make itself. As Haught 2000: 137
 says, ‘The notion of God as defenseless and vulnerable love
provides…an ultimate explanation of nature’s evolutionary
character.’ Although I think that this aspect of both
Polkinghorne’s and Haught’s explanations are subject to the
above criticisms, I think their books are valuable contributions
and well worth reading.

4. For a defense of dualism, see Plantinga 2006.

5. If extraterrestrial embodied moral agents exist in our universe,

they might also contribute to its redemption, in which case the
human contribution might be restricted to some localized
region. We will not discuss this issue here, however.

6. See David Bradshaw 2004 for lengthy philosophical and

historical discussion of how this participation is supposed to
occur.

7. Further evidence that the redemption of creation comes

through humans is given by two of the scriptures cited above:
Romans 8: 21, which says that creation shares in our
redemption, not something separate, and Colossians 1: 20,
which appears to imply that creation’s redemption is ‘through
the blood of the Cross’, which presumably only had a direct
redemptive e ect on humans.

8. To see this, suppose that one has a set of deterministic laws L

and at some time t*, a state S(t*) would be the perfect state.
Running the laws backward, one would obtain an initial state
S(t0) which, when the laws are played forward, would yield
S(t*). Thus, God could guarantee S(t*) by creating the universe
in an initial state S(t0). Furthermore, for any arbitrarily chosen
time t′, the initial state S(t0) will result in a state S(t′) at time
t′. There is no reason, however, to think that in general the
state S(t′) will be a perfect state. Of course, God will have
much exibility: God can create di erent laws L, and there
will typically be more than one perfect state at any given time.
Even with this exibility, however, it is unclear whether God
could bring it about that at every time t a perfect state is
realized merely by creating the right initial state. It is
especially unclear if the laws of nature must meet certain
other criteria, such as having enough regularity for them to
support free action by human beings and having enough
simplicity and other features for us to discover them and
hence develop technology. The closest we come in our
universe to a preprogrammed development of a complex
system is that of living organisms, such as a tree; such
development, however, is never perfect, and always requires a
compensating increase in entropy in the surrounding
environment. Thus, it at least seems plausible to suppose that
not even God could create a perfect world whose behavior was
solely determined by the previous states of the world and the
laws of nature.
 9. For a somewhat di erent proposal concerning the redemption
of all creation, see Polkinghorne 2003: ch 10.

10. Further, I suggest, this interweaving is deeper if the formation

of the character occurs over many choices, instead of in a few
choices, although this is not essential to my argument.

11. The theodicy is explicated in a yet to be published manuscript.

For the account of original sin, see Collins 2003.

12. Lest I be misunderstood, I do not claim that quantum

mechanics or general relativity show that this more organic
view of nature is correct, only that they undercut the
reductive, mechanical view of nature and indicate that at some
level nature is deeply interconnected; this at least opens the
door for taking a more ‘organic view’ of nature as a viable
possibility. (For those who read this critique in Collins 2006, it
should be noted that the explication of David Bohm’s
interpretation of quantum mechanics fails to mention the
‘classical potential’, thus resulting in a signi cant error in one
of the critiques of Bohm. This error, however, does not a ect
the argumentation in the rest of the chapter.)

REFERENCES

ALSTON, WILLIAM (1994). ‘Divine Action: Shadow or Substance?’, in

Thomas Tracy (ed.), The God Who Acts: Philosophical and
 Theological Explorations. University Park, Pa.: Pennsylvania State
University Press.

BRADSHAW, DAVID (2004). Aristotle East and West: Metaphysics and the
Division of Christendom. Cambridge: Cambridge University Press.

CHAN, WING-TSIT (1963). A Source Book in Chinese Philosophy.

Princeton: Princeton University Press.

COLLINS, ROBIN (2003). ‘Evolution and Original Sin’, in Keith Miller

(ed.), Perspectives on an Evolving Creation. Grand Rapids:
Eerdmans, 496–501.

——— (2006). ‘Philosophy of Science and Religion’, in Philip

Clayton and Zachary Simpson (eds.), The Oxford Handbook of
Religion and Science. Oxford: Oxford University Press, ch.20,
328–44.

HAUGHT, JOHN (2000). God After Darwin: A Theology of Evolution.

Boulder: Westview.

——— (2003). Deeper than Darwin. Boulder: Westview.

——— (2006). ‘God and Evolution’, in Philip Clayton and Zachary

Simpson (eds.), The Oxford Handbook of Religion and Science.
Oxford: Oxford University Press, ch. 41, 697–712.

LOSSKY, VLADIMIR (1976). The Mystical Theology of the Eastern Church.

Crestwood: St Vladimir’s Seminary Press.

MURPHY, GEORGE (2003) ‘Christology, Evolution, and the Cross’, in

Keith Miller (ed.), Perspectives on an Evolving Creation. Grand
 Rapids: Eerdmans, 370–89.

MURRAY, MICHAEL (forthcoming). Creation, Providence, and the Problem

of Animal Su ering. Oxford: Oxford University Press.

PLANTINGA, ALVIN (2006). ‘Against Materialism’, in Faith and

Philosophy 23/1: 3–32.

POLKINGHORNE, JOHN (1989). Science and Providence: God’s Interaction

with the World. Boston: Shambhala Publications [New Science
Library].

——— (1994). The Faith of a Physicist: Re ections of a Bottom-Up

Thinker. Minneapolis: Fortress.

——— (1998). Belief in God in an Age of Science. New Haven: Yale

University Press.

——— (2001). ‘Kenotic Creation and Divine Action’, in id. (ed.), The
Work of Love: Creation as Kenosis. Grand Rapids: Eerdmans, 90–
106.

——— (2002). The God of Hope and the End of the World. New
Haven: Yale University Press.

ROUGHGARDEN, JOAN (2006). Evolution and the Christian Faith:

Re ections of an Evolutionary Biologist. Washington: Island.

RUSSELL, ROBERT (2006). ‘Quantum Physics and Divine Action’, in

Philip Clayton and Zachary Simpson (eds.), The Oxford
Handbook of Religion and Science. Oxford: Oxford University
Press, ch. 34, 579–95.
SHELDRAKE, RUPERT (1988). The Presence of the Past: Morphic
Resonance and the Habits of Nature. New York: Vintage Books.

SWINBURNE, RICHARD (2004). The Existence of God, 2nd edn. Oxford:

Oxford University Press.

TRACY, THOMAS (2006). ‘Theologies of Divine Action’, in Philip

Clayton and Zachary Simpson (eds.), The Oxford Handbook of
Science and Religion. Oxford: Oxford University Press, ch 35,
596–611.

WARD, KEITH. Divine Action. London: Collins.

——— (2001). ‘Cosmos and Kenosis’, in John Polkinghorne (ed.),

The Work of Love: Creation as Kenosis. Grand Rapids: Eerdmans,
152–66.

WARE, KALLISTOS (1976). The Orthodox Way, rev. edn. Crestwood: St

Vladimir’s Seminary Press.
 CHAPTER 12

DIVINE PROVIDENCE

THOMAS P. FLINT

INTRODUCTION

THROUGHOUT history, most Christians, like most Jews and Muslims,

have consistently professed their belief in divine providence.1 Our
creator is not a distant demiurge who fashioned us without caring
how we turn out or who left us to fend for ourselves. On the
contrary, God is a creator whose omniscience, omnipotence, and
perfect love are manifested in the care he shows for his creatures, a
care exhibited in his plan to bring his beloved children to a good
end.

Traditionally, two elements of this notion of providence have

been highlighted. First, God is in control of his universe. Everything
that occurs is either intended or at least permitted by him; nothing
takes place that is beyond his power to prevent. Hence, the world,
with all its inhabitants and all their doings, is radically dependent
upon God. Second, God knows all there is to know about his
universe. Nothing in the past is noetically lost to him; nothing in the
future is beyond his cognitive reach.
 Theories of providence arise because the two elements just noted
appear to lead to problems, especially problems concerning the
freedom of God’s creatures. Take rst the notion of God’s
sovereignty. If we are genuinely free, how can God be genuinely in
control? Doesn’t our freedom set limits to his sovereignty? Isn’t the
world dependent on us, not just on God, for its features and its
history? Mustn’t God’s plan be constantly revised in response to our
actions?

The questions that freedom poses for divine control are thus
varied and serious. No less serious are those that freedom raises for
God’s knowledge. If we’re free, don’t we have to set de nite limits
to God’s knowledge, especially to his foreknowledge? How could he
know what a free being is going to do before that free being decides
to do it? If God already knows, infallibly and unchangeably, what
I’m going to do long before I do it—even long before I’m born—how
could I possibly do otherwise? Mustn’t we, then, concede that a God
who leaves some of his creatures free surrenders not only the
control, but also the knowledge that the traditional picture of
providence a ords him?

Christian advocates of providence have not been blind to such

questions, and through the years have o ered a number of di erent
responses. Though categorizing responses is always a dangerous
(and potentially quite misleading) business, I don’t think it would be
overly simplistic to see the attempts to deal with these questions as
falling into three broad categories.
 First, some Christians have suggested that the problems arise
from our assuming a misguided picture of freedom—that which is
often called libertarianism. Abandon this picture—acknowledge that
it is both philosophically bankrupt and theologically untenable—
and our problems dissolve; human freedom, properly understood, is
fully compatible with God’s complete control and universal
foreknowledge.2 Though a number of names have been ascribed to
this school of thought (including the Augustinian view, the Banezian,
the Calvinist, and others), I will (with some trepidation) follow the
tradition among many Catholic thinkers in referring to this as the
Thomist solution.

Second, other Christians, insisting that the libertarian picture of

freedom ought not be surrendered, have claimed that our notion of
providence needs to be modi ed in order to meet the objections. If
we acknowledge that God’s control over his world is more limited
than past enthusiasts have a rmed, and if we recognize that perfect
knowledge does not entail foreknowledge, we see that true-blue
human freedom is not endangered. In recent years, the most ardent
advocates of this approach have called themselves Open Theists, the
name I also will use in this chapter (along with the shorter
Openists).3

Finally, a third group of Christian theorists holds that neither the

strong, traditional picture of providence nor the libertarian notion of
freedom is negotiable; each is manifestly supported by reason
and/or by revelation. But, these Christians say, the appearance of
con ict between providence and freedom is merely an appearance.
Once we see that God’s sovereign control would operate through his
creatures’ free decisions, and (more speci cally) through his
knowledge of what those decisions would be, we see that none of
the radical moves advocated by the two camps described above
needs to be made. This attempt to reconcile freedom and providence
is usually called Molinism.4

In this chapter, I will attempt to spell out more clearly the

Thomist, the Openist, and (especially) the Molinist approaches to
divine providence, and to indicate as objectively as I can the
strengths and weaknesses of these three positions.5 Before doing so,
however, let me say just a bit more about both the traditional
notion of divine providence and the libertarian picture of freedom.

TRADITIONAL PROVIDENCE AND LIBERTARIAN FREEDOM

The heart of the traditional Christian notion of providence is nicely

stated at the start of ch. 5 of the Westminster Confession:

God, the great Creator of all things, doth uphold, direct, dispose,
and govern all creatures, actions, and things, from the greatest even
to the least, by his most wise and holy providence, according to his
infallible foreknowledge, and the free and immutable counsel of his
own will, to the praise of the glory of his wisdom, power, justice,
goodness, and mercy.
According to this tradition, God is not merely the one who brought
all things into being. All aspects of his creation depend upon his
sustaining presence, and all (great and small) are subject to his
control. As the last line of the passage makes clear, God’s plan for
the world is not merely all-encompassing, but fully in accord with
the divine perfections of justice, goodness, and mercy. And that
plan, the quotation clearly implies, is not one that God makes up as
he goes along. Rather, he foreknows all that will occur, and thus
(one would assume) is neither surprised by what happens nor forced
to alter his intentions as events move in directions unanticipated by
him.

This picture of providence has been dominant throughout the

history of Christianity, though (as we shall see) it has been called
into question in recent years by the rise of Open Theism. On
re ection, it’s not di cult to see why such a picture (which has its
roots both in biblical and in classical Greek sources) should seem at
least prima facie attractive to thoughtful Christians. After all, the
heart of Western monotheism is a conviction that God is the all-
knowing, all-powerful, all-good creator of all that is. How, one
might wonder, could such a deity who chooses to create not know
precisely what he’s doing and what he could have done? How could
he have created a world that wasn’t crafted to mirror his in nite
power, wisdom, and love? Mustn’t we assume that, in Cardinal
Newman’s memorable phrase, ‘He knows what He is about’,6 and
can’t we rest con dent both in his overall plan and in our place
therein?

As we have seen, part of this picture is that God governs all in

accord with ‘the free and immutable counsel of his will’. This, of
course, assumes that God has free will. And this assumption
naturally leads those of a philosophical bent to re ect upon just
what freedom might amount to, in God’s case or in ours. As noted
above, many (though hardly all) Christians throughout history have
insisted that the notion of freedom now commonly referred to as
libertarianism is the one that is most consonant with a Christian
outlook. Let’s now consider, just a bit more carefully, the
characteristic features of libertarian freedom.

Though a complete and precise characterization of libertarianism

is notoriously di cult to provide, the basic idea is that external
determination of a person’s action (especially causal determination
by some factor not subject to the person’s causal control) is
incompatible with that action’s being free. Libertarians insist that
some of our actions are free; hence, they deny that those actions are
determined, mediately or immediately, by events not under our
causal control. And what goes for us goes for God as well; his
freedom too must be understood in this libertarian way.

The central idea here seems to be that my actions (or at least my

free actions) are the ones that I initiate and control. And equally
central, at least for most libertarians, is the connection between
freedom so understood and moral responsibility. Actions caused by
something or someone external to me simply couldn’t be actions for
which I am responsible—for which I could properly be praised or
blamed. If I program an automaton to kill an innocent person, I, not
the automaton, am guilty of murder—assuming, of course, that I’m
not an automaton myself!

Again, what we’ve o ered here is a fairly coarse-grained

depiction of the libertarian picture. Libertarians disagree about
many things: the precise nature and degree of causal relations
present within a free act; the precise connection between the beliefs
and desires we have and the actions we perform; the frequency (or
infrequency) of free acts; whether some genuine actions are unfree;
and so on. But enough, I hope, has been said for readers to grasp the
type of position on freedom that libertarians champion.

I hope as well that enough has been said to see why there at
least appears to be a tension between providence (understood in the
strong traditional manner delineated above) and freedom
(interpreted in a libertarian manner). On the one hand, providence
postulates complete divine foreknowledge of and control over all
that occurs, including human actions. Libertarianism, on the other
hand, insists that external determination of an action is
incompatible with that action’s being free. To endorse providence,
then, it seems we need to deny that there are any free actions, at
least as understood by libertarians; to endorse libertarian freedom, it
seems we must deny (or at least limit) God’s providence. Given this
apparent tension, it is not surprising that Christians have attempted
to rectify matters by surrendering (or at least signi cantly
modifying) one or the other of the two positions that seem to lead to
our quandary. Let us now turn to the two directions in which these
attempts might be taken.

ABANDONING LIBERTARIANISM THE THOMIST ALTERNATIVE

As with the other pictures of providence we will examine, speaking

of the Thomist alternative is a bit misleading, since it implies a
degree of uniformity that simply does not exist. Still, what I am
calling the Thomist position, though a big tent covering a wide
variety of positions, is still a tent, one that shelters those who
respond to our quandary by saying that it is the libertarian account
of freedom, or at least the standard version of that account,7 that is
causing our problems. Liberate ourselves from libertarianism, they
say, and the problem of freedom and providence dissolves.

Libertarianism causes problems, the Thomists say, because it

con icts with one of the most central and indispensable principles
of traditional Christian belief: that God is genuinely in control of
everything that goes on in his universe. Libertarianism suggests that
we free creatures and God are competitors in a zero-sum game of
actualizing the world.8 God performs certain free actions, and
there’s nothing we can do about them. We perform other free
actions, and there’s nothing he can do about them (short, of course,
of simply putting us out of the free-action-performing business
entirely). Increase divine free activity and the sphere for free human
activity shrinks; magnify human freedom, and divine sovereignty
and control wanes. No true Christian, the Thomist says, can
comfortably accept so radical a diminishment of the creator. Hugh
McCann, one of the foremost contemporary advocates of the
Thomist view, nicely summarizes and expands upon this point:

to the extent God does not exert active control over my decisions,
whether through other events or direct involvement, He does not
control them at all. He can therefore achieve His ends only by
reacting to what I do, and to that extent His plans are subordinated
to mine. In addition to weakening His sovereignty, this situation
also threatens God’s omniscience. It suggests He can know how I
will act in the circumstances in which I am placed only by observing
my actions. As creator, He is in the dark. He can know what the
possibilities are, but if my freedom makes for more than one, then
even His knowledge of the world He is creating appears to depend
on my action—an unsatisfactory situation to say the least. These
problems can be avoided if God is able to exercise creative control
in my actual choice.9

For the Thomist, then, God’s sovereignty requires that God have
control over our actions greater than libertarians have typically
allowed. Some Thomists suggest that God’s arranging of natural
causes that determine our actions is fully in accord with their
freedom. But this, to my understanding, has never been more than a
minority view among Thomists. Libertarians are right, they say, in
resisting the common contemporary compatibilist view of freedom
—the view that actions that are functions of the laws of nature and
prior states of the natural world can still be free provided that the
crucial determining events are of the right sort (such as the agent’s
own, fully embraced beliefs and desires) and bring about the action
in the normal way.10 What libertarians have failed fully to
appreciate, though, is that God is not just another natural cause.
God’s relation to his universe is utterly unique, and his determining
of his creatures’ actions no more robs them of their freedom than
does (say) Euripides’ authorial determination of Medea’s actions
mean that she lacks freedom within the world of the play. As
Aquinas famously put it:

it does not of necessity belong to liberty that what is free should be

the rst cause of itself, as neither for one thing to be cause of
another need it be the rst cause. God, therefore, is the rst cause,
Who moves causes both natural and voluntary. And just as by
moving natural causes He does not prevent their actions from being
natural, so by moving voluntary causes He does not deprive their
actions of being voluntary; but rather is He the cause of this very
thing in them, for He operates in each thing according to its own
nature.11

Many Thomists feel comfortable (as Aquinas seems to here)

speaking of God as the source of the very being both of ourselves
and of our actions, and hence as the one whose will supernaturally
causes the free actions we perform. Others feel that such language is
at best misleading; God determines our decisions, they believe,
without there being any causal connection between his act of will
and our decisions. But whatever the precise explanation embraced
here (and I don’t mean to imply that the choice of explanation is
insigni cant), Thomists concur in maintaining that the type of
‘absolute metaphysical freedom’ endorsed by most libertarians, a
view that insists that I, not God, am the ‘ nal ontological arbiter’ of
my actions, cannot be sustained by the circumspect Christian.12

Two other elements of the Thomist position, both at least

suggested by the McCann quotation above, are worthy of mention.
First, Thomists insist that a God who truly enacts a plan for his
world does not need to look at that world to see how things are
going in it. If the creator is truly in charge, then the fact that
something (call it X) is incorporated into his plan just entails that X
occurs, and occurs when and as God wills. (And, of course, for the
typical Thomist, no event is not part of his overall plan.) Once God
determines his own will—once he decides upon a providential plan
for creation—he has no need to observe the world to see what
occurs. To the extent that libertarians imply that God’s knowledge is
dependent upon such observation, their picture of freedom
diminishes the nature and the quality of that knowledge.

The second element is closely related to the rst. In making his

providential plans for the world, God’s will determines how it is that
his free creatures will in fact use their freedom. But there’s no
reason to think that God’s knowledge of creaturely freedom is
exhausted by his recognition of how creatures will actually perform.
He knows not only what they will do, but also what they could have
done. And most Thomists will take this a step further. God knows
not only what they could have done, but also what they would have
done had he placed them in other situations. Take a certain free
creature named Framboise who visits an ice cream shop tomorrow,
and assume that she’ll freely order a raspberry sundae. She does so,
says the Thomist, because of God’s will that she order the sundae.
Consequent to forming his plan, God knows, of course, that
Framboise will order the raspberry sundae. But he also knows that
she could have ordered a strawberry sundae instead. And perhaps he
also knows that, if the ice cream shop hadn’t o ered raspberry
sundaes, Framboise would freely have ordered a hot fudge sundae
instead. He might well know such a counterfactual of creaturely
freedom, but if he does know it, he knows it only because its truth
was also determined by his all-sovereign will; i.e. part of that all-
encompassing act of will was that Framboise would have freely
ordered the hot fudge sundae were raspberry unavailable. Again,
none of this knowledge depends upon his observing what goes on in
the world.

Critics of the Thomist view raise both metaphysical and moral

objections to its picture of God and his relation to creation. The
picture of freedom that it o ers, they suggest, is hopelessly awed.
As noted above, Thomists o er a variety of explanations of the
relation of God’s will to our actions. Some say that the former
initiates a causal chain that ends with the latter; others suggest that
a unique type of supernatural causal connection immediately relates
the two; still others maintain that our very actions are the content of
God’s act of will, and thus not strictly speaking causally connected
to it. Most libertarians respond by saying that the di erences here,
though of arcane metaphysical interest, are uniformly unhelpful in
safeguarding human freedom. Whichever of the various Thomist
schemes we adopt, doesn’t it remain the case that, with regard to
human actions, it’s God and God alone who’s really calling the shots
here, not us? Even if there’s no external event that’s causing me to
act as I do, isn’t the existence of an external agent whose will fully
and inevitably determines my action su cient to rob that event of
freedom?

The moral objections have seemed, to most critics of the Thomist

position, even more severe than the metaphysical ones. For what the
Thomist position implies is that there were no non-logical
limitations upon God’s creative activity. Assuming (as Thomists do)
that God is necessarily sovereign, it follows that any world that is so
much as possible is a world God could have created, since all
contingent events are determined by his will.13 But if this is so, how
can we possibly account for the presence, the amount, and the
horrendous nature of the evil in our world? If God is as unfettered
as the Thom sts say, if he could have created any possible world
whatsoever, then it seems he had available to him worlds much
better than the one in which we live. Adam freely ate the apple, and
the result was unlimited su ering for himself and for his
descendants. But why did he make that free choice? According to
the Thomist, the ultimate explanation lies in God’s will. But then,
why did God will as he did? Why didn’t he will that Adam resist
temptation, and maintain the sinless status of the Garden with
which he was entrusted? Don’t the Thomists, in seeking to magnify
the power and sovereign control of the deity, make him alone
ultimately responsible for the evil that his creatures do, thereby
belittling his moral grandeur? Indeed, by suggesting (as most
Thomists, being traditional Christians, do) that God would punish
for eternity those whom he has determined will never repent for the
sins he has also determined that they commit, haven’t Thomists
depicted a God whose actions display not the loving-kindness of the
God of the Bible, but rather (as Molina puts it) ‘cruelty and
wickedness’?14

Thomists are not without responses here. Perplexity with regard

to the motivations behind God’s permission of evil or uncertainty
concerning the correct analysis of freedom, some suggest, shouldn’t
lead us to soften our conviction that God is truly in charge of all
that occurs.15 As the so-called skeptical theists of today remind us
(see ch. 17 on ‘Skeptical Theism’ in this volume), we should hardly
expect to understand why God acts as he does; hence, our inability
to answer the questions posed above gives us no reason to think
there aren’t su cient responses that simply lie beyond our ability to
grasp. Others have suggested that perhaps we can understand why
God wills evil: to create a world in which both his mercy (in saving
sinners) and his retributive justice (in punishing unrepentant
evildoers) are graphically apparent.16 And others have reminded us
that the e cacy of God’s will is perfectly consistent with the
existence of genuine created causes. Suppose that Pamplemousse, a
young French statesman eager to follow in the ways of his mentors,
cheats on his wife. God’s willing that Pamplemousse be unfaithful
determines the cheating; still, it’s Pamplemousse, not God, who’s
doing the cheating, and it’s Pamplemousse, not God, who’s morally
accountable for that sinful act.17

There are, then, many directions in which the Thomist might go,
and the arguments quickly and predictably become quite complex.18
Still, many (probably most) Christians nd this approach
unsatisfactory. For them, surrendering a libertarian picture of
freedom is too high a price to pay to resolve the tension between
freedom and providence. If one still feels that the apparent
incompatibility between the two is genuine, the only remaining
route of escape is to amend the strong, traditional notion of
providence that we earlier described. And this route takes us to
Open Theism.

AMENDING PROVIDENCE: THE OPEN THEIST ALTERNATIVE

Open Theism, as an explicit alternative to the Thomist and Molinist

accounts of providence, is a fairly new position. Because the strong,
traditional notion of providence seemed so central a part of
orthodox Christian belief, defending Christianity by altering our
picture of providence struck many as a pointless endeavor, or at
least would so have struck them had they even considered it, which
few did. But within the last few decades, many philosophers and
theologians have argued that it is only by revising our idea of how
God interacts with his creatures that we can o er a coherent
account of providence while avoiding the metaphysical and moral
perplexities that both Thomists and Molinists face.

As Openists see it, a rming divine sovereignty in so strong a

fashion as Thomists (or, as we’ll see, Molinists) do is simply
incompatible with our freedom, which can plausibly be understood
only along libertarian lines. Divine foreknowledge of and control
over our free actions both need to be understood in a modi ed way
if our status as free agents is to be safeguarded. But these
modi cations, they insist, should not be seen as denying God’s
status as the all-knowing, provident lord of the universe. On the
contrary, they allow us to fashion a picture of a perfect creator that
is both more solidly based in Scripture and more resilient in the face
of philosophical objections.

Let’s rst consider the limitations on God’s knowledge that Open

Theists embrace. Unlike most Thomists and many Molinists,
Openists typically see God as a temporal being; the idea of an
eternal God who has no relation to the temporal realm is, they
argue, unbiblical or philosophically untenable or both. But if God is
in time, can he foreknow what we will freely do? Openists argue
that he cannot. Though developed arguments for this conclusion
generally get rather complex,19 the basic idea is easy enough to
understand. Open Theists agree with the traditional ascription of
perfection to God, including doxastic perfection: there’s no way God
can believe a proposition which isn’t true. So suppose God had
always known, and hence always believed, that Framboise would
order that raspberry sundae tomorrow. Since his belief is infallible,
there’s no way he could have this belief and Framboise not order the
sundae. So, when Framboise enters the ice cream shop, there’s a fact
about the past (regarding God’s prior belief) that entails her action.
How, then, could she be free? If she could change what it was that
God believed in the past, perhaps we could make sense of her
having real alternatives. But how could she do something tomorrow
to change what it was God believed thousands of years ago? What
God believed back then is a settled fact about the past; hence, it’s
incoherent to think that there’s something she can do about it
tomorrow. So God’s believing in advance that a free action will be
performed entails that the action in question isn’t free after all. And
from this it follows that such belief isn’t possible at all.20

So divine foreknowledge has to be denied to make room for

human freedom. But it’s not only knowledge of what will happen
that needs to go. Remember those conditional propositions that, as
we saw, Thomists typically see God as determining—propositions
(counterfactuals of creaturely freedom, as we called them) not about
what will happen, but about what would happen under various
hypotheses. Such conditionals can be seen as answers to questions
about potential creaturely actions. If Framboise were to go to the ice
cream shop, what would she freely order? If Framboise were to nd
the ice cream shop had no raspberry sundaes, what would she freely
order instead? Thomists generally want to insist that there are
de nite answers to such questions, answers dependent upon God’s
will. But Open Theists reject the claim that there are true
counterfactuals of this type. For suppose there were such truths, and
suppose that God, being omniscient, knew them—all of them. This
means that God would know what any free being would do in any
situation in which it might be placed. But if God knew a
counterfactual such as If Adam were to be placed in the Garden, he
would eat the apple, and then were to decide to place Adam in the
Garden, he would know that Adam will eat the apple. In other
words, divine foreknowledge seems undeniable once we grant there
are true counterfactuals of creaturely freedom known by God. And
since we’ve already seen (according to Open Theists) that
foreknowledge is incompatible with our freedom, we need to reject
counterfactuals of creaturely freedom as well.21

Now, it might seem that the Openists’ God, knowing neither

what will happen nor what would happen (under various
hypotheses), would have scant providential control over his world.
But Open Theists suggest that God, though not the all-controlling
deity of the Thomists, would still have enough to go on. For even if
God doesn’t know what Adam would do if placed in the Garden, he
might well know what Adam would probably do. And knowledge of
such ‘would-probably’ conditionals, say Openists, gives God enough
information to operate providentially. Take Framboise once again.
Perhaps God believes that Framboise (and others) would bene t
greatly were she to meet a marriageable young gent. And maybe he
sees that there’s a situation he can put her in such that, if in that
situation, she’d probably freely decide to go to the ice cream shop.
Once there, she’d probably freely decide to order a raspberry sundae.
And, upon ordering it, she’d probably strike up a conversation with
that nice young sundae-maker (and raspberry a cionado)
Maraschino, who’d probably ask her out on a date, which (since God
sees that the two are in fact a perfect match) she’d probably accept,
and so on. The point is that knowledge of probabilities is enough for
God to act knowingly to achieve his good ends.

Note, though, that there are no guarantees in this picture of

providence. God can’t (as the Thomists’ God can) simply decide how
Framboise will freely react in these various situations. And, since
improbable things do happen, there’s a genuine possibility that
Framboise will act so as to frustrate God’s good plan for her. So the
God of Open Theism is a genuine risk-taker. He aims for the best,
but with the full knowledge that his free creatures may disappoint
him, and block his endeavors to lead us toward what is best for
ourselves and for his world. The moral evil that occurs in the world
is to be explained not via the contortions needed to justify the all-
determining God of the Thomists, but as the all-but-inevitable result
of God’s gracious creation of beings whose free actions he can
neither foresee nor control.

Openists can develop this picture of a risk-taking God in a

number of di erent directions. Some embrace it rather reluctantly
and try to minimize the degree of risk God faces, by curtailing the
quantity of signi cant free actions or by maximizing the
probabilities God knows. (If God knows that there’s a .999999998
probability of Framboise’s ordering a raspberry sundae, he knows
there’s not much of a chance that she’ll escape Maraschino’s
attention.) Others openly prefer the picture of a God who values our
freedom so much that he’s willing to share with his free creatures a
very large degree of the responsibility for fashioning our world.
Openists also di er on how exactly God would utilize his knowledge
of would-probably conditionals. Some think that God would from
the start develop an unlimited number of contingency plans ready to
be enacted should things take an unexpected turn. (‘OK, I’m going to
get Framboise into that ice cream shop tomorrow. She’ll probably
order a raspberry sundae, getting Maraschino’s attention. But if she
doesn’t, I’ll arrange for her to spill her water glass. That will
probably get Maraschino to come over and try to make a joke of it,
thereby getting them to talk. But if he doesn’t come over, then
I’ll…’)22 Many Open Theists, though, seem to feel that such
extensive preplanning would be neither necessary nor tting for
God. On their view, God has ample resources, given his knowledge
of would-probably conditionals, to respond to any surprising
development if and when it occurs. Rather than meticulously
preprogramming his reactions, he responds to his creatures in real
time, using all the knowledge he has to nudge humankind toward
the good no matter how they might behave.23

So Open Theism can be developed in a number of di erent

directions; as with Thomism and (as we’ll see) Molinism, a
multitude of distinct positions are available. Still, the heart of the
Openist view is clear enough: with respect to free creaturely actions,
God has in advance only would-probably knowledge, and thus must
take risks in his providential interactions with the free beings he
creates.

The development of Open Theism over the past two or three

decades has led to stormy debates, especially within certain
evangelical organizations and colleges. The debates have often
enough turned to practical as well as theoretical issues: membership
for Open Theists within Christian academic societies has been called
into question, and teaching jobs have been lost. Though unseemly,
these con icts are in a sense unsurprising. For Open Theism
constitutes a radical break with the picture of providence that has
been dominant within Christianity throughout its history. And
defenders of that dominant picture have not been reticent in
pointing to what they see as the religious and philosophical
de ciencies of the Openist alternative.
 Many critics of Open Theism charge that the view, though
allegedly developed in part to o er a more biblically accurate
picture of God’s relationship to his people, in fact does the opposite.
The strong, traditional picture of providence, these critics insist, is
actually demanded by the Bible; the hesitant, odds-playing God of
the Openists, they say, can hardly be reconciled with the meticulous
providence so evident in Scripture, and so repeatedly endorsed by
councils and catechisms throughout the history of the church. Part,
though not all, of this criticism centers on the question of prophecy.
If God doesn’t know what someone will freely do, how can he reveal
what he will do? If no one can know how Peter will freely react to
Christ’s arrest, how can Jesus so con dently predict Peter’s denials?
Open Theists have o ered several responses to this criticism. Some
predictions may be of actions that are not in fact free. Some may be
implicitly conditional; e.g. what Jesus was really telling Peter was
that if the apostle acted in accord with the character he’d formed up
to that point, he d deny Jesus. And some prophecies may be
revealing what God intends to do regardless of how his creatures
act. Not surprisingly, critics of Open Theism have not been
persuaded by such responses.24

Equally questionable, from the traditionalists’ perspective, is the

role God’s knowledge of probabilities plays in the Openists’ account
of providence. Some critics have called into question whether an
Openist can consistently a rm that God knows even would-
probably conditionals.25 Even if such objections can be parried,
though, serious questions arise when we consider the level of the
probabilities of which God is supposedly aware. For God to be a
genuine risk-taker of the sort so many Openists celebrate, the
probabilities God knows cannot uniformly be extremely high. Again,
if the probability of Framboise’s ordering a raspberry sundae if she
goes to the ice cream shop is .999999998, God (assuming he wants
her to order the sundae) isn’t taking much of a risk if he gets her to
go. If all the probabilities God knew were of this type, he would be
nearly as manipulative and controlling as is (according to Openists)
the God of the Thomists—and evil would remain nearly as big a
problem. On the other hand, if the probabilities God knows are
more varied, it’s hard to see how God could have anything more
than a very short-range plan for his world, and thus anything more
than an extremely emaciated type of providential control.26

As one might expect, the Open Theists’ embrace of a risk-taking

deity has also proved controversial. While we may admire humans
who are willing to take risks to bene t others, few of us (the critics
say) see risk-taking as inherently valuable, especially with respect to
matters of life and death. The heart surgeon who employs a risky
new procedure when a (comparatively) risk-free one is available
would be viewed not as courageous or daring, but as foolish and
perhaps even malevolent. A God who (pace Einstein) does play dice
with the universe, at least with free macro-objects if not with
undetermined micro-ones, is more to be pitied than acclaimed.
Indeed, far from lessening the problem of evil, the Openist stance
actually intensi es it. For God could surely make a wonderful world
without including free creatures in it. If the cost of creaturely
freedom is the genuine possibility that some, or most, or all of his
creatures will use their freedom to make things worse and worse for
themselves and for their fellows, a truly loving God would deem
such a cost too high.27

Open Theists have, of course, responded to these and other

criticisms of their position.28 While their spirited defense of this
view has led to many converts, many (probably most) Christians
who have considered the issues retain grave doubts about the
adequacy of the Openist alternative. For those who harbor such
doubts but who also cannot bring themselves to accept the Thomist
position, the only remaining live candidate for their consideration is
Molinism.

LIBERTARIANISM AND TRADITIQNAL PROVIDENCE: THE MOLINIST

ALTERNATIVE

As our discussion has already indicated, Molinism is the position of

those who decline to surrender either the libertarian notion of
freedom or the traditional picture of providence—and who insist
that the two can be combined without contradiction. Let me brie y
describe the Molinist view and then indicate how, according to
Molinists, it dissolves the alleged tension between freedom and
divine sovereignty.29
 Like their Thomist brethren, Molinists believe that the alterations
to the notion of providence advocated by Open Theists are
unacceptable. Divine control over all that occurs, along with both
foreknowledge and knowledge of counterfactuals of creaturely
freedom, are non-negotiable elements of a sound doctrine of
providence. But, as libertarians, Molinists side with their Openist
brethren in condemning the Thomist embrace of an all-determining
divine will. What we need to see, say Molinists, is that God’s
knowledge and control operate through creatures who enjoy full
libertarian freedom.

Consider, Molinists say, God’ s knowledge of his world. Some of

the truths God knows—e.g. mathematical truths such as two plus
three equals ve—are necessary truths, ones that could not have been
other than they are and that are in no sense the result of any free
decision on God’s part. Knowing such truths can be seen as part of
God’s very nature; hence, Molina labeled this God’s natural
knowledge. On the other hand, many truths of which God is aware—
e.g. Framboise will freely order a raspberry sundae tomorrow—are
neither necessary nor beyond God’s power to control. There are
many possible worlds in which Framboise never has a sundae
tomorrow, and there are many ways in which God could have
prevented her having one (most radically, by deciding never to
create Framboise at all). Contingent truths of this sort, then, are true
only because God freely allowed them to be true. Such propositions,
said Molina, are elements of God’s free knowledge. And this,
obviously, is where foreknowledge of what will occur in the world
belongs.

But where do the counterfactuals of creaturely freedom t in?

Recall that, on the Thomist picture, God knows what any free
creature would do in any situation in which it might be placed. If
Framboise were to go to the ice cream shop tomorrow and nd that
raspberry sundaes were unavailable, there’s a fact of the matter as
to what she would freely do (e.g. order a hot fudge sundae), and
God, being omniscient, knows that fact. Molinists agree, but (unlike
Thomists) believe that the existence of such truths force us to posit a
third category of divine knowledge. For these conditionals, like
elements of free knowledge, are contingent truths; but, like elements
of natural knowledge, they’re not in any way under God’s control,
given the fact that the creatures they are about have libertarian
freedom. (As Molinists are wont to say, such truths are prevolitional,
meaning simply that they re true independent of any exercise of
God’s free will.) So the contingent but prevolitional truths really
belong in a middle category between the necessary, prevolitional
truths that make up natural knowledge and the contingent,
postvolitional (i.e. those that are subject to God’s will) truths that
constitute free knowledge. Not surprisingly, Molina gave the name
middle knowledge to this third category.

The notion of middle knowledge is the cornerstone of the

Molinist position. For it allows us to see how God can both
foreknow and exercise control over all that occurs. Since both
natural knowledge and free knowledge are prevolitional, they can
be thought of as present to God when he’s deciding what creative
act to perform.30 Given his middle knowledge, God knows exactly
how his free creatures would react in any situation in which he
might place them. He knows, for example, that if he were to create
Framboise and she were to visit the ice cream shop, she’d freely
order a raspberry sundae; he knows that if raspberry were
unavailable, she’ d opt for hot fudge instead; and so on. God knows
such truths about every possible creature and every situation in
which she might be placed.31 So, once he decides which creatures to
create and which situations to place them in, he foreknows exactly
what that creature will do. For example, by middle knowledge, he
knows what Framboise would do if she were in the ice cream shop;
once God decides to put her in the ice cream shop, it follows that he
knows what she will do. So middle knowledge, when combined with
God’s full creative decision (concerning which beings to create in
which circumstances), simply entails foreknowledge. Given middle
knowledge, then, foreknowledge no longer appears to be a mystery.

Nor does it seem to be a threat to our freedom. For

foreknowledge is not (as with the Thomists) simply a function of
God’s free choice; rather, it’s a function of his choice concerning
what to create (over which he does have control) and truths about
what we would freely do (truths over which he has no control
whatsoever). If one thought of the counterfactuals about Framboise
as utterly beyond her control, one might have doubts about whether
or not middle knowledge safeguards her freedom. But no Molinist
worth his or her salt thinks of the counterfactuals that refer to a
creature as manacles she is powerless to escape. For such
conditionals say only what the creature would in fact freely do, not
what she would have to freely do. It may be true that Framboise
would freely order a raspberry sundae, but it’s also true that, when
in the ice cream shop, she has full libertarian freedom to (say) order
a vanilla milkshake instead, and were she to do so, the
counterfactual that is in fact true about her would never have been
true. In other words, God is powerless regarding which
counterfactuals of creaturely freedom are true, but we aren’t.32
Hence, middle knowledge is fully consistent with our freedom.

Furthermore, say Molinists, it’s also consistent with a strong

notion of divine providential control. True, given the reality of
libertarian freedom, God can’t exercise the type of unilateral control
promoted by the Thomists. Even so, say Molinists, middle
knowledge a ords God a much stronger type of control than the
emaciated Openist version. For the Molinist God, thanks to middle
knowledge, isn’t a risk-taker. He never has to deal with acting in
situations where he knows only what probably will result from his
actions. Rather, he knows precisely how his free creatures would
react to any action on his part. If he were to put Framboise in
situation A, she’d freely do X; if he were to put her in situation B,
she’d do Y; and so on. God can then decide which situation to put
her in (say, A or B) depending upon which result (X or Y) can more
readily be woven into a world that satis es his creative intentions.
God needn’t worry about things heading o in an unintended
direction—about his being surprised by unexpected actions on the
part of his creatures. With middle knowledge, he truly does know
what he is about; by exercising his providential control through the
free actions of his creatures, he (and we) are assured that the world
that results will fully manifest his wisdom and love.

The advantages of the Molinist view (in saving both the

libertarian notion of freedom and the classical picture of
providence) are evident, and attempts to apply it to a number of
speci c Christian beliefs—the problem of evil, prophecy, petitionary
prayer, biblical inspiration, original sin, and the salvation of the
unevangelized, among others—have been o ered by a number of
Molinists.33 Potentially the most fruitful such application may prove
to be a Molinist Christology, which alone seems to o er the prospect
of providing us a coherent picture of the incarnate Son that ensures
both his genuine libertarian freedom and his true impeccability.34
Further re ection on the implications of this Christology (which is
controversial even among Molinists) can be expected in the coming
years. Much light might be thrown on the ultimate destiny of those
who make up the Body of Christ.

Molinism has always been an extremely controversial position,

and disagreements concerning the coherence and the plausibility of
the view continue. Before turning to the major objection that has
been raised, let me rst note two lessdiscussed but signi cant
questions concerning this theory of providence.

First, what exactly is to be included in the antecedent of a

counterfactual of creaturely freedom? Molinists have typically
argued that the basic elements of middle knowledge should be
conditionals with complete antecedents. To see what they mean by
this, note that most everyday counterfactuals are radically
incomplete with respect to their antecedents, though the context
usually makes the direction in which completion would be made
relatively easy to see, even if impossible for us fully to state. For
example, with conditionals such as

If I’d asked you at noon yesterday to join me for lunch, you’d

have agreed

we naturally assume that the idea is that, had the world gone as it
in fact did up until noon yesterday, and had I then asked you to join
me, you’d have agreed. We’re not, for example, asking how you
would have reacted had I asked you to lunch at noon after
performing some action (say, insulting you vilely at 11:45) that, as a
matter of fact, I didn’t perform. With some counterfactuals, of
course, the consequent is stated in a manner no more precise than
the antecedent. Take, for example, the memorable words sung by
Tevye in Fiddler on the Roof:

If I were a rich man,

Ya ha deedle deedle, bubba bubba deedle deedle
 dum.
All day long I’d biddy biddy bum.
If I were a wealthy man.35

The general thrust of Tevye’s counterfactual may be clear enough to

us, but its precise meaning is tough to discern.

None of these interpretative problems, it seems clear, could

plague the counterfactuals God knows via his middle knowledge.
The counterfactuals would need to have complete antecedents, in
part because, at the logical moment when such conditionals are of
use to God, there is no fact of the matter as to how the world has
gone up to the relevant point in time, and thus no way for the
context to allow completion of the antecedent. Besides, one would
think that God’s creative decisions would best be guided by (and
could adequately be guided only by) conditionals with the widest
antecedents and the most exact consequents—‘If I were to create a
world in which things had gone precisely this way up to time t, and
Tevye were free at t, he’d freely do exactly X.’

Nevertheless, determining what that ‘precisely’ amounts to—i.e.

specifying exactly what goes into the antecedent C of a
counterfactual of creaturely freedom (which we might represent as
C→ A)—is no mean feat, one to which Molinists have heretofore
perhaps devoted less time than the task deserves, and one on which
there is no reason to think all Molinists will agree. It makes sense to
think that God’s foreknowledge of the relevant action would not be
included in the antecedent of such conditionals. If C includes ‘God
foreknows that A’, then (C → A) will be necessarily true; if it
includes ‘God foreknows that ∼ A’, then it will be necessarily false.
In neither case will it be the type of contingent, prevolitional
conditional that would seem to be of much action-guiding use for
God. Furthermore, for the Molinist, foreknowledge is simply an
e ect of God’s middle knowledge and creative action, not part of the
situation in which the action is performed, and thus seems peculiar
to include in the antecedent. So leaving foreknowledge out of C
makes sense. But what, for example, of cases of prophecy? Here one
could see Molinists torn in both directions. On the one hand, a
prophet’s publicly declaring ‘You will do A’ seems an obvious
actual-world exercise of causal power on his or her part, and thus an
exceedingly odd factor to exclude from C. On the other hand, a
genuinely prophetic utterance, if included in C, would give us an
antecedent that entails the consequent as surely as would including
divine foreknowledge, and thus would render the conditional
necessarily true, and so no part of middle knowledge.

More careful and detailed consideration of this issue by Molinists

seems to be called for.36

Another issue Molinists might well need to face more directly is

a challenge suggesting that they have given insu cient reason to
reject the Thomist view that sees God as the cause of the truth of
contingent counterfactuals of creaturely freedom.37 As noted above,
Molinists typically say that God knows which such counterfactuals
are true, but that their truth in no sense constrains the relevant
agents, since they retain (Molinists say) counterfactual power over
their truth. Even if (C → A) is true, the agent in question, if placed
in C, retains the power to do something (i.e. bring it about that ∼
A) such that the counterfactual would have been false. But then, the
challenge goes, why not just say that God at-out causes certain
counterfactuals to be true, but that his causing in no way constrains
the relevant agents, since they retain counterfactual power over
their truth? For example, God, prior to creation, causes (C → A) to
be true, but the agent in question, if placed in C, retains the power
to do something (bring it about that ∼ A) such that the
counterfactual would have been false. Of course, the counter
factual’s being false would entail that God never caused it to be
true, and so having counterfactual power over its truth would entail
having counterfactual power over God’s prior causal activity. But
many Molinists have been willing to countenance such power in
other instances—take, for example, Plantinga’s celebrated example
of Jones’s having the power to act in such a way that God would
have done something he in fact didn’t do, namely, prevent some
ants from moving into Jones’s yard yesterday.38 Molinists, it would
seem, can parry this Thomist thrust only if they can o er strong
reasons to think that we couldn’t have counterfactual power over
God’s causation of counterfactuals. Though I suspect that such
reasons can be o ered, it is admittedly a task that no Molinist has
yet fully undertaken.
 Rather than focus on such questions, though, opponents of
Molinism have over the last thirty years devoted most of their
attention to the ‘grounding’ problem. Counterfactuals of creaturely
freedom are supposed to be true prior to the actions (or even the
existing) of the agents in question. How, the skeptics ask, could such
conditionals be true at all? What would cause, or bring about, or
ground their truth? God, traditional Molinists have insisted, can’t
make the conditionals that are true to be true; he knows but doesn’t
control their truth. But no one and nothing else exists at the logical
moment (prior to creation) when God knows and uses these truths
in his creative deliberations. But then, how could anyone or
anything make these propositions to be true? And if there is no
ground, no truthmaker, no cause of truth that we can identify, how
can we possibly allow that there are any such truths?

Now, speaking of the grounding objection, as some critics of

Molinism tend to do, is obviously misleading, since such terms as
ground, bring about, cause, make true, and so on are not clearly
synonymous, and since each individually can be interpreted in a
number of di erent ways. Trying to spell out all the available
variations on the basic form of the grounding objection by
considering all such interpretations would in all likelihood be a
tiresome venture—and, I suspect, one that’s not worth the e ort.
After all, hardly anyone claims that his or her grounding argument
constitutes a refutation of middle knowledge, and even the most
ardent proponents of the grounding objection acknowledge that
there are responses available to render Molinism coherent.39 Still,
looking at one recent version of the argument may be helpful
insofar as it o ers a decent idea of the types of move each side
makes in this sometimes tedious debate.

William Hasker has suggested that the key intuition of the

grounding objector might be captured via the following principle:

(GP) Any true contingent proposition is true in virtue of the

existence or non-existence of some concrete state of
a airs.40

Hasker explains that by a concrete state of a airs he means ‘the

exempli cation of an occurrent property by a substance at a time, or
of an occurrent relation by two or more substances at a time’. The
notion of existence in (GP), Hasker says, is to be taken ‘trans-
temporally’: ‘a thing “exists” if it exists now, or has existed, or will
exist’. The ‘in virtue of’ relation is left unexplicated.41

Given (GP), Hasker argues, it follows that there can be no true

counterfactuals of creaturely freedom. For such conditionals (if true
at all) would have to be contingent truths, yet there are no concrete
states of a airs in virtue of which they are true. This is, perhaps,
most glaringly true (he suggests) for counterfactuals with
antecedents that turn out to be false. If, for example, Framboise isn’t
in the situation of being told that raspberry sundaes are unavailable,
then there simply isn’t anything that she (or anyone else) does that
could be seen as grounding any counterfactual of freedom about
what she would do. So all such conditionals, Hasker concludes, are
necessarily false.

In response to any grounding objection, there are two directions

in which a Molinist might be inclined to go. Grounding objections
are akin to shouts, directed at the Molinist, of ‘Your Conditionals
Aren’t Grounded!’ where some speci c grounding principle such as
(GP) is assumed.42 Responses are then apt to be akin to counter-
shouts either of ‘Are So!’ or of ‘So What?’ Sometimes, in fact,
Molinists might be inclined to o er both ripostes, indicating that one
is appropriate if the argument is taken one plausible way, the other
if the argument is understood in a di erent plausible way. Let’s see
which (if either) of these two replies the Molinist might reasonably
use against Hasker.

Could an ‘Are So!’ response be o ered? That is, could a Molinist

plausibly claim that (GP) is true, but that counterfactuals of
creaturely freedom pass the (GP) test for grounding? That depends
in part upon how that unanalyzed relation ‘in virtue of’ is
understood. Alas, constraints of space preclude a full discussion of
this matter. My own guess, though, is that, with regard to this
version of the grounding objection, ‘Are So!’ replies probably would
ring hollow.

Would ‘So What?’ counter-shouts be more viable? That is, is it

plausible to claim that counterfactuals of creaturely freedom are
ruled out by (GP), but that this doesn’t matter, since (GP) is false?
There is much that might be said here about whether, say, ‘would-
probably’ conditionals or laws of nature do in fact (as Hasker
asserts) pass muster under (GP). Should they fail to do so, the
plausibility of (GP) would obviously be reduced. Discussion of these
issues, though, would take us too far a eld. Let us instead focus on a
di erent approach a Molinist might take to call (GP) into question.

Recall that Hasker relies upon a ‘trans-temporal’ reading of the

notion of existence in (GP); the concrete state or event in virtue of
which a proposition is true ‘exists now, or has existed, or will exist’.
For clarity’s sake, let’s rephrase (GP), with the transtemporality of
‘existence’ made explicit, thus:

(GP) Any true contingent proposition is true in virtue of some

concrete state of a airs that does exist, or has existed, or
will exist.

Now, pretend that you (unlike Hasker) are a skeptic concerning non-
conditional future contingent propositions (such as Framboise will
freely order a raspberry sundae tomorrow). You’re perfectly willing to
agree that things were and are a certain way, but not willing
(because of human freedom and/or God’s freedom and/or quantum
indeterminacy and/or whatever) to grant that things will be a certain
way. If so, you’d be apt to think that (GP) is unduly complicated, for
there are no concrete states of a airs that will exist. (There are lots
that might exist, but none that de nitely will.) So you decide to help
Hasker out. Instead of saying that a thing ‘exists’ just in case it
‘exists’ now, or has existed, or will exist, you suggest, we should
really say that something ‘exists if and only if it exists now or has
existed—period. Instead of the painfully swollen (GP), you say, all
we need is a thinner, more modest principle:

(GP–) Any true contingent proposition is true in virtue of some

concrete state of a airs that does exist or has existed.

(GP–), you note, seems to work just as well against the Molinist as
the original (GP) did. So Hasker, you conclude, should look upon
the move to (GP–) as but a friendly amendment strengthening his
argument.

It’s fairly obvious, though, that Hasker would view this alteration
as anything but friendly. Excluding counterfactuals of creaturely
freedom from the realm of the contingently true is not enough to
make a grounding principle plausible for Hasker; it also needs to
include all those contingent propositions that should rank as
grounded. Since (GP–) wouldn’t allow us to say that Framboise will
freely order a raspberry sundae tomorrow is true, and since we need a
grounding principle that does allow such truths, we need the bee er
(GP), not the comparatively emaciated (GP–).

Now, I would have no complaints about Hasker’s responding in

such a manner. But note what it suggests about grounding
principles. Starting with the assumption that contingent truths need
grounding, we further assume that we know at least some of the
more important classes of such truths. We know, for example, that
there are non-conditional contingent truths about the past and about
the present; we know (Hasker assumes) that there are laws of nature
and ‘would-probably’ conditionals; and we know that there are non-
conditional contingent truths about the future. We don’t start with a
grounding principle and use it to decide whether or not the
members of one of these classes of propositions are in fact
contingent truths. Things work the other way around. We start with
the classes of propositions we feel con dent about and fashion a
grounding principle that will (at a minimum) not rule out any of
them. Contingent truths about the future would be ungrounded if
(GP–) were an adequate account of grounding, but so what? Because
it doesn’t let them in, it follows that it’s not adequate.

But, of course, what’s good for the Openist goose is good for the
Molinist gander. Suppose the Molinist assumes that contingent
truths need grounding. He, like Hasker, will assume as well that he
knows some of the major classes of such truths. He knows that there
are non-conditional contingent truths about the past and about the
present; he knows (let’s assume) that there are laws of nature and
‘would-probably’ conditionals; he knows that there are non-
conditional contingent truths about the future; and he knows that
there are true counterfactuals of creaturely freedom. Like Hasker, then,
he won’t start with a grounding principle and use it to decide
whether or not the members of one of these classes of propositions
are in fact contingent truths. Things work the other way around.
He’ll start with the classes of propositions he feels con dent about
and fashion a grounding principle that will (at a minimum) not rule
any of them out. Counterfactuals of creaturely freedom would be
ungrounded if (GP) were an adequate account of grounding, but so
what? Since it doesn’t let them in, it just follows that it s not
adequate. What we need is a bee er principle, one that
acknowledges that grounding states or events include not only ones
that do, did, or will exist, but ones that would exist (under speci ed
conditions). What we need, in other words, is something more on
the order of

(GP+) Any true contingent proposition is true in virtue of some

concrete state of a airs that does exist, or has existed, or
will exist, or would exist (under speci ed conditions).

And with (GP+), of course, we have a principle that the Molinist

can endorse without hesitation.

Not every version of the grounding objection, again, can be met

by the Molinist with the same response. A Molinist needs to look
very carefully at the details of a speci c grounding objection before
determining what response is appropriate. Still, for every version of
the grounding objection of which I am aware, some elaboration of
the ‘Are So!’ or of the ‘So What?’ response can be made by the
Molinist. Hence, though one can safely predict that variations on the
grounding objection will continue to appear, and that a signi cant
number of philosophers will nd one or another of them convincing,
the Molinists’ track record in coming up with responses suggests
that variations on prior responses will likely be available.

CONCLUSION
I have argued that theories of divine providence are of three basic
types. Each such type has its advantages and its disadvantages. Each
has had numerous able and creative defenders. As with most
philosophical disputes, one can hardly expect this debate to come to
an end. I suspect that the eld of battle may shift more clearly in the
coming years to considerations of which view, when applied to
speci c doctrines (such as the Incarnation), o ers us the most
satisfying overall position. Still, it seems quite likely that all three
positions will continue to be defended (and attacked) for the
foreseeable future.

NOTES

I am grateful to Mike Rea for comments on an earlier draft of

this chapter.

1. For the rest of this chapter, I will speak of the Christian notion
of providence. In general (though with some obvious
exceptions), what I say would apply equally well within the
Jewish or Muslim traditions.

2. By foreknowledge I mean knowledge of what lies in our future.

Whether God is best to be thought of as existing in time is a
much disputed question, one that I think can be largely
ignored for our discussion.

3. Process theists, who di er from Open Theists in a number of

respects, would also t into this general category. My
 discussion, though, will ignore process theism, largely because
(as I see it) Open Theism o ers the more promising (and, in
recent years, surely the more discussed) version of this general
approach to providence. It’s also, perhaps, worth noting that a
number of features characteristic of (even if not essential to)
Open Theism—e.g. the general rejection among Open Theists
of such traditional divine attributes as eternity and simplicity
—will be largely ignored in our discussion.

4. The name comes from Luis de Molina, the sixteenth-century

Jesuit theologian who rst explicitly proposed this solution.

5. Readers should be forewarned that perfect objectivity may be

too much to expect from the author of a book called Divine
Providence: The Molinist Account (Ithaca: Cornell University
Press, 1998). Though I am unabashedly an advocate of the
Molinist position, and though I shall devote more space to
discussing that position than to the two alternatives, I shall not
endeavor to underestimate the real challenges that Molinism
continues to face, nor to ignore the genuinely alluring
elements o ered by its rivals.

6. John Henry Newman, Prayers, Verses and Meditations (Ft.

Collins: Ignatius Press, 2002), 339.

7. I will, by and large, henceforth dispense with this quali er, but
the reader should understand it to be tacitly present in most of
the following discussion.
 8. The language here is borrowed from my colleague David
Burrell. See e.g. his Freedom and Creation in Three Traditions
(Notre Dame: University of Notre Dame Press, 1993), 112.

9. Hugh McCann, ‘Divine Sovereignty and the Freedom of the

Will’, Faith and Philosophy 12 (1995), 586.

10. For a clear expression of the Thomist rejection of contemporary

compatibilism, see e.g. Theodore J. Kondoleon, ‘The Free Will
Defense: New and Old’, The Thomist 46 (1983), 19.

11. Thomas Aquinas, Summa Theologica, I q. 83 a. 1 ad 3.

Translation by the Fathers of the English Dominican Province.

12. The words quoted are from McCann, ‘Divine Sovereignty’, 593.

13. I assume, here and throughout this discussion, the standard

contemporary notion of a possible world. For a magisterial
explication of possible worlds, see Alvin Plantinga, The Nature
of Necessity (Oxford: Clarendon, 1974), esp. ch. 4.

14. Luis de Molina, On Divine Foreknowledge: Part IV of the

Concordia, trans. Alfred J. Freddoso (Ithaca: Cornell University
Press, 1988), disp. 50 sect. 14 (p. 139).

15. See e.g. the classic discussion in Reginald Garrigou-Lagrange,

The One God, trans. Dom. Bede Rose (St Louis: B. Herder,
1944), 469–71.

16. For further discussion and references, see my ‘Two Accounts of

Providence’, in Thomas V. Morris (ed.), Divine and Human
 Action: Essays in the Metaphysics of Theism (Ithaca: Cornell
University Press, 1988), 168–70.

17. For an interesting recent exchange on this point, see Hugh

McCann, ‘The Author of Sin? Faith and Philosophy 22 (2005),
144–59, and Katherin Rogers, ‘God is Not the Author of Sin:
An Anselmian Response to McCann’, Faith and Philosophy 24
(2007), 300–10.

18. For a fuller discussion, see my Divine Providence, 84–94.

19. See e.g. William Hasker, ‘Foreknowledge and Necessity’, Faith

and Philosophy 2 (1985), 121–57. For one response, see my ‘In
Defense of Theological Compatibilism’, Faith and Philosophy 8
(1991), 237–43.

20. Though Openists agree that God cannot know (or even believe)
propositions about what free creatures will freely do, they
di er as to whether or not there are any such propositions that
are in fact true. For an interesting discussion of the
alternatives, see Dale Tuggy, ‘Three Roads to Open Theism’,
Faith and Philosophy 24 (2007), 28–51.

21. Note that the picture isn’t of a God who is ignorant of

something that could be known. Openists are united on this
point: there simply isn’t anything that could be known here.
Could an Openist believe that there are true counterfactuals of
creaturely freedom, but that they’re one and all necessarily
 unknowable, even by God? Technically, yes. But this is not a
popular Openist stance.

22. Such an approach is at least suggested in Gregory Boyd, God of

the Possible (Grand Rapids: Baker Books, 2001), 61, 127. See
also Thomas Oden, The Living God: Systematic Theology, Volume
One (San Francisco: Harper, 1987), 306. For a discussion of
the advantages and disadvantages of such an approach, see my
review of William Hasker’s Providence, Evil and the Openness of
God, in Philosophia Christi 8 (2006), 493–6.

23. It’s also worth noting that an Open Theist could view a God
who constantly uses his knowledge of would-probably
conditionals to interfere with human events as a tad too
manipulative. For an interesting presentation of such a view,
see James Rissler, ‘Open Theism: Does God Risk or Hope?’,
Religious Studies 42 (2006), 63–74.

24. For a fuller discussion of this issue, see William Hasker, God,
Time, and Knowledge (Ithaca: Cornell University Press, 1988),
194–6, and my Divine Providence, 100–2.

25. See Michael Robinson, ‘Why Divine Fore knowledge? Religious

Studies 36 (2000), 25175, and Jennifer Jensen, ‘The Grounding
Objection to Molinism’ (Doctoral dissertation, University of
Notre Dame, 2008), ch. 5. Hasker has responded to Robinson
in Providence, Evil and the Openness of God (London: Routledge,
2004), 206–11.
26. For more on this topic, see my Divine Providence, 102–5.

27. For a fuller discussion of these issues, see ibid. 105–7.

28. See especially Hasker, Providence, Evil and the Openness of God,
‘Appendix: Replies to My Critics.

29. For a much more extensive and detailed presentation of the

Molinist position, see my Divine Providence, especially ch. 2.

30. The language here suggests that God is in time—that there’s a

time when he’s deciding what to do, and a later time at which
he does it. Many Molinists, though, think of God as atemporal,
and view the use of temporal metaphors (based, obviously, on
the fact that our decision-making is typically a temporally
extended process) as a useful but potentially misleading way
of describing what is essentially a dependence relation, not
essentially a temporal one. For more on this point, see ibid.
37, 174–6.

31. I’m speaking a bit loosely here in referring to possible creatures.

Molinists are not committed to the odd claim that, in addition
to all the actual creatures in existence, there are also a horde
of merely possible creatures hanging around! For a stricter
statement of the Molinist view, see ibid. 46–7.

32. At least, we aren’t powerless with respect to counterfactuals

with true antecedents. Molinists needn’t think that Framboise
can do something about counterfactuals where the antecedent
describes a situation she’s in fact never in. All they need insist
 is that, if she were in such a situation, she would have power
over both the counterfactual and her own action.

33. In addition to chs. 8–11 of Divine Providence, see esp. William

Lane Craig,’ “No Other Name”: A Middle Knowledge
Perspective on the Exclusivity of Salvation Through Christ’,
Faith and Philosophy 6 (1989), 172–88, as well as his
‘“MenMoved by The Holy Spirit Spoke from God” (2 Peter 1:
21): A Middle Knowledge Perspective on Biblical Inspiration’,
Philosophia Christi 1 (1999), 45–82. Regarding original sin’, see
Michael C. Rea, ‘The Metaphysics of Original Sin, in Peter van
Inwagen and Dean Zimmerman (eds.), Persons: Human and
Divine (Oxford: Clarendon, 2007), 319–56. For the classic
presentation of a ‘free will defense’ in response to the
argument from evil, see Plantinga’s The Nature of Necessity, ch.
9.

34. See e.g. my ‘“A Death He Freely Accepted”: Molinist Re ections

on the Incarnation’, Faith and Philosophy 18 (2001), 3–20, as
well as my ‘The Possibilities of Incarnation: Some Radical
Molinist Suggestions’, Religious Studies 37 (2001), 125–39. For
a very di erent Molinist view, see William Lane Craig, ‘Flint‘s
Radical Molinist Christology Not Radical Enough’, Faith and
Philosophy 23 (2006), 55–64.

35. The lyrics are by Sheldon Harnick.

36. My preference is to include instances of prophecy in the

antecedents. Space constraints, though, preclude a fuller
 discussion of this issue.

37. My thinking on this subject has been prompted by remarks by

Brian Leftow and Robin Collins.

38. See Alvin Plantinga, ‘On Ockham’s Way Out’, Faith and
Philosophy 3 (1986), 235–69.

39. See e.g. Robert Adams’s remark in n. 22 of the version of

‘Middle Knowledge and the Problem of Evi’ printed in his The
Virtue of Faith and Other Essays in Philosophical Theology (New
York: Oxford University Press, 1987), 77–93. See also Hasker s
Providence, Evil and the Openness of God, 197.

40. Hasker, Providence, Evil and the Openness of God, 195.

41. Ibid. 195.

42. For a sustained argument to this e ect, see Jensen, The

Grounding Objection to Molinism.
 CHAPTER 13

PETITIONARY PRAYER

SCOTT A. DAVISON

TRADITIONAL theists believe that there exists an all-knowing, all-

powerful, perfectly loving, and perfectly good God. They also
believe that God created the world, sustains it in being from
moment to moment, and providentially guides all events, in
accordance with a plan, toward a good ending. Historically, most
traditional theists have believed that God sometimes answers
prayers for particular things. In keeping with the literature on this
subject, I shall call such prayers ‘petitionary prayers’.

In this chapter, I discuss several problems related to the concept

of the traditional theistic God’s answering petitionary prayers. To
simplify matters, and for the sake of convenience, I shall speak as if
traditional theism were true, but I should not be understood to be
arguing or claiming this is so.1 Also, I shall not discuss here any
speci c teachings about petitionary prayer derived from any
particular theistic religious traditions.2 Finally, wherever possible, I
shall avoid speci c assumptions about what traditional theists
should say about the nature and extent of divine power, knowledge,
and goodness, although at certain junctures in the argument it may
be impossible to avoid some commitments in these areas.

1. SOME FLAWED ACCOUNTS OF ANSWERED PRAYER

In what circumstances might it be true to say that God has answered

a prayer? Discussions of the di culties surrounding petitionary
prayer in the literature typically ignore the complexities involved in
providing an adequate answer to this most basic question. I shall
begin with a simple example.

Suppose that a particular person prays that a certain event

happen, and that the event in question actually happens. Should we
say, in this kind of case, that God has answered the prayer? Not
necessarily. For example, suppose that you and I are competing in a
science contest, and I pray that you will be struck by lightning and
incapacitated so that I can win the contest. Imagine now that you
are struck by lightning and incapacitated, exactly as I imagined you
might be when I prayed about it. Does this mean that God answered
my prayer?

No. It is an important restriction on answered prayers that they

concern only good things. This restriction follows from the nature of
God’s moral perfection. God would not cooperate in the production
of evil for its own sake, or for bad reasons. In terms of the example
just described, it would be wrong for God to strike you with
lightning and incapacitate you just so that I could win the science
contest, unless something else were at stake in this situation. So if
people pray for bad things to happen, and they actually do happen,
then we should not conclude that the prayers in question were
answered. In other words, what we might call a ‘pure correlation’
account of answered prayer is mistaken. God answers only good
prayers, we might say, or at least prayers for things that are not bad.
(Someone might suppose that there are neutral things, things that
are neither good nor bad; I shall ignore this point in what follows,
because nothing of substance turns on it.) It is just a coincidence
that bad things sometimes happen after people have prayed for
them to happen.

Consider instead the following suggestion about when it is true

to say that God answers a prayer: whenever a person prays for
something good to happen, and it happens, then God has answered
a prayer. We could call this a ‘modi ed correlation’ account of
answered prayer. Is this suggestion satisfactory? Not really. After all,
in the same way that there can be coincidences involving prayers for
bad things, there can be coincidences involving prayers for good
things, too. It could happen, for example, that the event in question
was a very important part of God’s plan for the world, in such a way
that God’s bringing it about was not related at all to anyone’s
prayers.

So just because a person prays for something good and it

happens, this by itself is not su cient for saying that the person’s
prayer was answered. What else is necessary? Taking a cue from the
cases of coincidence mentioned above, perhaps what is needed is
some condition such as this: in order for God to answer a prayer, it
must be the case that if the person had not prayed for the event in
question, then it would not have occurred. Let us call this a
‘counterfactual dependence’ account of answered prayer. Almost
every treatment of the topic of petitionary prayer in the literature
endorses (or assumes) something like the counterfactual dependence
account.3 Is this approach satisfactory?

No. On the one hand, this account demands too much. Imagine—
just for the sake of the argument; I shall argue below that this
scenario is problematic—that I pray for a certain person to recover
from a serious illness, and that God answers my prayer by healing
this person. But suppose that as it happens, if I had not prayed for
this, then you would have prayed for it instead, and God would
have answered your prayer by healing this person in exactly the
same way. Do we really want to say that my prayer was not
answered in the actual sequence of things, just because the recovery
in question would have happened even if I had not prayed for it?
Clearly not. So in order for a prayer to be answered, it is not
necessary that if the person had not prayed for the event in
question, then it would not have occurred.4

On the other hand, the suggested account is also too weak.

Suppose that I pray for something to happen, and it does, and it
would not have happened if I had not prayed for it to happen. This
is not su cient for saying that my prayer was answered, because
there are many ways in which it might be true that the event in
question would not have happened if I had not prayed for it to
happen. For example, it might be the case that the event in question
was somehow caused by my very act of praying. Lawrence Masek
(2000: 279) provides an example of this (unintentionally, I believe)
in the following passage, which is about helping distant victims of a
hurricane:

Perhaps my prayer for the hurricane victims makes me more aware

of their su ering, which leads me to donate money to help them.
My friend might see this action and donate money, and his friend
might see his action and do likewise. Hence, my prayer can lead to
comfort for the hurricane victims that would not have occurred
without my prayer.

If something like this were to happen, then it would be true that the
event in question would not have happened if I had not prayed for
it. But in this kind of case, we need not say that my prayer was
answered by God, since the very act of praying for the victims could
lead to comfort for the hurricane victims all by itself, even if God
did not exist.5 So in order for a prayer to be answered, it is not
su cient that if the person had not prayed for the event in question,
then it would not have occurred.

What we should say here, I think, is something like the

following: A person’s prayer for something is answered by God if
and only if (1) the person prays for the thing in question, (2) the
thing in question is good,6 (3) God brings about the thing in
question, and (4) God brings about the thing in question at least in
part because the person prays for it.7 In order to make sense of
condition (4), we will need to appeal to something like God’s
reasons for doing things.8 I shall call this type of account a ‘reasons
account’ of answered prayer.

2. A REASONS ACCOUNT

A reasons account of answered prayer is clearly superior to a pure

correlation account, a modi ed correlation account, and a
counterfactual dependence account. But it isn’t completely clear as
it stands. The most pressing problem involves trying to make sense
of condition (4): what does the word ‘because’ mean here? One of
the di culties we face stems from the fact that creatures have no
direct causal impact upon God, at least as traditionally conceived.
Another, more pressing di culty stems from the fact that since God
is good, God already has a reason to bring about all good events,
whether or not anyone ever prays for them to occur. So condition
(2), which requires that answered prayers concern only good events,
actually makes it harder to understand the word ‘because’ in
condition (4).

Compare a related situation concerning the free and rational

choices of human beings. Suppose that I face a choice between
several alternatives, each of which is very good. For example,
imagine that I can choose to attend medical school, to attend law
school, or to attend art school, but that I can choose to do only one
of these things. Suppose further that there is no morally overriding
reason for me to choose any particular option. What should I do?

Let us suppose that in the end, I chose to attend art school. How
might I explain why I made this decision? I will want to nd a
statement of the form ‘I chose art school because—’, where I can ll
in the blank with a plausible reason. What constraints would we
want to impose upon di erent ways of trying to complete this
statement? At the very least, we would like the statement as a whole
to be a true description of my reasons for choosing art school over
the other alternatives.

There is a danger, though, that my explanation might work too

well. If the explanation works too well, then it will seem impossible
that I should have chosen medical school or law school instead. That
would appear to rob me of my freedom, at least according to a
libertarian conception of freedom.9 For example, if I were to say
that I chose to attend art school because I was pathologically afraid
of both doctors and lawyers, then it would seem that I did not
choose art school freely at all.

If my choice to attend art school is rational, it must be a

reasonable response to the value of attending art school. But if my
choice to attend art school is free, then there must be alternative
options open to me. There is a vast literature on this topic, the
problem of trying to explain how actions might be both rational and
free at once.10 I don’t have anything new to say about this question,
except to point out that the same question arises in an especially
acute way in connection with the idea of answered prayer, since
God’s decisions are supposed to be both rational and free.

On the one hand, God’s decisions are supposed to be rational

because they are always made in light of a perfect grasp of whatever
reasons or values are at stake in a situation.11 The traditional
doctrine of divine providence implies that God’s actions in the world
are not random, but instead based upon knowledge and love,
guiding creation wisely in accordance with a good plan. At the very
least, God’s decisions with respect to prayers are rational in the
sense that God answers prayers only for what is good (or at least not
bad).

On the other hand, traditional theists typically suppose that

God’s decisions with respect to prayers must be free because
creatures have no direct causal impact upon God and no way to
compel God to act in any particular way. Some people have a rather
magical view of the power of prayer according to which God is
literally compelled to answer certain prayers, but this view is highly
at odds with traditional theism.12

But if God’s decisions with respect to prayers are both rational

and free, then it is especially di cult to understand what it means
to say that God answers a prayer. If I pray for something good to
happen, then God already has a reason to bring it about, whether or
not I pray for it, since it is a good thing. Suppose that I do in fact
pray for something good and that God in fact brings about the thing
in question. In order for God’s bringing about this thing to count as
an answer to my prayer, as noted above, it must be the case that God
brought this about at least in part because I prayed for it. But what
does this mean?

Let us suppose that, independently of my prayers, God has a

number of reasons for bringing about the good thing in question.
Now when I pray for this thing, God has an additional reason to
bring it about, which I shall call reason x. When we say that God
brings about the thing in question at least in part because of x (that
is, because I prayed for it), we are saying something about the role
that x plays among God’s reasons for bringing about the thing in
question. How central a role must x play in God’s reasons for
bringing about the thing in question in order for God’s bringing it
about to count as an answer to my prayer?

It is hard to be precise here, of course. Even in the case of human

beings, where we feel more con dent conceptually, it is hard to
answer similar questions. For example, suppose that my neighbor
asks me to remove some poison ivy from my fence that faces his
yard, and imagine also that I have many reasons for doing this
already. If I do decide to remove the poison ivy from the fence,
what role must the reason provided by my neighbor’s request play
in my decision in order for my action to count as a response to my
neighbor’s request? There is the clear case where the neighbor’s
request plays no role at all (so that my action would not count as a
response to the request), and the equally clear case where the
neighbor’s request plays a necessary role (so that my action would
count as a response to the request). But in between these cases, so to
speak, there are many unclear possibilities.

In what remains of this chapter, I shall develop in more detail

the problem related to petitionary prayer just outlined, along with a
few others. Thanks to the reasons account of answered prayer, it is
possible to formulate and evaluate these problems with a degree of
precision that was previously not possible. These problems are deep,
perhaps even intractable, but I shall argue in the end that traditional
theists should not nd this result very troubling.13

3. THE DIVINE FREEDOM PROBLEM

If we consider more carefully the claim that God’s decision whether

or not to answer a prayer is free in the libertarian sense, then we
run into a problem.

Imagine for the sake of the argument that someone asks God to
bring about the occurrence of some event E. Suppose also that God
has a number of good reasons for bringing about E already, and now
that someone has prayed for E to occur, God has yet another reason
for bringing about E.14 To complete the story, let us imagine that
God freely brings about the occurrence of E (in the libertarian sense
of ‘freely’). Could this qualify as a case of answered prayer?

Since God brings about E freely, God could have decided instead
not to bring about E, in those very same circumstances, even though
God had all the same reasons for bringing about E that God actually
possesses.15 But then how can the sum total of God’s reasons for
bringing about E possibly explain God’s decision to bring about E?
After all, those same reasons are compatible with God’s choosing
not to bring about E, so they do not explain why God chose to bring
about E as opposed to choosing not to bring about E.16 Since the
sum total of God’s reasons for bringing about E cannot explain why
God chose to bring it about, neither can any subset of those reasons,
including the o ering of the prayer for E. So it is not the case that
God brought about E at least in part because of the prayer, which
implies that this is not a case of answered prayer after all. Let us call
this ‘the divine freedom problem’ of petitionary prayer.

The divine freedom problem of petitionary prayer is a speci c

instance of the general problem of how reasons can possibly explain
free action, and it clearly stems from a libertarian account of
freedom. This problem does not arise for an important alternative
picture of human action, according to which reasons are causes of
action, and hence explain it in a very straightforward way.17 Of
course, there is considerable controversy (to say the least) over
which kind of account of human action is the best one, over
whether or not human beings are free in the libertarian sense. But
traditional theists seem committed to the claim that God is free in
just this sense. This is because of the traditional view that creation is
free (as opposed to a kind of necessary emanation, of the sort
described by Neoplatonists) and the obvious fact that nothing in the
world could determine God’s actions.18 So the divine freedom
problem is a serious problem for traditional theists who believe that
God could answer petitionary prayers.19

4. THE DIVINE GOODNESS PROBLEM

Some traditional theists have held the view that God must always
do what is best,20 that God is obligated to do everything possible to
maximize value in every situation.21 If this were so, then of course
God would be obligated to do anything for which anyone prays if
such prayers happen to specify the maximum value available in a
given situation. But in nearly all these cases,22 God will be obligated
to bring about those same states of a airs even if nobody prays for
them,23 which implies that God’s bringing about those states of
a airs would not qualify as answers to prayers, since the o ering of
prayer would not play an important enough role among God’s
reasons for bringing them about (according to the reasons account
of answered prayer developed above).

Even if we set aside this highly stringent view of God’s

obligations, still one might wonder why God would not bring about
something good in a given situation simply because nobody prayed
for it.24 It is helpful to view this question as identifying a version of
the problem of evil. In response to the problem of evil, traditional
theists often argue that particular evils are permitted by God in
order to realize greater goods or to prevent worse evils. Defenders of
the practice of petitionary prayer typically adopt this strategy as
well (see below, sect. 6). But as I shall argue, the explanations of the
necessity of petitionary prayer developed in the literature to date
are completely unconvincing. Apart from the speci c teachings of
any particular religious tradition, which I shall not discuss here,
there seems to be no reason to expect the God of traditional theism
not to bring about some speci c good just because no prayers were
o ered for it. But if God would have brought about the good things
for which people prayed anyway, even if prayers had not been
o ered, then it follows from the reasons account that God’s actions
do not count as answers to those prayers, since the o ering of the
prayers cannot have played a signi cant enough role in God’s
decision. Let us call this ‘the divine goodness problem’ of petitionary
prayer. Since traditional theism includes the claim that God is
perfectly good (and perfectly loving), the divine goodness problem
is a serious problem for traditional theists who believe that God
could answer petitionary prayers.

5. THE REASONS-SKEPTICAL PROBLEM

Given the complexity involved in saying that God answers a prayer,

as indicated by the reasons account developed above, it seems
impossible, apart from direct revelation, for any human being ever
to know (or even be justi ed in believing) that any particular
prayers have ever been answered. In order to see that this is so,
suppose that every event is produced by God in some sense, as
traditional theists have typically insisted.25 Even if this is the case,
apart from direct revelation, there is no way to have enough access
to God’s reasons for bringing things about to be able to distinguish
between answered prayers, on the one hand, and events that God
brought about for other reasons after prayers happened to have
been o ered, on the other hand.26 And since no human being could
possibly distinguish these things from each other, the only
reasonable thing for human beings to do would be to withhold
judgment on the matter. Let us call this the ‘reasons-skeptical
problem’ of petitionary prayer.

Even in the case of so-called miraculous events, where one might

think that there are good reasons for believing that God is the direct
cause of some particular event, still one would have no reason to
think that God had done those things because of prayer. In fact, the
more impressive the miracle, and the more good that it
accomplishes, the more likely that it would be just the kind of thing
that God would have done anyway, whether or not anyone had
prayed for it to occur.27

Exceptions to this general rule about miraculous events might

include the cases discussed by Michael J. Murray28 (Elijah’s
confrontation with the prophets of Baal in 1 Kings 18) and Thomas
P. Flint29 (the healing of the blind man by St Peter in Acts 3). In
these cases, one could argue (as Flint does) that the o ering of the
prayer actually changes the circumstances in which God acts, and
hence makes a di erence by ‘raising the stakes’ in a situation, so to
speak. But then God is responding in these cases to the change in
the circumstances caused by the o ering of the prayer (the raising
of the stakes), rather than responding to the prayer per se. In other
words, it is not the fact that a prayer was o ered for E that explains
why God brings about E in these cases; had a di erent kind of event
(namely, one that did not involve prayer at all) changed the
circumstances in just the same way by raising the stakes, then it
would have given God the very same reasons for acting.

In addition, the unusual cases mentioned by Flint and Murray

are clearly the exception rather than the rule. From the perspective
of a traditional theist, the pattern exempli ed in these cases is not a
pattern that one should be encouraged to create on one’s own. One
could even argue that deliberately trying to raise the stakes in such
situations is a way of trying to force God to act, or that it involves
the kind of testing of God that is forbidden by the
Judaeo/Christian/Islamic theistic religious traditions.30 By contrast,
in the typical cases in which traditional theists believe that their
prayers have been answered, there is no convincing evidence that
something miraculous has occurred, no good reason to think that
the event in question would not have happened if no prayers had
been o ered, and most importantly, no way to discern what role
prayers might have played in God’s decisions (and hence no way to
tell whether a given event is an answer to prayer or not).

A di erent version of the skeptical problem has been noted in

the literature before, a version based on a counterfactual
dependence account of answered prayer. This version of the
problem is not quite as powerful as the reasons-skeptical version
developed above (which emerges only after a reasons account of
answered prayer has been identi ed). Even so, the force of this
weaker skeptical problem has been underestimated. In a reply to
David Basinger, Michael Murray states this problem clearly (‘How
could we ever know that this apparent answer to prayer would not
have happened anyway, even if nobody had prayed for it?’), and
then argues in reply that in the case of Elijah and the prophets of
Baal, ‘the indirect evidence makes it clear that the consumption of
the sacri ce was a response by God to Elijah’s petition’.31 He adds
that ‘Many theists claim, similarly, that indirect evidence makes it
plausible that particular events have occurred in response to their
petitionary prayers. And while these judgments certainly will be
false in some cases, there is no reason to think that they are always,
or even often, unjusti ed’ (Murray 2004: 265). Murray’s empirical
claim about the quality of evidence available to many theists is
simply false. The typical person never has good ‘indirect evidence’
for answered prayer, let alone anything that remotely resembles the
kind of evidence available in the case of Elijah and the prophets of
Baal.

Murray also claims that ‘God enlightens the mind of the

petitioner to make certain features of the world salient (features
related to the provision or failure thereof), and to see the reasons for
the provision or its failure’ (ibid. 249). But he provides no reason to
think that God does this on a regular basis (or fails to do it when the
prayers in question were not answered but God decided to bring
about the events in question anyway). In other words, Murray seems
to suggest that God enlightens the minds of petitioners just because
something like that would be required in order for them to know
that their petitions have been answered, but this is clearly ad hoc.32

All of this means that Murray’s answer to the weaker skeptical

problem is unsatisfactory. More importantly, there seems to be no
way at all to answer the stronger reasons-skeptical problem of
petitionary prayer, and this has important implications for the
justi cations of the practice of petitionary prayer o ered in the
literature to date.

6. IMPLICATIONS OF THE SKEPTICAL PROBLEM

Since it is impossible to know whether or not a given prayer has

been answered, apart from direct revelation, many of the rational
justi cations for the practice of petitionary prayer o ered in the
literature collapse. For example, Murray argues that petitionary
prayer keeps us from the idolatrous belief that we are self-su cient.
This is because God makes the provision of certain goods dependent
on the o ering of prayers, and being ‘truly dependent on petitioning
is what allows many, and maybe all, to “recognize God as the
ultimate source of all the goods that we enjoy” in the rst place’
(ibid. 247). Like a father who gives his son an action gure only if
he asks for it, God provides certain things for believers only if they
ask for them (ibid. 246–7), and teaches us ‘a number of things about
[H]is own good nature and purposes in the world by responding one
way or another to our petitions’ (ibid. 249). In a similar vein, Isaac
Choi claims that seeing some of our prayers ‘clearly and objectively
answered’ can serve as a powerful reminder of God’s reality, nature,
and love (Choi 2003: 12).

But even when ‘the antecedent probability of what we requested

happening by chance is extremely low’ (ibid. 12 n. 35), this does not
provide good evidence for the conclusion that God has answered our
prayers. This is because we are never in a position to say what the
antecedent probability of a given event is when considered in the
light of God’s reasons and knowledge of a situation (since we do not
have access to God’s reasons or knowledge of a situation).33 And so
we cannot draw the lessons Murray and Choi describe from our
experiences with any justi cation.34

One objection that might be o ered in response to Murray’s

explanation of the necessity of petitionary prayer is that many
people seem to ourish without o ering any petitionary prayers at
all (the so-called ‘fat pagan’ objection). By way of reply, Murray
(2004: 253) says, ‘All we can infer from the model I have o ered is
that there are some times when God makes provision dependent on
petition, and in those particular cases, those who fail to pray will
fail to receive what is dependent on petition. Nothing in the
empirical evidence could show us that this never happens.’ By way
of reply, it only seems fair to note that nothing in the empirical
evidence could show us that this ever does happen, either. This is
because the empirical evidence is simply inadequate to justify a
conclusion either way.

In fact, Richard Swinburne, a defender of the practice of

petitionary prayer, actually claims that ‘if God answered all prayers
for the removal of bad states of a airs promptly and predictably,
that would become evident’, and this would be bad because it would
eliminate the ‘epistemic distance’ between ourselves and God.35
Swinburne (1998: 206) thinks that such distance is necessary in
order for us to have a free choice between good and evil; otherwise,
we would be like the child whose mother is watching, for whom the
temptation to do wrong is ‘overborne’. ‘The more uncertainty there
is about the existence of God, the more it is possible for us to be
naturally good people who still have a free choice between right and
wrong’ (ibid. 207). Even if Swinburne is wrong about the necessity
of epistemic distance for signi cant freedom, he echoes the point
that the available evidence for answered prayer is ambiguous at
best.

Swinburne defends petitionary prayer by arguing that ‘If human

responsibility is good, then this extension to it—of exerting
in uence on (though not of course compelling) God to change
things [through petitionary prayer]—would surely also be good’
(ibid. 115). But if it is impossible to know whether or not one’s
prayers are ever answered, then it seems unlikely that one is
responsible (in any substantial sense) for the results of answered
prayer. This is because in general, one’s degree of responsibility for
the obtaining of some state of a airs depends upon the degree to
which one could foresee its obtaining, the degree to which one
intended that it obtain as a result of one’s actions, and the degree to
which one’s actions contributed causally to its obtaining.36 So cases
in which one person petitions another person to act freely in speci c
ways over time, especially when one does not know the outcome of
such petitions, are cases in which one’s responsibility for the
obtaining of the state of a airs in question is dramatically
diminished.37

Eleonore Stump, in the most widely discussed treatment of our

question, argues that the practice of petitionary prayer keeps human
beings from being either totally dominated or spoiled by God, and
thus safeguards the divine/human friendship from one or another
kind of using.38 On the one hand, Stump claims that if God bestows
certain goods only in response to petitionary prayers, then this will
guard against unwelcome interference from God, which could
produce an unbalanced friendship leading to a slavish or ‘lackey-
like’ follower. By way of reply, it should be pointed out that if we
don’t know which goods God bestows upon us as a result of
petitionary prayer (as opposed to those goods God bestows on us
just because we need them, whether or not we ask), then we won’t
know whether or not God is ‘respecting our boundaries’, so to
speak.39

On the other hand, with respect to the second danger, that of

spoiling the creature, Stump (1979: 143) says the following about a
prospective petitioner:

If he gets what he prayed for, he will be in a position to attribute his

good fortune to God’s doing and to be grateful to God for what God
has given him. If we add the undeniable uncertainty of his getting
what he prays for, then we will have safeguards against what I will
call (for lack of a better phrase) overwhelming spoiling.

The problem with this suggestion, of course, is that if the petitioner

does not know whether his prayers have been answered, he is lucky,
or God was going to do these things anyway, then he will not be in
a position to attribute his good fortune to answered prayer after
all.40

All the attempts to provide a rational justi cation for the

practice of petitionary prayer discussed above are undermined by
the skeptical problem. But the remaining rational justi cations in
the literature are not above reproach for other reasons. For instance,
in connection with the idea that petitionary prayer tends to guard
against the idolatry of self-su ciency, Murray (2004: 246) says that
‘With each petition, the believer is made aware that she is directly
dependent on God for her provisions in life.’ It is worth noting that
the same justi cation could be given for the practice of o ering
prayers of thanks, which face none of the problems discussed in this
chapter (see sect. 7 for more on this). But as Choi points out, what
Murray says about petitionary prayers would be true even if God
never answered any of them, as long as people falsely believed that
they were necessary and answered.41 The same point applies to
Murray’s claims that God withholds certain goods until people pray
for one another, thus forcing interdependence and unity among
believers (ibid. 250) and providing them with a reason to share
their needs with one another (ibid. 252).

Choi claims that God might not maximize the goodness in every
human life in order to leave room for the improvement of our lives
due to divine rewards and acts of human love. E cacious
petitionary prayers express praiseworthy attitudes, which God
would naturally choose to reward (2003: 9–10).42 Choi claims that
‘[petitionary] prayerlessness betrays a practical lack of faith and
trust in God’ (ibid. 11), but as we shall see below, this is not
necessarily the case. And there are ways other than petitionary
prayer to manifest praiseworthy attitudes.

Choi also claims that petitionary prayers provide us with a way

to love those whom we cannot directly a ect (ibid. 9–10).
Ironically, C. S. Lewis points out that often people pray for others
when they should be helping them instead.43 But suppose that we
con ne our attention to prayers for those whom we cannot help
ourselves. In this case, either there are others who can help those in
need, or there are not. (1) If there are others who can help, then
they should; but if they choose not to, is there any point in asking
God to change their minds? If Choi thinks that there is a point in
asking God to change their minds, then his main point is undercut,
namely, the point that God does not maximize our well-being
because this provides an opening for creatures to improve each
other’s lives through acts of love. After all, if God is willing to
override my choices about whether or not to help in response to
your prayers, then God is not taking seriously my choices about
whether or not to improve the world. (2) Finally, if no people are in
a position to help those in need, Choi has given us no reason to
think that God would not provide for them, whether or not anyone
asks for this. Here we run into the problem of divine goodness again
(see above, sect. 4), and the problem of evil.

A more fundamental problem for Choi’s position stems from the

concepts of action and responsibility. Since God can freely decide
not to answer any particular prayer, as noted above, it is not clear
that we can describe answered prayers as acts of love performed by
the petitioners. The o ering of a petition might be an act of love
performed by the petitioner, but the answering of the prayer would
be an act of love performed by God, not by the petitioner. (If the
o ering of the petition is itself an act of love, then it is a good thing,
but this will be true whether or not it is answered; see sect. 7 below
for more on this.) Choi’s point is that God ‘leaves room’ for us to
make a di erence in the lives of others, but it would be the
answering, not the petitioning, that would make a di erence in the
world, and the petition does not cause the answering so long as God
is free. Hence God should get the credit for the answered prayer, not
the petitioner, which undercuts Choi’s point that God leaves room
for creatures to make a di erence in the world.
 7. HOW TRADITIONAL THEISTS SHOULD RESPOND

A reasons account of answered prayer seems to be the right kind of

account, but it leads to very serious problems of petitionary prayer
for traditional theists and undercuts all the extant defenses of the
practice in the literature. Suppose that there are good reasons to
believe that petitionary prayers are never answered, and that even if
this is wrong, still it is impossible to know when prayers are
answered. How should traditional theists respond to this
conclusion? Several responses seem appropriate.44

First, consider the claim that one’s relationship with God would
be de cient in some way without answered prayers.45 Why should
this be so? What should matter here is not that God brings about
good things for people in response to petitionary prayers, but rather
that God loves them and provides for them, whether or not they ask
for anything speci cally. There is something inappropriately
egocentric to insist that one be a cause (in some sense) of God’s
action in the world, rather than simply being grateful for God’s
provident care. Traditional theists believe that every good thing
comes from God (ultimately, if not directly), and none of the
problems outlined here with petitionary prayer applies to other
kinds of prayers, such as prayers of thanksgiving.

In fact, contrary to what Choi says, one could argue that

petitionary prayer re ects a lack of trust in God. Rather than
trusting in God’s love, power, and knowledge to care for everyone
(unless there is a good reason not to do so—see the literature on the
problem of evil), those o ering petitionary prayers seem to think
that God needs suggestions, reminders, or repeated badgering in
order to act. As noted (see the discussion in n. 13 above), these
things might be appropriate when dealing with fellow human
beings, but surely an all-knowing, perfectly loving God needs no
such prompting.

Dwelling on particular people and their needs is an expression of

love and concern all by itself, whether or not it is done in the
presence of God, so to speak, and whether or not it results in the
o ering of petitionary prayers. It would not be strange to nd the
heartfelt, thoughtful sympathy of an atheistic friend much more
comforting than the petitionary prayers of an inattentive theistic
friend, for example. Quite frequently we are unable to change the
world, but a large part of being a good person is standing
symbolically on the side of the good anyway.46 Quite often the best
we can do in a given situation is to express love and to identify
symbolically with those in need. It is interesting to note that these
things seem to be the main functions of the o ering of petitionary
prayers in public. These are very good things to do, even if prayers
are not e cacious in the sense of moving God to act.47

For many people, petitionary prayer is an involuntary response

in times of need, whether or not they believe that it will be
e cacious, and traditional theists will probably nd something
appropriate about this (even if they believe that such prayers do not
move God to act di erently). So petitionary prayer is not pointless,
even if it is not e cacious in the sense of moving God to act, since
it serves various important functions. As a number of authors have
suggested,48 perhaps we should think of prayer in general terms not
so much as the attempt to bring God’s mind into line with our minds
as the attempt to bring our minds into line with God’s mind.

But it would be a mistake to assume that failing to pray in the

petitionary way evinces a lack of concern for others. Instead, it
could be the case that one does not petition God on behalf of a
particular person’s need simply because one knows all too well how
pressing that need is, one knows that God sees this truth even more
clearly than one does, one knows that God loves this person even
more than one does, and one trusts that God will do what is best,
whatever that might be. If this is right, then it may actually betray a
lack of trust in God in some cases to pray for speci c things to
occur.

What should a traditional theist do who formerly believed in the

e cacy of petitionary prayer, but now nds it pointless because of
the sorts of arguments advanced in this chapter? Those who believe
in the e cacy of petitionary prayer sometimes pray that God would
take care of those for whom we have forgotten to pray. In a similar
vein, those who nd it di cult to pray in the petitionary way any
longer might o er a conditional prayer to the e ect that if they are
wrong about the pointlessness of prayer, then would God do all
those things for which they would have prayed had they continued
to believe in the practice of petitionary prayer? If it turns out that
the arguments of this chapter are mistaken and that petitionary
prayer is e cacious after all, then why wouldn’t God answer this
particular prayer? Why would God punish a person, in e ect, for
having false beliefs about petitionary prayer, as long as they were
based on (somewhat decent) arguments? Rather than taking God for
granted, such a prayer would express faith in God’s good will,
con dence in God’s providence, and trust in God’s power.

NOTES

For very helpful comments concerning earlier versions of this

chapter, including many clever objections that I could not
answer here, I wish to thank Thomas P. Flint, Michael Rea,
Eleonore Stump, Michael Murray, Ronald L. Hall, Paul Draper,
William Hasker, Kate Rogers, William Rowe, George
Mavrodes, and participants in the 2006 Meeting of the Society
for Philosophy of Religion in Charleston, South Carolina.
Research toward the completion of this paper was supported
by a course load reduction granted by Morehead State
University, for which I am also grateful. Finally, it is worth
noting that although I have argued here against the e cacy of
petitionary prayer based on general considerations, I am also a
practitioner of a particular theistic religious tradition that
a rms it, so I look forward to future debate concerning these
issues.
1. Of course, these questions would be interesting in their own
right even if traditional theism turned out to be false.

2. This restriction renders my discussion rather limited and

arti cial, I admit, but a complete discussion would require
something like a book-length treatment of the question.

3. For instance, see Swinburne 1998: 115; Flint 1998: 222, 226;
Murray 2004: 243; and Basinger 2004: 255. The exception to
the rule here is Flint (1998: 227 n. 22), who thinks that
counterfactual dependence is probably not su cient for
answered prayer.

4. The structure of this counterexample is due to Harry

Frankfurt’s well-known attack on the so-called principle of
alternate possibilities: see Frankfurt 1969.

5. We should draw the same conclusion in a case in which I

prayed for strength, and the very act of praying for strength
gave me strength all by itself.

6. As Michael Rea has pointed out to me, ‘good’ here must mean
something like ‘good, all things considered’, as opposed to
‘intrinsically good’.

7. Thomas P. Flint has pointed out to me that when I say that

God ‘brings about’ something, as I do in condition (4), this
should not be taken to imply that God directly causes the thing
in question. Instead, my use of ‘brings about’ is intended to
 include what Plantinga (1974:172–3) has called ‘weak
actualization’.

8. This means that we will need to appeal to that which Richard

Swinburne (1979) has called ‘personal explanation’, or what
others call ‘folk-psychological explanation’. For a lively
discussion of the di culties facing folk-psychological
explanations as applied to human beings, see Stich 1983.

9. According to a libertarian conception of freedom, an act is free

only if nothing makes that action causally necessary at that
time. For further discussion of this point, see Campbell 1957;
Frankfurt 1969; Chisholm 1982; Fischer 1982; van Inwagen
1983; Flint 1988; Rowe 1991, 2004; and Davison 1994a,
1999a.

10. For a few contemporary examples, see Campbell 1957;

Davidson 1963; Chisholm 1982; Fischer 1982; Dennett 1984;
Kane 1985; Fischer and Ravizza 1998; and Wolf 1990.

11. I shall ignore here the question of whether or not God creates
the values at stake in all situations (see Davison 1991a). I do
not assume in this chapter that God is morally obligated to
maximize the value in each situation; I shall say more about
this below, in sect. 4.

12. See the discussions of this point in Philips 1981: ch. 6;

Swinburne 1998: 115; and Flint 1998: 222; I say more about
divine freedom in sect. 3 below.
13. In addition to the theoretical problems discussed below, there
are two interesting practical puzzles that are worth noting that
arise in connection with petitionary prayer. For instance, it is
widely assumed that the better a thing would be, the more
likely it is that God would answer prayers for its occurrence.
But the reasons account of answered prayer suggests
otherwise. For the better something would be, the stronger
God’s reasons are for bringing it about anyway, and hence the
less likely it is that prayers played a signi cant role in God’s
decision. I call this ‘the puzzle of increasing value’. Another
practical puzzle has to do with what, exactly, a person should
pray for. Defenders of petitionary prayer are quick to point out
that God does not answer some prayers because answering
them would not be good for us, despite appearances to the
contrary (see Flint 1998: 217). Of course, it is often very
di cult to know what the consequences of any particular
event might be. Thinking about this tends to make one’s
prayers more and more general. More pressure in this
direction comes from God’s nature. Petitionary prayer
sometimes gives the appearance of o ering God advice about
what to do, reminding God about a situation, or trying to
explain to God why God should care about someone. All of
these things would make sense if another human being were
the object of such petitions, but God’s complete knowledge of
every situation and perfect love for everyone involved make
them inappropriate. So the more clearly one thinks about the
 nature of God, the more general one’s petitionary prayers
become, and then the less clear it is whether or not one’s
prayers make a di erence. I call this ‘the puzzle of
particularity’.

14. In this chapter, I shall speak as if God’s decision process has

stages that occur sequentially in time, but this is a dispensable
convenience; without a ecting the argument, we could
imagine instead that God sees from eternity or knows by
middle knowledge what a person will or would pray, and takes
this into account when deciding what to do. For more on the
notions of middle knowledge, providence, and the di culties
with this picture, see Flint 1988, 1998; Hasker 1989; Davison
1991b, 1999b, 2003, 2004, 2005.

15. For those who are fond of talking this way, if God acts freely in
the actual world, then there is a possible world in which God
does otherwise, and God’s reasons and circumstances are the
same in that possible world as they are in the actual world
before the point of the decision to do otherwise.

16. Here I am appealing to something like the following principle:

If an agent possesses reasons for acting R1, R2,…, Rn, and the
agent performs action A, and the agent could have performed
action B in exactly the same circumstances, including the
possession of all and only reasons R1, R2,…, Rn, then the
possession of reasons R1, R2,…, Rn does not explain why the
agent performed action A instead of action B.
17. See Davidson 1963; Goldman 1970; and Dretske 1988 for
examples of this type of account. Of course, such accounts face
di culties of their own (see Davison 1994a for a discussion of
Dretske’s account, for example).

18. I say here that ‘nothing in the world’ could determine God’s
actions, but some theists are inclined to believe that God’s
actions are determined by God’s own nature, thereby avoiding
both a libertarian account of God’s freedom and the view that
God’s actions are determined by something in the world.
(Thanks to Kate Rogers for reminding me of this fact.) This
view implies that if a person prays for the right thing in the
right circumstances, then given God’s nature, God cannot
refuse to answer the prayer. I nd this conclusion to be highly
at odds with traditional theism, but there is no doubt that the
problems of petitionary prayer may force traditional theists to
take a hard look at the nature of divine freedom. For more on
this question, see the discussions mentioned in n. 8 and the
detailed and provocative discussion of divine freedom in Rowe
2004.

19. Michael Rea has pointed out to me that there are several
responses one could make to this argument: (a) libertarianism
is false, in which case God needn’t be free in the libertarian
sense; (b) libertarianism is true, and the problems can be
solved (even though we can’t yet see how), in which case
there’s no special worry in the divine case. I can add two other
 possible replies to his list: (c) we can live with this problem in
the human realm, but not with the puzzles it creates with
respect to God and answered prayer; (d) traditional theists
who believe (naturally enough) that God is free in the
libertarian sense have inconsistent beliefs, but it’s not clear
what they should do about it. These are all interesting
responses. For the record, I am not arguing here that there is a
special problem for freedom and explanation in the divine
case, just that there is a problem in the divine case, given the
way that most traditional theists think about divine freedom,
and that the reasons account enables us to appreciate it in a
new way.

20. This section contains a re ned version of the argument

presented in Basinger 1983 and further re ned in Basinger
2004.

21. Here I shall ignore the di erence between saying that God
must do what is best and saying that God is obligated to do
what is best. Thanks to Michael Rea for reminding me of the
di erence.

22. I say that this is true in nearly all cases because Thomas P. Flint
has pointed out that there are cases in which the very o ering
of the prayer changes the circumstances, and hence changes
what would be the best outcome in a situation (see Flint 1998:
222). In these cases, though, God is responding to the change
in the circumstances caused by the o ering of the prayer,
 rather than responding to the o ering of the prayer per se, and
so they do not seem to be cases of answered prayer, either,
according to the reasons account. (I shall say more about this
kind of case below in sect. 5.)

23. As Basinger 2004: 260 . points out. Basinger develops what I

have called the divine goodness problem of petitionary prayer
in more detail than I do here, and in a slightly di erent
direction, by discussing in detail several di erent views of
divine obligations.

24. This line of thinking suggests that the divine goodness problem
might lead to another practical problem for petitionary prayer
for those who nd it hard to believe that God would adopt this
sort of policy; see n. 13 above for a discussion of other
practical problems of prayer.

25. Traditional theists have insisted on this by embracing doctrines

of divine creation, conservation, and concurrence with the
operation of secondary causes; for a philosophical introduction
to these doctrines, see the essays in Morris 1988.

26. This will be true given any plausible account of knowledge or

justi cation, whether internalist or externalist. From an
internalist point of view, human beings simply lack access to
information about God’s reasons for acting in speci c
circumstances, so nothing could justify our beliefs (however
true) about answered prayer. From an externalist point of
view, it seems completely implausible to suggest that there is a
 belief-forming process that is activated only when God
answers prayers, but not when God happens to bring about
that for which a person prayed, but for independent reasons.
For more on this, see the discussion below of Murray’s reply to
the weaker form of the skeptical problem and the helpful
discussion in Lehrer 1990.

27. See the discussion of the puzzle of increasing value in n. 13

above.

28. See Murray 2004: 249.

29. See Flint 1998: 222.

30. It is worth noting that we are very suspicious of people who

claim to be able to discern which prayers God will answer, and
not just because it is hard in general to know what outcomes
would be best in a given situation.

31. See Murray’s response to Basinger 2004 Murray (2004: 264).

32. Since there is no way to know that God regularly enlightens the
minds of petitioners when a prayer is answered (and fails to do
so when the prayers in question were not answered but God
decided to bring about the events in question anyway), one
could always wonder whether one’s belief that God had
answered a prayer was true or justi ed. If God had
enlightened one’s mind, then it would be justi ed (but only
according to an externalist conception of justi cation: see n.
26 above) and true. But if one had simply tried to explain to
 oneself what had happened after the o ering of a prayer in a
psychologically or spiritually satisfying fashion, even though
God had not enlightened one’s mind because God had brought
about the event in question for independent reasons, then
one’s belief would be false and unjusti ed.

33. Of course, we cannot know the antecedent probability (given

God’s reasons and knowledge of a situation) of God#8217;s
bringing about an event even if we do not pray for it; we also
cannot know the antecedent probability of the same event’s
occurring if we pray for it (given God’s reasons and knowledge
of a situation). So we cannot fruitfully compare the two, which
is what Choi’s suggestion would require.

34. It is worth noting in passing that people actually draw all kinds
of ridiculous conclusions about God on the basis of their
experiences involving prayer. The point here is that none of
these conclusions is justi ed.

35. Swinburne 1998: 118.

36. See Davison 1999a and 1994b ch. 5.

37. So if God really wants to extend human responsibility for the

world, there must be a better way.

38. See Stump 1979.

39. If in fact this is important: see the reply to Stump in Ho man

1985.
 40. But if he attributes all good things to God, as a traditional
theist should, then he will be thankful whether or not good
things come in response to answered prayer. I shall say more
about this below.

41. See Choi 2003. In practice, of course, people pray only for
those things that lie beyond their immediate control; nobody
prays that God would pass the salt, for instance.

42. See ibid. Lawrence Masek 2000: 274 . makes a similar point.

43. ‘It’s so much easier to pray for a bore than to go and see him’:
Lewis 1963: 66.

44. For another interpretation of petitionary prayer that does not

require its e cacy, see the provocative and insightful
discussion in Phillips 1981 esp. ch. 6.

45. Perhaps because it would degenerate into an ‘impersonal

deistic relationship’, as Choi 2003: 12 says.

46. For more on this, see Adams 2002.

47. In fact, it seems to me that many people (perhaps even most of

them?) actually o er petitionary prayers, even privately, in
order to identify symbolically with those in need, not to move
God to act. Nothing I have said in this chapter counts against
this practice.

48. See Lewis 1963, for example.

REFERENCES
ADAMS, ROBERT M. (2002). Finite and In nite Goods. Oxford: Oxford University Press.

BASINGER, DAVID (1983). ‘Why Petition an Omnipotent, Omniscient,

Wholly Good God?’, in Religious Studies 19: 25–42.

——— (2004). ‘God Does Not Necessarily Respond to Prayer’, in

Michael L. Peterson (ed.), Contemporary Debates in Philosophy of
Religion. Malden, Mass.: Blackwell, 255–64.

CAMPBELL, C. A. (1957). ‘Has the Self Free-Will?’ in On Selfhood and

Godhood. London: George Allen & Unwin, 158–79.

CHISHOLM, RODERICK (1982). ‘Freedom and Action’, in Gary Watson

(ed.), Free Will. Oxford: Oxford University Press, 24–35.

CHOI, ISAAC (2003). ‘Is Petitionary Prayer Super uous?’ unpublished

manuscript presented at the Eastern Division Meeting of the
Society of Christian Philosophers, Asbury College, 5 December.

DAVIDSON, DONALD (1963). ‘Actions, Reasons and Causes’ Journal of

Philosophy 60: 685–700; repr. in Donald Davidson (1980), Essays
on Actions and Events. Oxford: Oxford University Press.

DAVISON, SCOTT A. (1991a). ‘Could Abstract Objects Depend Upon

God?’ in Religious Studies 27 (Dec.), 485–97.

——— (1991b). ‘Foreknowledge, Middle Knowledge, and “Nearby”

Worlds’ International Journal for Philosophy of Religion 30/1
(Aug.), 29–44.

——— (1994a). ‘Dretske on the Metaphysics of Freedom’, Analysis

54/2, NS 242 (Apr.), 115–23.
——— (1994b). ‘The Metaphysics of Moral Responsibility’,
unpublished doctoral dissertation, University of Notre Dame.

——— (1999a). ‘Moral Luck and the Flicker of Freedom’, American

Philosophical Quarterly 36/3 (July), 241–51.

——— (1999b). ‘Divine Providence and Human Freedom’ in

Michael J. Murray (ed.), Reason for the Faith Within. Grand
Rapids: William B. Eerdmans.

——— (2003). ‘Divine Knowledge and Human Freedom’ in

Raymond Martin and Christopher Bernard (eds.), God Matters:
Readings in the Philosophy of Religion. New York: Longman, 12–
24.

——— (2004). ‘Craig on the Grounding Objection to Middle

Knowledge’ Faith and Philosophy 21/3 (July), 365–9.

——— (2005). ‘Prophecy’, Edward N. Zalta (ed.), The Stanford

Encyclopedia of Philosophy (Summer Edition),
<http://plato.stanford.edu/archives/sum2005/entries/prophec
y/>, accessed 19 May 2008.

DENNETT, DANIEL C. (1984). Elbow Room. Cambridge, Mass.: MIT.

DRETSKE, FRED (1988). Explaining Behavior: Reasons in a World of

Causes. Cambridge, Mass.: MIT.

FISCHER, JOHN MARTIN (1982). ‘Responsibility and Control’, Journal of

Philosophy (Jan.): 24–40; repr. in John Martin Fischer (ed.),
 Moral Responsibility. Ithaca, NY: Cornell University Press, 1986,
174–90.

——— and RAVIZZA, MARK, SJ (1998). Responsibility and Control: A

Theory of Moral Responsibility. Cambridge: Cambridge University
Press.

FLINT, THOMAS P. (1988). ‘Two Accounts of Providence’, in Thomas

V. Morris (ed.), Divine and Human Action. Ithaca, NY: Cornell
University Press, 147–81.

——— (1998). Divine Providence: The Molinist Account. Ithaca, NY:

Cornell University Press.

FRANKFURT, HARRY G. (1969). ‘Alternative Possibilities and Moral

Responsibility’, Journal of Philosophy 66: 829–39; repr. in id.,
The Importance of What We Care About. Cambridge: Cambridge
University Press, 1988, 1–10.

GOLDMAN, ALVIN (1970). A Theory of Human Action. Englewood Cli s,

NJ: Prentice-Hall.

HASKER, WILLIAM (1989). God, Time and Knowledge. Cornell University

Press.

HOFFMAN, JOSHJJA (1985). ‘On Petitionary Prayer’ Faith and

Philosophy 2: 30–7.

KANE, ROBERT (1985). Free Will and Values. Albany, NY: State
University of New York Press.
LEHRER, KEITH (1990). Theory of Knowledge. Boulder, Colo.:
Westmont.

LEWIS, C. S. (1963). Letters to Malcolm: Chie y on Prayer. New York:

Harcourt Brace Jovanovich.

MASEK, LAWRENCE (2000). ‘Petitionary Prayer to An Omnipotent and

Omnibenevolent God’, American Catholic Philosophical Quarterly
74: 273–83.

MORRIS, THOMAS V. (ed.) (1988). Divine and Human Action. Ithaca,

NY: Cornell University Press.

MURRAY, MICHAEL J. (2004). ‘God Responds to Prayer’ in Michael L.

Peterson (ed.), Contemporary Debates in Philosophy of Religion.
Malden, Mass.: Blackwell, 242–54.

PHILLIPS, D.Z. (1981). The Concept of Prayer. New York: Seabury.

PLANTINGA, ALVIN (1974). The Nature of Necessity. Oxford: Oxford

University Press.

ROWE, WILLIAM (1991). Thomas Reid on Freedom and Morality. Ithaca,

NY: Cornell University Press.

——— (2004). Can God Be Free? Oxford: Oxford University Press.

STICH, STEPHEN (1983). From Folk Psychology to Cognitive Science: The

Case Against Belief. Cambridge, Mass.: MIT.

STUMP, ELEONORE (1979). ‘Petitionary Prayer’ American Philosophical

Quarterly 16 (Apr.), 81–91.
SWINBURNE, RICHARD (1979). The Existence of God. Oxford: Clarendon.

——— (1998). Providence and the Problem of Evil. Oxford: Oxford

University Press.

WOLF, SUSAN (1990). Freedom Within Reason. New York: Oxford

University Press.

VAN INWAGEN, PETER (1983). An Essay on Free Will. Oxford: Oxford

University Press.
 CHAPTER 14

MORALITY AND DIVINE AUTHORITY

MARK C. MURPHY

THE QUESTION FORMULATED

I will discuss morality and divine authority in the context of the

question of whether God—that is, God’s existence, nature, or
activity—explains morality.1 By way of introduction, let me make
some clarifying remarks about what I will mean by ‘God’, ‘morality’,
and ‘explains’.

By ‘God’ I mean an absolutely perfect being. I will assume such a

being to have the traditional divine perfections—necessity, aseity,
omniscience, omnipotence, goodness, and freedom—and I will
assume that this being is a creator, at least of every contingent
being. When we say that God has some property or stands in some
relation, sometimes we mean that any being that quali ed as God
has that property or stands in that relation; sometimes when we say
that God has some property or stands in some relation, we mean
that there is a being, whom we call ‘God’, who is thus quali ed and
who has that property or stands in that relation. I will reserve the
expression ‘fact about God’ or ‘theistic fact’ for those facts whose
obtaining we mean to assert when we speak of God in the second
way; by ‘fact about God’ or ‘theistic fact’ I mean a fact that there is
an omniscient (etc.) being who has that property or stands in that
relation.

By ‘morality’ I mean the set of all of the valid norms of the form
‘x is morally required to φ’, where x ranges over human beings and
φ over action-types, and where to be morally required is to be
categorically reason-giving, overriding, and justi able from an
impartial point of view.2 (I assume this set of norms is nonempty.)
Some such norms may be universally quanti ed, and unquali ed
(e.g. for all x, x is morally required to φ); some such norms may be
universally quanti ed, and quali ed (e.g. for all x, any x who is P is
morally required to φ); some such norms may be particular (e.g. A
[who is an x]is morally required to φ). To be a moral fact is to be an
obtaining state of a airs of the form ‘x’s being morally required to
φ’, whether of this universal or particular form. On this stipulative
de nition, then, assuming that they obtain, all the following states
of a airs are moral facts: everyone’s being morally required not to kill
the innocent; the well-o ’s being morally required to aid to the less-well-
o ; and Bob Dylan’s being morally required not to break promises.

I cannot give an adequate account of what it is for one fact to

explain another. But when we ask whether God explains morality,
we are asking whether there is an answer to the question ‘Why is
morality as it is?’ and, if so, whether that answer includes facts
about God. One addition that I can make to this ordinary
understanding is that the answer to this ‘Why’ question must appeal
to some relationship between facts that is not observer-relative, that
is, that is not dependent on one’s mastery or failure of mastery of
concept-use and that is not dependent on one’s knowledge or
ignorance of the relevant facts. This is also stipulative. One fairly
common understanding of ‘explains’ is at least partly epistemic: to
explain is, on this understanding, to give new information. But this
more epistemic reading of ‘explains’ is parasitic on its metaphysical
aspect: the way that explanations give new information is by
making one aware of, or calling one’s attention to, some relationship
between facts that would hold independently of such awareness. It
is that relationship between facts about God and facts about
morality that I want to explore in this chapter.

So the question is whether God explains morality, understanding

‘God’, ‘explains’, and ‘morality’ in these particular senses. It will be
helpful to have a particular thesis to evaluate. Here it is:

Theistic Explanation of Morality (ThEM): For every moral fact,

there is some fact (or facts) about God that explains it.

Defenses of this thesis might appeal to rather di erent sorts of

relationship between moral and theistic facts, and I catalog some of
those di erences below. But here let me note a distinction between
two broad classes of theistic explanation of morality. Suppose that
we distinguish between grounds of moral requirements and their
validating conditions.3 This distinction between grounds and
validating conditions cannot be spelled out in logical terms alone
(for example, in terms of necessary or su cient conditions); the
distinction is, rather, that between a cause, or source, and the
circumstances in which a cause can operate or a source can issue its
product.4 Would-be defenders of ThEM might choose to support
ThEM by holding that theistic facts are needed in order to explain
how morality is grounded; alternatively, they might choose to
support ThEM by holding that theistic facts provide for the
validating conditions for moral norms. My discussion below focuses
exclusively on how theistic facts might ground moral requirements,
but philosophers have also followed this alternate route of appeal to
theistic facts in order to o er validating conditions explanations for
morality. For example: Kant holds that the condition for moral
requirements to be binding on rational agents such as us, who also
naturally aim at our own happiness, is that there exists a being who
is capable of ensuring that virtue is proportioned to happiness. He
also holds that a condition of our being capable of acting in
according with the demands of morality—and this is important to
validate moral requirements, as we cannot be bound to do what we
are unable to do—is that we receive divine assistance.5 This Kantian
account is at least a partial theistic explanation of morality: God, on
this view, makes the world safe for morality, makes the world a
place in which moral norms can exert their due normative force. But
I will be concerned here with God’s explanatory role in the
grounding of moral norms rather than in ensuring that their
validating conditions obtain.
 EXPLAINING MORAL FACTS: MORAL SUBSUMPTION

Our question is whether every moral fact is (or must be, or even can
be) explained by facts about God. What, though, are the available
strategies for explaining moral facts?

Consider the following moral fact, and how one might explain it.
It is a fact that Mark Murphy is morally required to show up to
teach his classes at Georgetown University. There is an obvious
explanation for this fact. The obvious explanation is that people are
morally required to do what they agree to do, and Mark Murphy has
agreed to show up to teach his classes at Georgetown University.

This moral fact and its explanation are trivial. What is important
is the structure of the explanation for moral facts suggested by the
example. One way that we often explain a moral fact is by
subsuming that moral fact under another moral fact of broader
extension, together with a supplementary non-moral fact. The
relevant relationship between the more particular moral fact to be
explained and the more general moral fact that does the explaining
is that of instantiation, in that the more particular moral fact is an
instance of the more general fact: because Mark Murphy has made
the agreement to show up to teach his classes, Mark Murphy’s being
morally required to show up to teach his classes is an instance of Mark
Murphy’s being morally required to do what he has agreed to do. Call
this model of explanation of moral facts explanation by way of
moral subsumption.6
 I want to make three points about this model of explanation. The
rst is that one might well explain a number of moral facts by
appeal to non-moral facts about God’s existence, nature, or activity
conjoined with a general moral fact about how persons are morally
required to respond to God or to facts about God’s existence, nature,
or activity. One might claim that it is wrong to torture humans
because it is wrong to express disrespect for what is an image of the
divine perfection, humans are images of the divine perfection, and
to torture expresses disrespect for humans. One might claim that it
is wrong to ignore the plight of those undeservedly in need because
God has commanded us to look after those undeservedly in need,
and people are morally required to do what God commands. It is
clear that in these cases, the requirements not to torture and to
assist the needy are explained by appeal to non-moral facts about
God (humans are images of God; God has given a certain command)
that allows a more speci c moral norm to be subsumed as an
instance of a more general norm.

The second point is that this model of explanation is obviously

intrinsically limited. If we can explain moral facts by appealing to
more general moral facts and a non-moral fact regarding
instantiation, then it is clear that there will remain some moral facts
—at least one—that cannot be thus explained using this model of
explanation. (I will consider in more detail the question of the
extent to which moral facts can be explained by way of subsumption
under the moral fact people are morally required to obey God later in
this chapter.7)

The third point is a conditional one. Suppose that we consider

the set of moral facts that are explainable in this way. These are
moral facts that can be subsumed under more general moral facts. It
seems plain that even if a moral fact can be explained via moral
subsumption without appeal to God’s existence, nature, or activity,
if it were the case that the general moral fact under which it is
subsumed is explained by God’s existence, nature, or activity, then
the particular moral fact counts as being explained by God’s
existence, nature, or activity.8 In the example above, the obtaining
of Mark Murphy’s being morally required to teach classes is explained
by moral subsumption under the obtaining of the more general state
of a airs people’s being morally required to honor their agreements.
There is no theistic appeal here, and one might say that it thus
follows that there are at least some moral facts—for example, that
Murphy is morally required to teach his classes—that are not
explained in theistic terms. But given the conditional point above,
we cannot conclude this. For if people’s being morally required to
honor their agreements is a state of a airs whose obtaining is
explained by appeal to God’s existence, nature, or activity, then the
obtaining of the state of a airs Mark Murphy’s being morally required
to teach his classes should count as being theistically explained as
well.
 EXPLAINING MORAL FACTS: THE BRUTE AND THE SELF-EXPLANATORY

Let us turn, then, to those moral facts that are not explainable by
way of moral subsumption. Suppose it is true that all these moral
facts are theistically explained. If so, then ThEM will be true. For
every moral fact explainable by way of moral subsumption is
ultimately explained by way of some fact not explainable by moral
subsumption. If all the facts not explainable by moral subsumption
are theistically explained, and if the conditional point in the
previous paragraph is correct, then ThEM will turn out to be true.

For each moral fact not explainable by way of moral

subsumption, there are, prima facie, several possibilities with
respect to its explanation. That moral fact might be simply brute,
lacking any explanation. It could be self-explanatory. Or it could be
explained by some other fact.

If a moral fact is a brute fact, there is nothing to say about why

that moral state of a airs obtains. Obviously, in that case there will
be no theistic explanation, and ThEM will be false. The notion that
some moral facts are brute facts may seem perfectly respectable, so
long as they are necessary. A brute contingent moral fact, by
contrast, is anathema. It seems bizarre to think that a moral state of
a airs could obtain in one world, fail to obtain in another world, yet
there be no explanation for the di erence. While brute necessary
moral facts—one might hold, for example, that those in a position to
help ought to do something to relieve undeserved su ering is simply a
brute, necessary moral truth—are not anathema, they are on the
surd side. If we are committed to the existence of moral facts, then
we must accept brute moral facts, if no explanation of them is
possible. But that is the last resort. We should look for explanations
for these, and it is an open question whether the best such
explanations will be theistic explanations.

Here is another option. One might claim that of the moral facts
not explained via subsumption, some of them are self-explanatory.
Alexander Pruss writes that a fact is self-explanatory if
understanding the fact, and that it obtains, is su cient for
understanding why it obtains.9 As I mentioned above, I want the
notion of explanation here not to be observer-relative, and Pruss’s
de nition seems to make it so. But this is easily handled. We can say
that we are in a position to judge that some fact is self-explanatory
if we are in a position to know, upon understanding some fact and
that it obtains, why it obtains. Our understanding why something
obtains upon knowing that it obtains is our basic test for the self-
explanatory, though it is surely prone to err by failing to identify
some facts as self-explanatory that really are.

One might think it obvious that if some of these moral facts are
self-explanatory, then ThEM must be false: for if some of these
moral facts are self-explained, then they are not explained by God’s
existence, nature, or activity, and that contradicts ThEM. But there
is at least an initially plausible way of bringing ThEM and the thesis
that some moral facts are self-explanatory into consistency.
 Here’s a model for the consistency of ThEM and the self-
explanatory character of some moral facts. Suppose that there is a
single moral fact that explains all moral facts distinct from it.
Suppose, for example, that the fact that people are morally required
to obey God explains all other moral facts; for every other morally
required act-type, it has its status as such in virtue of being
commanded by God. Why are we morally required to refrain from
telling lies? Because we are morally required to obey God, and God
commanded us not to tell lies. And so forth.

One might plausibly make two claims about this moral fact. The
rst is that it is self-explanatory. If one grasps that the state of
a airs people’s being morally required to obey God obtains, then one is
in a position to see why it obtains. James Rachels writes:

To bear the title ‘God’ …a being must have certain quali cations.
He must, for example, be all-powerful and perfectly good in
addition to being perfectly wise. And in the same vein, to apply the
title ‘God’ to a being is to recognise him as one to be obeyed…. And
to recognise any being as God is to acknowledge that he has
unlimited authority, and an unlimited claim on one’s allegiance….
That God is not to be judged, challenged, de ed, or disobeyed, is at
bottom a truth of logic.10

Rachels holds both that people are morally required to obey God is
self-evident (its truth is immediately knowable a priori) and self-
explanatory (why it is true is a matter of logic).
 The second thing that one might claim about this moral fact is
that it is also a theistic fact. It is a fact about God: God is such that
people are morally required to obey him. This is not a mere rigged-
up contingent fact, like God’s being such that there are three paper
clips on Murphy’s desk, or even a rigged-up necessary fact, like
God’s being such that three is a prime number. This is a fact about
God, one might claim, a fact about how people are to relate to him.
And so, on this view, it is a fact about God’s existence and nature,
and thus quali es as a theistic fact.

If the view that I have described so far is coherent, then we have

a way to render ThEM at least consistent with the self-explanatory
character of some moral facts. Both of the claims made here are
hard to believe, though. It is hard to believe that the claim that God
is to be obeyed is genuinely self-explanatory. It seems that one can
grasp the state of a airs people’s being morally required to obey God
without thereby seeing why it obtains. For it is not clear why being
a perfect creator involves created rational beings’ being subject to a
moral requirement of obedience; this needs further explanation.11
Even if it were self-explanatory, it is not clear that it would count as
a theistic fact. A theistic fact, as I am using the term, is a fact that
involves God’s existing; if a fact is a theistic fact, then its obtaining
cannot be coherently asserted by an atheist. But this moral fact may
not involve God’s existing; it seems that the obtaining of this fact
could be coherently asserted by an atheist. That God is to be obeyed
entails that we are to render obedience to whatever being quali es
as God—to whatever being is the most perfect being possible. This
moral fact could hold even without God’s existing, and even though
God does exist, that does not make this moral fact God’s-existence-
involving. (Indeed, Rachels uses this alleged de dicto logical truth
about God to argue that there is no such being as God.)

As far as I can see, there is only one route of escape for one who
wants to claim that people are morally required to obey God is a
theistic fact. My argument has been that it does not follow from the
moral requirement’s telling us how to respond to God that it is a
theistic fact. But there are other ways that one might argue for this
being a theistic fact. If the property being morally required turns out
to be itself a theistic property, then the moral fact people are morally
required to obey God would be God’s-existence-involving. But on this
view the moral fact’s being God’s-existence-involving is not due to
its being a fact about how we should respond to God, as opposed to
how we should respond to each other, or ourselves, or dogs, or the
environment. (I will discuss this account of moral properties in
further detail below.) And even if it were to turn out that the best
account of the property being morally required is that this property is
a theistic property, that would do nothing to deal with the objection
that the moral fact in question is not self-explanatory.

THEISTIC EXPLANATIONS OF NON-SUBSUMED MORAL FACTS

Let us put to the side for the moment the possibility of defending
ThEM in terms of moral facts that are both theistic and self-
explanatory. There are various ways to divide up the sorts of theistic
explanations of moral facts that remain for consideration. One way
to divide them up is by the sort of theistic facts included in the
explanans: whether these theistic facts are facts about God’s
intrinsic nature, for example, or whether they are facts about God’s
free activity. Another way to distinguish them is by the sort of
explanatory relationship that holds between the theistic and moral
facts: whether the explanatory relationship is causal, for example, or
perhaps constitutive. Another way to divide them up is by whether
the explanatory relationship between theistic and moral facts is
immediate or mediated by other sorts of facts.

Quinn’s Account: Immediate Divine Causation

Here is one model of theistic explanation of morality that appeals to

divine action. This is perhaps the simplest theistic account. For
every moral fact, God causes that moral fact to obtain ex nihilo.
Philip Quinn describes and defends such a view, on which God’s
bringing it about that one is morally obligated to φ occurs in a
particularly robust way: the causal relationship between God’s will
and moral facts is total, exclusive, active, immediate, and
necessary.12

Quinn o ers an argument from divine sovereignty for this

view.13 The view of divine sovereignty that Quinn favors combines
divine aseity with the dependence of non-divine fact on God’s
agency. (For Quinn, a fact is non-divine if it neither involves nor
logically entails God’s existence.) Quinn thinks that from this view
ThEM follows—at least, ThEM follows if we exclude from the scope
of ThEM all moral facts that involve or logically entail God’s
existence.14 As we saw above, though, it is plausible that no non-
subsumed moral facts involve or entail God’s existence, at least not
in virtue of their being about how one ought to respond to God. If
the success of the argument from divine sovereignty does entail
ThEM, then it is clear that the conclusion to be drawn is not only
that ThEM is true, but that it is necessarily true: if divine
sovereignty obtains only if all non-divine-facts depend on the
exercise of God’s causal powers, and God is necessarily sovereign,
then ThEM is true necessarily.

It is important to note that Quinn’s view does not entail the

necessity or the contingency of non-subsumed moral facts. It might
be thought that since such moral facts are, on this view, the result of
God’s free creative choices, such moral facts must be contingent. But
it may be that God, though free, has a character such that God
necessarily causes non-subsumed moral facts to obtain.15 More
plausible is the view that, even if non-subsumed moral facts are
contingent, it is a necessary truth that if God creates human beings,
then God causes certain moral facts to obtain. God’s necessary love
for the beings that he creates may well ensure that God freely wills
the norms of morality to obtain that are the most loving.16 If one
doubts this, the likely source of the doubt would be that there is no
set of moral norms that counts as the most loving norms that God
could bring about.

What seems plausible about Quinn’s view is that God’s

sovereignty entails that all non-divine-facts are ultimately to be
explained by reference to the exercise of God’s causal power. But it
is hard to make sense out of Quinn’s view that moral facts are made
to obtain simply in virtue of an exercise of God’s causal power. The
relationship between moral facts and non-moral facts is, we tend to
think, a rational relationship: if a non-moral fact contributes to
making a moral fact obtain, it is in virtue of that non-moral fact’s
constituting the moral reason to perform a certain action, or in
virtue of that non-moral fact’s serving as an enabling condition for
that reason, or in virtue of that non-moral fact’s serving as a
defeater-defeater for that reason.17 But the explanatory relationship
that Quinn posits between a non-moral fact (God’s causal activity)
and any moral fact is not a rational relationship; it is a merely
causal relationship.

Perhaps those who would explain morality theistically would

have reason to accept the strangeness of causation of moral facts ex
nihilo if the doctrines of divine aseity and of dependence of the non-
divine on the divine required it. But these doctrines do not require
this view. For even if God’s sovereignty entails that all non-divine
facts are ultimately to be explained by reference to the exercise of
God’s causal power, Quinn’s argument does not successfully show
that the exercise of divine causal power is the immediate cause of the
obtaining of non-subsumed moral facts.

Here is an analogy. Consider the question of the relationship

between God’s sovereignty and the physical fact that salt dissolves
in water. There are some philosophers—‘occasionalists’—who hold
that it would be an a ront to the divine sovereignty to hold that the
fact that salt dissolves in water has any explanation in addition to
that of the divine will’s causing the dissolving.18 Indeed, those
defending this view would deny that there was any explanation of a
particular case of salt’s dissolving in water other than the divine
will’s causing that dissolving. But this view is a distinctly minority
position, because to most philosophers it seems plain that it would
not be contrary to divine sovereignty for God to operate in nature
through secondary causes. It seems to be perfectly in keeping with
the dependence of all creation on God that God could create a world
in which there are beings that have, themselves, causal powers, and
to which we can appeal in explaining physical facts. That salt
dissolves in water is to be explained (in part) by the water’s causal
powers. This is not an a ront to the divine sovereignty, so long as
the water’s having the causal powers that it has is somehow
explained in theistic terms.19

Now, one might make the same case on behalf of theistic

explanations of morality. One might hold, pace Quinn’s speci c
account but nonetheless relying on the doctrine of divine
sovereignty, that the divine will can explain morality while not
immediately causing moral facts to obtain. The model for this would
be: moral facts are explained by some other facts; and these other
facts are themselves theistically explained.

Natural law theory: mediated theistic explanation

Here is an example of this sort of strategy. Consider the classic

formulation of natural law ethics. According to natural law ethics,
the fundamental norms of morality are necessarily true; one could
not be a human being and fail to be bound by these norms.20 The
status of these norms as binding and as exhibiting their particular
content is explained most immediately by reference to human
nature. Human nature might be understood here in an Aristotelian
way—as the set of distinctive potentialities the possession of which
constitutes one’s humanity21—or in a more Hobbesian way—as the
set of innate drives that are characteristic of human beings.22 In
either case, the norms of morality are as they are due to facts about
human nature. In asking why lying is wrong, it would be a mistake
to bypass facts about human nature and go directly to God. For, on
the theistic formulations of these views, it is by God’s making
human beings with the particular natures that they have that
humans are morally required to act as they are.23

More than one detail of this account remains vague. What

exactly is the relationship supposed to be between facts about
human nature and moral facts? Surely it is not merely causal. The
most promising idea, it seems to me, is that moral facts can be
informatively identi ed with, that is, reduced to, certain facts about
human nature. One might say, for example, that being morally
required just is being backed by reasons for (human) action (of a
certain kind), and being a reason for (human) action just is being
humanly good, and being humanly good just is being related (in a
certain speci c) way to human nature (to the actualization of a human
potentiality, as in the Aristotelian version, or to the satisfaction of
an essential human desire, as in the Hobbesian version).

More important for our purposes is the relationship between

facts about God and facts about human nature. If we take it that
theistic facts are not to be invoked to explain the explanatory
connection between the facts about human nature and moral facts,
then there would be only one place to locate theistic explanation
within a natural law view: in the existence of the nature, human,
that entails the relevant moral facts.

How might God be responsible for the existence of the kind

human, so that facts about God (God’s making it the case that there
is such a thing as human) can explain moral facts (e.g. humans are
morally required not to lie) by way of the entailment from facts about
human nature to moral facts? There are a variety of ways, some of
them more interesting, some of them less interesting. If one holds
the view that a kind cannot exist unless some members of that kind
exist, then one can hold that it is simply God’s creative act of
bringing some humans into existence that explains why the kind
human exists and exhibits God’s role in explaining moral facts. How
does God’s activity explain why humans are morally required not to
lie? Because being human explains being morally required not to lie,
and without God’s activity, there would be no humans, and so no
such thing as being human. (Note well: this view is not the mere
view that without God there would be no humans, and so no beings
to which the moral norm humans are morally required not to lie can
apply. This latter account is consistent with its being a moral fact
that humans are morally required not to lie, even without there being
any humans, and so God’s role in creating humans would not be
su cient to support ThEM.)

This would do the job of vindicating ThEM, and of vindicating

ThEM as a necessary truth, so long as we add that it is a necessary
truth about any possible set of moral facts that it is explained by the
kind to which the beings to whom those norms apply belong and
that it is a necessary truth that no creaturely kind exists unless God
brings about instances of that kind. But one can defend ThEM in
terms of God’s role in xing kinds without holding that a creaturely
kind does not exist unless God brings about instances of that kind.
One might claim that a kind can exist without there being any
instances of it, but only insofar as that kind is xed by a divine idea
—a particular creative possibility represented by the divine mind.
Human is a certain creative possibility; there are no doubt other
such possibilities involving created rational material beings. But if
the kind is ontologically dependent on the divine mind, ThEM is still
vindicated. For this view of the relationship between creaturely
kinds and the divine ideas would su ce, in combination with the
general natural law idea, to entail that all moral facts must be
explained by facts about God.

But if either of these formulations of the natural law view is

accepted, one might complain that while the letter of ThEM has
been preserved, its spirit has been violated. One might have hoped
that the way that facts about God explain morality has something
distinctive about it, or that the connection between God’s nature or
activity and moral facts would be direct. Instead, on this particular
formulation of the natural law account, facts about God explain the
fact that humans are morally required not to lie in precisely the
same way that facts about God explain the fact that water dissolves
salt. God’s role in the explanation goes only to accounting for the
existence of the kinds (human, water, salt) in question.

There are ways to make God’s role in explaining moral facts

stronger without going so far as the normative occasionalism to
which Quinn commits himself. One might argue, for example, that
what is distinctive about the kind human (in comparison to the kind
salt and the kind water)—its ability to ground explanations of moral
facts—is due to some additional theistic explanation, not just some
brute or self-explanatory facts about that nature. Recall that on the
Aristotelian and Hobbesian varieties of natural law theory it is the
role of proper actualization of a set of potentialities and necessary
human desires (respectively) that x the moral facts to be explained.
One might claim that the status of some states of a airs as the
proper actualization of beings of a certain type or the state of some
states of a airs as the ful llments of necessary human desires
requires or permits distinctive theistic explanation.

Take the Aristotelian view rst. One might claim that just as we
appeal to the intentions of a minded being to explain what counts as
the successful functioning of an artifact, we need to appeal to the
intentions of a minded being to explain what counts as the proper
actualization of the human being.

By contrast, on the Hobbesian view: one could say that what lls
the role that God’s ‘maker’s intentions’ have in the Aristotelian view
is the presence of desire in the human—so that there is a gap
between the way humans are, and the way that they must be if they
are to be ful lled. What’s more, some of these desires are necessarily
possessed by humans, and central to their motivational structures.
The special explanation of the moral in the Hobbesian scheme is to
be located in God’s making some creatures that have appetites,
indeed appetites that are characteristic of members of that kind.24

The dispute between the sort of normative occasionalism

defended by Quinn and the natural law view comes down to a
dispute on whether ThEM should be defended by treating theistic
facts as an unmediated or mediated explanans vis-à-vis moral facts.
Quinn treats the explanatory relationship as unmediated causation,
and understands the theistic facts that immediately cause moral
facts as facts of the form ‘God wills that A be morally required to
φ’.25 There are, however, other views on which the relationship
between theistic and moral facts is unmediated, though the
relationship is not a causal one.

Adams’s Account: Unmediated Constitutive Explanation by

Appeal to Divine Command

Consider, for example, the view defended by Robert M. Adams. On

Adams’s view, non-subsumed moral facts of the form ‘any x who is P
is morally obligated to φ’ are explained by theistic facts of the form
‘God commands all those who are P to φ’. The relationship between
God’s commanding those who are P to φ and its being morally obligatory
for those who are P to φ is not a causal relationship, but a
constitutive one: to be morally obligatory just is to be commanded
by God. So, for each moral fact that one is morally required to φ, it
can be explained by the fact that God commands one to φ and being
morally required to φ just is being commanded by God to φ. This is a
constitutive explanation of the moral fact.

Adams’s reductive account would justify ThEM, and as a

necessary truth: every moral fact must be theistically explained.26
One might dispute this, claiming that on Adams’s account, the moral
fact That being morally required to ϕ just is being commanded by God
to φ is not theistically explained, and so ThEM turns out false. But
this claim is mistaken. That being morally required to p just is being
commanded by God to φ is not a moral fact, and so need not be
theistically explained for ThEM to be true. It does not count as a
moral fact because it does not a rm that any action has any moral
status; one can be a complete moral nihilist, denying that any moral
requirements exist, while also a rming that being morally obligated
to φ just is being commanded by God to φ.

Adams’s argument for the reduction of the morally obligatory to

the divinely commanded is that the role of the obligatory—that is,
the set of commonplaces about the obligatory that are recognized by
competent users of that concept—is best lled by divine commands.
By conceptual analysis alone we can know only that wrongness is a
property of actions (and perhaps intentions and attitudes); that
people are generally opposed to what they regard as wrong; that
wrongness is a reason, perhaps a conclusive reason, for opposing an
act; and that there are certain acts (for example, torture for fun) that
are wrong. But given traditional theistic beliefs, the best candidate
property to ll the role set by the concept of wrongness is that of
being contrary to (a loving) God’s commands. For that property is an
objective property of actions. Further, given orthodox theistic views
about the content of God’s commands, this identi cation ts well
with widespread pre-theoretical intuitions about wrongness; and
given Christian views about human receptivity to divine
communication and God’s willingness to communicate both
naturally and supernaturally, God’s commands have a causal role in
our acquisition of moral knowledge.27 Adams also claims that it is
part of the meaning of ‘obligation’ that obligations are social in
character28 and involve actually made demands by one party in the
social relationship on another.29 It is the fact that a demand is
actually made that gives sense to the notion that one has to perform
an action, rather than merely that it would be good, even the best,
to do it.30

On Adams’s view, as on the view of his fellow divine command

theorist, John Hare, non-subsumed moral requirements can be
contingent: even though the content of the divinely commanded is
shaped by its being issued by a perfectly loving being, love does not
yield a determinate set of commands, and so God’s commands might
well have been di erent.31 But, again, their view on the explanation
of moral requirements is consistent with the view that moral
requirements are necessary, if one takes the divine nature to x the
commands that God gives to human beings.

There is a bit of a dilemma in the vicinity. It does seem on its

face implausible that God’s commands are entirely xed, so that it is
a necessary truth that, for each command God could give, given all
of the facts about the natures of human beings and the
environments in which we live, God must give precisely that
command. Even if there are certain ways that the world could be
such that, given that world, God can give only the precise set of
commands that God gives, surely there is some looseness between
God’s commands and other facts about the world such that it is
possible for God to give somewhat di erent commands while the
world is otherwise the same. But it seems to me that if we
acknowledge some divine freedom in commanding, we open the
door to a powerful argument against the view that being morally
required is identical with being commanded by God. Put the point this
way. For every fact that it is morally required to φ, it seems deeply
implausible that every fact not identical with (or including) its being
morally required to φ could be the same yet it not be morally
required to φ. For some fact of the form God’s commanding φ-ing, it
is extremely plausible that every fact not identical with (or
including) that fact of God’s commanding φ-ing could be the same
yet God not command φ-ing. But, if so, then there is a tension in the
thesis that its being morally required to φ is identical with God’s
commanding φ-ing: it is very plausible that its being morally
required to φ supervenes on the set of facts each of which is neither
identical with nor includes it; it is very plausible that there are some
cases of God’s commanding φ-ing that do not supervene on the set of
facts each of which is neither identical with nor includes it. Of
course, reductions often bring with them tensions of this sort; what
needs to be asked is whether the pressure for reduction, and the
other merits of the reduction, are su cient to justify resolving the
tension by rejecting our prior views either of the supervenience of
the moral on the non-moral or of God’s freedom in commanding.32

Zagzebski’s Account: Unmediated Constitutive Explanation by

Appeal to Divine Motivation

Consider, by contrast, the account o ered by Linda Zagzebski. On

her view, for an action to be morally required in given
circumstances is for it to be an action the failure to perform which is
contrary to a virtuous character; it is to be an action that
necessarily, in those circumstances, the virtuous person performs.33
But on Zagzebski’s view, the standard for virtue is set by God; it is
God’s motivations that determine what count as good human
motivations, and thus what count as human virtues.34 The
determination relation is not causal but constitutive: for a
motivation to be good is for it to be the motivation that God would
have in like circumstances. (This may strike one as a strange view,
given the di erences between God and human beings; Zagzebski
points out that at least Christians may not be in a position to make
this objection, since their view that God became human in Christ
commits them to the view that in some cases we know fully well
what God would do in certain circumstances of human life.35)

Why think this? Like Adams’s view, Zagzebski’s begins with a set
of claims that is not distinctively theistic: that in the construction of
ethical theories we should begin not with concept-analysis but with
cases, and in particular with exemplars of good persons.36 Starting
with clear cases of good persons, we can initially de ne good
outcomes, acts, traits, and motivations in terms of what good
persons would seek, want, and do. Such starting points are fully
compatible with a philosophical inquiry into what makes someone a
good person, just as starting our inquiry into tigers may rightly
begin with paradigm cases of tigers and starting our inquiry into
water may rightly begin with paradigm cases of water. But once we
take these starting points and situate them within a framework of
theistic belief, it is clear that our paradigm of good personhood is
chosen, so to speak, for us; we cannot but think of God as an
exemplar of a good person (unless one wants to deny God’s
personhood altogether), and we cannot think of God as simply one
good person among others. God will have to be at the center of the
view, and the best theory that places God at the center of the view,
Zagzebski thinks, will make God’s motivations central and de ning
of the good, and thus of moral requirement. All this turns out to
generate non-subsumed moral requirements that are necessary,
given (a) the necessity of the divine motivations, (b) the
characterization of good human motivation in terms of divine
motivation, and (c) the de nition of obligation as what must be
done by one who is properly motivated.

Zagzebski’s account is, however, hard to accept. There are

reasons to worry about the particular way that she appeals to
exemplars to get the ball rolling on explaining moral properties; it
seems to me that the sort of infallibility that she claims for our
designation of certain persons as paradigms of goodness is simply
incredible. She mentions Socrates, Buddha, and Jesus as exemplars,
and she writes of exemplars of a natural kind N that ‘Direct
reference ensures that our semantic community cannot be radically
mistaken about the members of the class of N’;37 ‘we may be
mistaken about some of them, but we cannot be mistaken about
very many’.38 Surely this cannot be right. It is epistemically
possible, for example, that the Buddha actually had a hidden life
that was sordid, that his motives were self-serving, that he was not a
particularly admirable person. It is epistemically possible that
Socrates was a jerk.39 To put the point as uncontroversially as
possible: if moral goodness is about motives, and even our own
motives are very much hidden from us,40 it hardly seems possible
that we could claim for ourselves anything like incorrigibility about
our exemplars. Perhaps we should distinguish real Buddha and real
Socrates from idealized abstractions based upon them: our
exemplars are Apparent-Buddha and Apparent-Socrates, and we thus
avoid messy worries about the actual details of their lives. But not
only does it seem obvious that we want to make this move because
we have some criteria, implicit or explicit, to which the contrasting
Real-Buddha and Real-Socrates might fail to measure up, but this
move would undercut Zagzebski’s idea that we learn about good
motivation by observation of these exemplars.41

Put this to the side. More di cult to grasp is Zagzebski’s account

of the goodness of objects and states of a airs in terms of God’s
motivations. One might have some sympathy for the view that God’s
motivations x the goodness of beautiful mountain ranges, the
relieving of undeserved su ering, or the writing of crisp
philosophical prose. What seems deeply objectionable, though, is
that on Zagzebski’s view even evaluative properties of God are xed
in this way: ‘What about God’s own thick properties? What does it
mean to say that God is lovable if that does not mean deserving of
love or having a property in advance that makes him lovable? It
means that God loves himself, and his lovability is a constituent of
his loving himself.’42 The di culty with this view is that it seems
clear that whatever makes it the case that God’s motivations are
privileged with respect to the xing of value is logically prior to
those motivations xing value; and surely what makes God thus
privileged is to be understood in evaluative terms. Zagzebski’s view
requires us to give up one or the other of these two immensely
plausible claims. And if we do end up saying that God’s own
goodness is not to be characterized in terms of divine motivations,
then we leave open the door for explanations of moral facts that
appeal directly to facts about God’s goodness that do not involve
divine motivations.

DIYINE AUTHORITY

I have been concerned thus far to see what sorts of strategies are
promising ways to defend ThEM. The strategies that I have been
discussing are global strategies, strategies to show that all non-
subsumed moral facts possess a common theistic explanation. But
there are ways to try to support ThEM, or to approach it, without
employing a global strategy. One might try to defend the full or
approximate truth of ThEM by appeal to speci c moral facts, the
application of which invariably requires a theistic fact. The most
common such strategy is an appeal to God’s status as legitimate
commander—we humans are under a moral requirement to obey
God, and so if God commands us to perform some action, then we
are morally required to perform that action.43

I will not here undertake an investigation into the nature of

divine commands, and how they can be given,44 but I do want to
consider the status of the putative moral fact that people are
morally required to obey God. To take this to be a moral fact is to
a rm the divine authority principle (DA)45—people are under a
moral requirement to take God’s commands as their guide to action.
We are familiar with authority relationships outside a theistic
content. It is commonly thought that parents are authorities over
their children in this way: children are morally required to obey
their parents, so that if I tell my son to help me to empty the
dishwasher, then my son is morally required to empty the
dishwasher, and the explanation includes the fact that I told him to.
The di erence between parental and divine authority is, we might
say, in the scope (God’s authority is supposed to extend to all
created rational beings, and to a wider range of actions) and
strength (the reasons for action given by God’s commands are
supposed to be weightier), but the formal features of the authority
relationship are the same.

DA, the principle that God has rightful authority over us, can be
employed in a number of ways—some more ambitious, some less
ambitious—in one’s overall account of how moral facts are to be
explained. One might treat the fact that DA is true as a supreme and
architectonic moral principle. On this view, which one might call
normative divine command theory, that DA is the case is the sole non-
subsumed moral fact, and so serves as that moral fact that explains
every subsumed moral fact. For example, if it is true that one is
morally required not to lie, the explanation is that people are
morally required to obey God, and God has told us not to lie. But
this is not the only way that one might appeal to DA. One might
treat its obtaining as a supreme moral fact without holding that it is
the sole non-subsumed moral fact. On this view, while there are
moral facts that are not subsumed under the fact that DA obtains,
DA is nevertheless superior to the other principles, either by
qualifying them (for example, every distinct moral principle carries
the rider ‘unless God gives a command incompatible with this’) or
by being lexically ordered over them (that is, one’s primary duty is
to obey God; only once this is accomplished can one move on to the
ful llment of other moral principles). On this view, the principle
that we are morally required to obey God controls the application of
all other moral principles (and is not itself controlled by any of
them), but does not ground the validity of those principles. On yet a
third view, DA is not supreme; it is one principle among others.
(One might nd implausible this view, thinking that it would
suggest that God could give a command that it would be permissible
to disobey; and one might balk at that suggestion. But the allegedly
untoward consequence does not at all follow: it may be that God
necessarily would not give a command that requires violation of
other moral principles.) On one version of this third view, DA is a
moral principle that is not subsumed under a more fundamental
moral principle or principles; on another version, it is so
subsumed.46

There are two obvious points to be made here. The rst is that,
as I indicated earlier, even the strongest version of DA, on which DA
is the sole non-subsumed moral principle, would not of itself
establish ThEM. For it could be that DA lacks a theistic explanation
—it could be that we do not need to resort to any theistic facts—
facts about God’s existence, nature, or activity—to explain why DA
is true. The second is that the extent to which DA can explain the
range of moral facts depends on where DA is placed within the
system of moral facts. So when we turn, as we now will turn, to the
question of whether we should hold that DA is true, we should also
ask what these arguments for DA’s truth would tell us about
whether (1) DA is supreme and uniquely non-subsumed, or (2) it is
supreme and non-subsumed, but not uniquely so, or (3) it is not
supreme and non-subsumed, or (4) it is itself subsumed.

I am going to argue that, at present, we lack any good

philosophical argument for the truth of DA, whether as a subsumed
or non-subsumed principle. I will then ask what implications the
failure of extant arguments for DA has for the objective of
explaining some moral facts by appeal to DA.

There are three bases from which arguments for God’s authority
have been launched. One of these bases we have already
encountered: a general account of how theistic facts are
explanatorily related to moral facts. On such an account, one aims
to show that given God’s explanatory relationship to the moral
order, it follows that God has authority over created rational beings.
A second of these bases is an appeal to the divine nature. On such
an account, one aims to show that given God’s perfection, it follows
that God is authoritative. A third is an appeal to widely held moral
principles by which DA is subsumed. On such an account, one
argues that given some correct moral principle or principles, alone
or in conjunction with facts that hold across all possible worlds in
which there are created rational beings, DA follows. But in my view
no extant argument tting within these schemas establishes the
correctness of DA.47

Arguments from Morality’s Dependence on God to Divine

Authority

One might suppose that the a rmation of one of the views of God’s
relationship to morality discussed in the previous section would
surely be su cient to establish God’s authority. If morality as a
whole is dependent on God, one might think, then surely a fortiori
God has authority over created rational beings. But this is a mistake.

Quinn’s causation view does not entail DA, for example. Even if
one holds that every moral state of a airs that obtains is caused to
obtain by a free act of God’s, that would not show that humans are
morally required to obey God. God might, for example, have willed
only the moral requirements humans are morally required not to
coerce one another and humans are morally required not to defraud one
another to obtain. If so, then there would have been no moral
requirement to obey God; the fact that God tells someone to do
something would not make any action morally required. Only if God
were also to will humans are morally required to obey God would
obedience be a required act. The same holds true of the natural law
theories that place mediating natural facts as o ering explanations
of moral facts: that God is ultimately responsible for the moral
requirements to which created rational beings are subject does not
entail that those created rational beings are required to obey God.

More promising in the defense of DA would be a view like

Adams’s, on which the property being morally required just is the
property being commanded by God. On this view, it follows that
whatever is commanded by God is morally obligatory. But even
putting to the side earlier objections to this view, we seem to fall
short of a defense of a moral requirement of obedience. Success in
performing all those actions that God has commanded one to
perform does not require believing in God’s existence. I might do
what you tell me to do, and non-accidentally so, without believing
that you exist or that you have commanded me to do anything. (My
father may have told me always to nish my supper, and I may have
long forgotten that he told me this; nevertheless, I might always
nish my supper, and I might always nish it because my father
gave me that command.) Indeed, this possibility is crucial for
Adams, who wants to say that atheists can act well, and non-
accidentally so, even while denying the existence of the facts that
are strictly identical with facts concerning moral obligation.48 But
obedience to someone requires, at least if it is intelligible, belief in
the existence of the person whom one is obeying. Obeying you (as
opposed to simply acting consistently with your commands) requires
my being guided by your commands as your commands, and so the
mere identi cation of the morally obligatory with the divinely
commanded is not enough to establish DA. One would have to
identify its being morally obligatory to φ with φ-ing’s being constitutive
of obedience to one of God’s commands, an identi cation that Adams
declines to make.

Arguments from the Divine Nature to Divine Authority

An alternative to these arguments for divine authority is a direct

appeal to the divine nature itself. Here is the idea. As we noted
above, God is to be conceived as perfect. One might argue that it
follows from God’s status as a perfect being that God possesses
authority over all created rational beings. The idea is plausible
enough, but the di culty is in making out the claim that God’s
perfection entails a moral requirement of obedience. None of the
traditional divine perfections, either alone or in combination,
provide the material for a successful argument that God is
authoritative.

Consider omniscience. One might argue that if God is omniscient,

then God knows all moral truths; and knowledge of these moral
truths will be re ected in God’s commands. We are morally
required, then, to do what God tells us to do. But this argument
fails. Even if God knows all moral truths and thus commands us to
do only what is morally required, God’s omniscience does not give
us any reason to think that God’s commands are themselves reasons
to obey God; God’s commands are indicators of what actions are
independently morally required, actions whose status as moral
requirements may be entirely independent of God’s say-so.

Perfect moral goodness fails as a basis for an account of divine

authority for similar reasons. Even if God is perfectly morally good,
that does not give us any basis to suppose that God’s commands give
new reasons for action, only that God’s commands are themselves
morally good. Think of analogies with knowledgeable or morally
good human beings. We want to imitate them, and to ask them for
advice; but we don’t take their knowledge or moral goodness to
place us under an obligation to obey them.

Neither does omnipotence provide an argument for divine

authority. Put to the side the fact that God’s omnipotence ensures
that God can threaten us with punishments for non-compliance, or
rewards for compliance; such will not give us moral requirements to
do God’s will. On re ection, omnipotence is not a promising source
of divine authority: for omnipotence is about what God can cause,
and as we have seen, a merely causal account of God’s relationship
to morality is insu cient for divine authority.49 There is no good
argument from the traditional divine perfections to divine
authority.50
Arguments from Other Moral Principles to Divine Authority

The arguments to DA from the general explanatory accounts of

moral facts by theistic facts and from the divine perfection fail. One
might, however, turn to arguments that hold that DA is an
implication of some other moral requirement. Even if this strategy
were fully successful, one would not be able to hold that DA is the
sole non-subsumed moral requirement, or even that it is a non-
subsumed moral requirement, for use of this sort of argument for DA
entails that DA is a subsumed moral fact. But even if it is a
subsumed moral fact, DA might be useful in explaining a variety of
moral facts that might otherwise go unexplained. (If one believes,
for example, that one is morally required to observe certain
religious rituals, rituals that have their status as binding only from
God’s say-so, an appeal to DA, however DA is explained, might be
very useful indeed.)

It turns out to be remarkably di cult, though, to nd a moral

principle with the relevant implications. Clearly, one cannot appeal
to any authority-generating moral principle whose application
conditions require a voluntary act on the part of the one subject to
authority (for example, that one ought to keep one’s promises), for
such a principle cannot yield the universality of DA (that is, that all
people are morally required to obey God). The most promising
candidate principles are those that might be held to apply simply in
virtue of God’s creating and conserving rational beings; for, even if
God creates contingently, the moral principles in question only
govern possible worlds in which God creates such beings.

Some have held, for example, that created rational beings are
God’s handiwork, and so are divine property. Of the number of
di culties for this position, the preeminent is that, of the plausible
rationales that we have for holding that being someone’s handiwork
establishes it the maker’s property, none of them goes through in
cases in which treating the handiwork as the maker’s property
necessarily involves an interference with another rational being’s
autonomy. The best case for holding that your making something
gives you rightful title to it and its use is that if in making that thing
you make others worse o in no way, including their freedom to act
at their discretion, it would be an unjust deprivation of your
freedom to be unable to make something for your own use and
enjoyment. But claiming rightful title over a human being itself
constitutes a deprivation of that human’s freedom to act at his or
her discretion. So on the best account of why making generates
rightful title, God’s making humans fails to explain God’s rightful
title over us.

Some have held, alternatively, that created rational beings are

under a debt of gratitude to God for being made and conserved in
being, and so one ought to repay this debt by yielding one’s
discretion over how to act to God. While some have argued that
there is no such debt of gratitude to God, or that debts of gratitude
never generate speci c requirements of action, it seems that the
bigger obstacle for this view is that even if the debt of gratitude
owed to God must be repaid by yielding one’s practical freedom to
God, that does not entail that one is under divine authority; it
entails at most that one ought to take steps to place oneself under
divine authority. If I were presently under a massive debt of
gratitude to you, it may well follow that I ought to give you many
good things, things to which I presently have a right. But you do not
come to have a right to them simply in virtue of my debt of
gratitude. At most, my failure to give them to you shows me up as
an ingrate; my failure does not itself trigger a transfer of ownership.
Similarly, even if I owe an incalculable debt of gratitude to God, at
most this would show that my failure to submit myself to God’s
authority marks me as an abominable ingrate; it does not of itself
make God authoritative over me.

Should we conclude that DA is false, and so cannot be used to

explain any other moral facts? Perhaps, but perhaps not. All these
arguments for divine authority are top-down: they try to move from
principles that are either more metaphysically or morally
fundamental than DA to DA as an implication. One might, by
contrast, take a more bottom-up approach. This would take DA to be
a knowable truth, though not one as yet explained, whether through
itself or by way of more fundamental metaphysical or moral
principles.

One might, for example, just posit divine authority as a divine

perfection. But this faces the immediate obstacle that authority
requires someone over whom to be authoritative, and the existence
of created rational beings is a contingent matter, a matter of God’s
discretion; but God possesses God’s perfections necessarily. Even if
one were to rig the perfection so that it is necessarily exhibited by
God—God is authoritative over whatever rational beings happen to exist
—so doing raises the question of such merely counterfactual
perfections. One would think that such merely counterfactual
perfections would be grounded in categorical ones. But, as we have
seen, there are no successful arguments from categorical divine
perfections to divine authority.

More promising would be the acceptance of the moral principle

people are morally required to obey God without proof from more
fundamental moral principles, or indeed by anything else. The
philosophical task would be simply that of de ning precisely the
God/created rational being relationship, and de ning precisely the
scope of the authority that God is supposed to have, all the while
guided (one might suppose) by the more stable considered
judgments about the rightness of obedience to God’s commands and
about how the principle requiring us to obey God’s commands
relates to other moral principles. On this view, our belief in DA
stands justi ed, if it does, by way of its being a clearly formulated
principle that remains once we have reached re ective
equilibrium.51 To defend DA in this way requires us to take no stand
on how DA is to be explained, whether DA is brute or self-
explanatory or otherwise explained, and indeed, if we have nothing
but this re ective equilibrium type argument for DA, we will be left
with no adequate account of why DA is true, or whether there is an
answer as to why DA is true.

This is less than we would hope for. But we should note that
other attempts to provide top-down philosophical accounts of
common authority relationships—in particular, the parent/child and
state/citizen—have su ered failures similar to that with which I
have charged the top-down arguments for divine authority.52 It may
well be that we are forced either reluctantly to accept the limited
usefulness of these bottom-up accounts of authority or to reject, in a
disturbingly wholesale way, the presence of these seemingly
paradigmatic authority relationships.

Even if we are willing to accept DA on these terms, we have to

concede that there is an explanatory gap that remains when we
appeal to DA to explain other moral facts. Even if, for example, one
thinks that such norms as one is morally required to love one’s enemies
and pray for those who hate one is a moral fact that can be explained
only by way of God’s command, the lack of an explanation for the
truth of DA would leave any moral fact explained by DA
inadequately accounted for. We would have to allow that we do not
quite know why we must love our enemies and pray for those who
hate us.

NOTES
 I am grateful to Mike Rea, Tom Flint, Trenton Merricks, Alex
Pruss, Errol Pierre-Louis, Ryan Lupton, and Chuck Mackel for
their comments on earlier drafts of this chapter.

1. I will not consider here the question of the relative merits of

theistic and non-theistic explanations of morality.

2. This is no doubt stipulative and slides over important issues

about the scope of morality’s application to non-human beings
(e.g. God, angels, Martians), the formal features that
distinguish moral from other sorts of requirements, and the
existence of non-deontic moral facts (that is, facts about moral
goods rather than moral requirements), but it will serve our
purposes to have a relatively clear and limited explanandum.

3. See, for a helpful discussion of this point in the context of

Thomas Hobbes’s theistic explanation of morality, Warrender
1957: 14–15.

4. A non-moral analogy: consider the event of the catching re of

a piece of paper. The distinction between the causal role of the
lit match’s being in contact with the paper and the causal role
of the paper’s not being damp corresponds to the di erence
between grounds of a moral requirement (what makes an act
morally required) and its validating conditions (the
circumstances in which those grounds are able to make that
act morally required).
 5. This is Kant’s (1998) task in Religion within the Boundaries of
Mere Reason; his account of the need for happiness to be
proportioned to virtue, and of the necessity of theistic facts to
make this possible, may also be found in the Critique of
Practical Reason (1996b), 5. 114–32. These Kantian themes
have been taken up both exegetically and in the service of a
theistic account of ethics in the work of John Hare: see e.g.
Hare 1996.

6. For a brief discussion of this model, see Murphy 2002b.

Something like this account of explanation of moral norms has
gured prominently in Mark Schroeder’s (2005, 2007) work
(on the explanation of reasons for action generally) in what he
calls the ‘Standard Model’.

7. See the ‘Divine Authority’ section below.

8. I am not appealing here to some general (and almost certainly)

dubious claim about the transitivity of explanation in making
this point. My point is speci c to explanation via subsumption:
since subsumption works as a mode of explanation by noting
that some action can be an instance of a broader action type, if
some fact explains the moral requirement to perform the
broader action type, then that fact explains the moral
requirement to perform the instances of it.

9. Pruss 2006: 123.

10. Rachel’s 1981: 43–4, emphasis in original.

11. Recall that on the conception of ‘God’ with which we are
working, ‘God’ is simply to be understood as the most perfect
being possible. Rachels’s argument requires us to add to this
that God has authority; it seems to me that it is not obvious
that a perfect being has authority, and indeed we have reason
to doubt it. See the ‘Divine Authority’ section below.

12. Quinn1999: 54–5.

13. Quinn1990, 1999: 63–5.

14. Quinn 1990: 293–5.

15. See Quinn 1999: 69–71, for thoughts in this direction.

16. For worries about this strategy, see Chandler 1985.

17. The fact of my promise can (partially) constitute a reason to

pay you on Tuesday for a hamburger today; my being able to
pay you on Tuesday can enable that reason to obtain; my
having threatened you with a beating unless you release me
from my promise can undercut the typical defeat of my reason
that your releasing me from my promise would have.

18. See, for illuminating discussions of the pros and cons of

occasionalism, Freddoso 1988 and Quinn 1988.

19. For accounts of God’s role in the dissolving of the salt by the
water alternative to that o ered by the occasionalist, see
Freddoso 1994.

20. For an overview, see Murphy 2002 c.

21. See Aristotle (1985), Nicomachean Ethics, 1097b22–1098a22.

22. See Hobbes (1994), Leviathan, ch. vi para. 4.

23. I o er as an example of this Aristotelian approach an account

on which it is God’s relationship to the agent’s nature that
gives God a role in explaining moral facts. But one might have
a patient-centered view, on which it is God’s relationship to the
natures of the entities to which the moral agent is responding
that gives God that role. Or one could hold a combination of
these views, on which morality is explained both by the nature
of the agent and the nature of the entities to which he or she is
responding; God’s role in the explanation could involve both
the natures of the agent responding and the entity to which
the agent responds.

24. Talk of natural desires and natural law in Hobbes makes

Hobbes sound a lot more like Aquinas and the predecessor
natural law theorists than is customary. This isn’t my fault; it’s
Hobbes’s. He’s the one who talks about natural desires and
characterizes his own moral theory as a theory of the laws of
nature, precepts that in themselves are rational but viewed in
light of their giver are divine laws. See Leviathan, ch. xv para.
41.

25. At di erent points in his career Quinn held di erent views on

this nature of the divine act that explains moral requirements.
He originally held that the divine act that brings about the
moral fact A is morally required to φ is God’s commanding A to
 φ (Quinn 1979); later (Quinn 1990) he held that it is God’s
willing that it be morally obligatory for A to φ; later (Quinn
1999), he held that it is God’s intending (antecedently) that A
φ that causes the resultant obligation. (For the idea of
antecedent intention of which Quinn was making use, see
Murphy 1998.) But Quinn nevertheless continued to hold that
the ‘bringing about’ relation between the theistic fact and the
moral fact is an unmediated causal relationship. In my view
this was a mistake: the view that the theistic explanans is
God’s antecedently intending that A φ ts better with a
reductive account, on which an action’s being morally
required is identi ed with its being intended (antecedently) by
God. For an account of the di culties with Quinn’s intention
account, see Murphy 2002b.

26. Adams (1999: 281–2) does hold, though, that were there no
God, or if God were not good, then something else (for
example, the hypothetical intentions of an ideal observer)
might t the semantically indicated role of the morally
obligatory.

27. Ibid.257.

28. Ibid.233.

29. Ibid.245–6.

30. Ibid.232.

31. Ibid.255–6; see also Hare 2001: 66–75.

32. For a more complete version of this argument, see Murphy
2002a: 82–92. For responses, see Almeida 2004 and
Wainwright 2005: 87–9.

33. Zagzebski 2004: 159–60.

34. Ibid.129–220.

35. Ibid.233, 237.

36. Ibid.40–50.

37. Ibid.51.

38. Ibid.41.

39. I am persuaded by some remarks of Peter Geach’s (1969: 118–

19) that, at the very least, Socrates as portrayed in the
Euthyphro is a not particularly admirable person.

40. As Kant (1996a) writes, ‘In fact, it is absolutely impossible by

means of experience to make out with complete certainty a
single case in which the maxim of an action otherwise in
conformity with duty rested simply on moral grounds and on
the representation of one’s duty....[ W]e like to atter
ourselves by falsely attributing to ourselves a nobler motive
[than self-love], whereas in fact we can never, even by the
most strenuous self-examination, get entirely behind our
covert incentives, since, when moral worth is at issue, what
counts is not actions, which one sees, but those inner
principles of actions that one does not see’ (Groundwork for the
Metaphysics of Morals, 4. 407). Zagzebski might well deny that
 moral value is not concerned with actions, but given her
emphasis on motivation as the ultimate explainer of value,
Kant’s point about the hiddenness of our motivations surely
undercuts any pretense that we have any sort of infallibility
about our moral exemplars merely from their being our
exemplars.

41. Zagzebski 2004: 253.

42. Ibid.225–6.

43. This is distinct from Adams’s strategy, which is metaethical—

he is arguing that the moral property of being morally required
is itself the theistic property being commanded by God. The
appeal to divine authority here is a normative strategy. It takes
no stand on the nature of moral properties, holding only that
the action-type obeying God exhibits the property being morally
required.

44. For a discussion of how God commands, see Adams 1999: 262–
70; for a more general account of God’s performance of speech
acts, see Wolterstor 1995.

45. This is a relatively strong version of a divine authority

principle. One might distinguish between one formulation of
DA that requires one’s actions to constitute obedience in order
for the principle to be satis ed and another that requires one’s
actions simply to be compatible with God’s commands in order
for the principle to be satis ed: respectively, people are under
 a moral requirement to take God’s commands as their guide to
action and people are under a moral requirement to perform
those actions that God tells them to perform. Both may
plausibly be called authority principles, because both are such
that God’s say-so is a constituent part of any reason to perform
an action grounded in that principle. Only the former, though,
is such that in order to satisfy it one must act under the
description ‘doing what God says to do’.

46. Other permutations are possible, in which the principle that

God is to be obeyed is superior to some but not all other moral
principles; in which the principle that God is to be obeyed is
the source of some but not all important moral duties; and so
forth.

47. What follows is argued in much more detail in Murphy 2002a.

48. Adams 1999: 355.

49. One might say that it diminishes God’s power if God cannot
place us under moral requirements by giving us commands.
But it is no diminution of God’s power to say that there are
some things that God cannot do by way of doing other things
—at least, that God cannot do some things by way of other
things unless God has realized the further conditions that
enable this connection to hold. It does not diminish God’s
power to say, and to say truly, that God can create stop signs
by making octagonal red pieces of metal only if the world that
God has made includes the social practice of having red
 octagonal pieces of metal as signs to stop. The point is that
whether God’s commands bring about moral requirements
may depend on certain background conditions obtaining (for
example, the existence of a distinct sort of rational being),
background conditions that are up to God’s discretion, the
failure of which to obtain precludes God’s commands from
being moral requirements.

50. Given the failure of traditional divine perfections to provide

the basis for an argument to divine authority, one might be
tempted to posit authority as a divine perfection. See below
for a discussion of this strategy.

51. ‘Re ective equilibrium’ is that condition in which our

considered general and particular moral judgments form a
coherent, mutually supportive whole; see Rawls 1999: 18–19.

52. For a skeptical treatment of parental authority, see Slote 1979;

for such treatments of the authority of the state, see Simmons
1979; Raz 1979; and Green 1990.

REFERENCES

ADAMS, ROBERT M. (1999). Finite and In nite Goods. Oxford: Oxford

University Press.

ALMEIDA, MICHAELJ. (2004). ‘Supervenience and Property-Identical

Divine-Command Theory’, Religious Studies 40: 323–33.
ARISTOTLE (1985). Nicomachean Ethics, trans. Terence Irwin.
Indianapolis: Hackett.

CHANDLER, JOHN (1985). ‘Divine Command Theories and the Appeal

to Love’, American Philosophical Quarterly 22: 231–9.

FREDDOSO, ALFRED J. (1988). ‘Medieval Aristotelianism and the Case

against Secondary Causation in Nature’, in Thomas Morris (ed.),
Divine and Human Action. Ithaca: Cornell University Press, 74–
118.

———(1994). ‘God’s General Concurrence with Secondary Causes:

Pitfalls and Prospects’, American Catholic Philosophical Quarterly
68: 131–56.

GEACH, PETER (1969). ‘The Moral Law and the Law of God’, in God
and the Soul. London: Routledge & Kegan Paul, 117–29.

GREEN, LESLIE (1990). The Authority of the State. Oxford: Oxford

University Press.

HARE, JOHN E. (1996). The Moral Gap. Oxford: Oxford University

Press.

——— (2001). God’s Call. Grand Rapids: Eerdmans.

HOBBES, THOMAS (1994). Leviathan, ed. Edwin Curley. Indianapolis:

Hackett.

KANT, IMMANUEL (1996a [1785]). Groundwork for the Metaphysics of

Morals, trans. Mary J. Gregor, in Kant’s Practical Philosophy.
Cambridge: Cambridge University Press.
———(1996b [1788]). Critique of Pure Reason, trans. Mary J.
Gregor, in Kant’s Practical Philosophy. Cambridge: Cambridge
University Press.

———(1998 [1793]). Religion with the Boundaries of Mere Reason,

trans. Allan Wood and George di Giovanni. Cambridge:
Cambridge University Press.

MURPHY, MARK C. (1998). ‘Divine Command, Divine Will, and Moral

Obligation’, Faith and Philosophy 15: 3–27.

———(2002a). An Essay on Divine Authority. Ithaca: Cornell

University Press.

———(2002b). ‘Theological Voluntarism’, Stanford Enyclopedia of

Philosophy, <http://plato stanford.edu/entries/voluntarism-
theological/>, accessed 20 May 2008.

———(2002c). ‘The Natural Law Tradition in Ethics’, The Stanford

Encyclopedia of Philosophy,
http://plato.stanford.edu/archives/win2002/entries/natural-
law-ethics/, accessed 20 May 2008.

PRUSS, ALEXANDER R. (2006). The Principle of Su cient Reason.

Cambridge: Cambridge University Press.

QUINN, PHILIP (1979). ‘Divine Command Ethics: A Causal Theory’, in

Janine Marie Idziak (ed.), Divine Command Morality. New York:
Edwin Mellen, 305–25.
———(1988). ‘Divine Conservation, Secondary Causes, and
Occasionalism’, in Thomas Morris (ed.), Divine and Human
Action. Ithaca: Cornell University Press, 50–73.

———(1990). ‘An Argument for Divine Command Ethics’, in

Michael Beaty (ed.), Christian Theism and the Problems of
Philosophy. Notre Dame: University of Notre Dame Press, 289–
302.

———(1999). ‘Divine Command Theory’, in Hugh LaFollette (ed.),

Guide to Ethical Theory. Oxford: Blackwell, 53–73.

RACHELS, JAMES (1981). ‘God and Human Attitudes’, in Paul Helm

(ed.), Divine Commands and Morality. Oxford: Oxford University
Press.

RAWLS, JOHN (1999). A Theory of Justice, rev. edn. Stanford, Calif.:

Harvard University Press.

RAZ, JOSEPH (1979). The Authority of Law. Oxford: Oxford University

Press.

SCHROEDER, MARK (2005). ‘Cudworth and Normative Explanations’,

Journal of Ethics and Social Philosophy1.

———(2007). ‘The Humean Theory of Reasons’, Oxford Studies in

Metaethics 2:195–219.

SIMMONS, A.JOHN (1979). Moral Principles and Political Obligations.

Princeton: Princeton University Press.
SLOTE, MICHAEL (1979). ‘Obedience and Illusions’, in Onora O’Neill
and William Ruddick (eds.), Having Children. Oxford: Oxford
University Press. WAINWRIGHT, WILLIAM A. (2005). Religion and
Morality. Aldershot: Ashgate.

WARRENDER, HOWARD (1957). The Political Philosophy of Hobbes: His

Theory of Obligation. Oxford: Oxford University Press.

WOLTERSTORFF, NICHOLAS (1995). Divine Discourse. Cambridge:

Cambridge University Press.

ZAGZEBSKI, LINDA TRINKAUS (2004). Divine Motivation Theory.

Cambridge: Cambridge University Press.
 CHAPTER 15

THE PROBLEM OF EVIL

PAUL DRAPER

INTRODUCTION

SUFFERING, immorality, and other evils generate a wide variety of

practical, a ective, and doxastic problems. Some of these problems
can be stated in the form of a question and some of these questions
are relevant to this handbook because they are theological—they are
connected in some relevant way to God or to faith in God or to the
metaphysical theory that God exists.

Practical problems of evil are concerned with what one ought to

do in response to evil. They are theological when the issue is what
actions one ought to perform, in light of one’s beliefs about God, the
assumption being that those beliefs should make a di erence to how
one responds to evil. If the question is whether or not one should
(continue to) rely on or trust God, then this assumption is obviously
true. It is also true, however, if the question is how best to oppose
evils such as oppression or how best to alleviate evils such as
su ering. One’s beliefs about God can make a di erence to how one
answers such questions as these as well.
 A ective problems of evil are concerned with whether the
emotions one feels in response to evil are appropriate. They are
theological when the issue is how one ought to respond emotionally
to evil, in light of one’s beliefs about God. Many theists have
emotional responses to evil that appear to con ict with their beliefs
about God. For example, evil may cause a believer to feel despair or
hopelessness or anger toward God or abandoned by God, and it is
hard to reconcile any of these feelings with the belief that an
almighty God loves us. Even an issue as apparently unproblematic
as how one should feel about the death of a loved one can become
controversial if one really believes that the loved one is in ‘a better
place’ and that one will be literally reunited with him or her after
one’s own death.

Both practical and a ective problems of evil raise doxastic

questions about evil. To determine what one ought to do or how one
ought to feel in response to evil, one must rst determine what one
ought to believe, including what one ought to believe about God, in
light of the evil in the world. It is important to notice, however, that
there are many distinct doxastic problems of evil, because beliefs
can possess or lack any of a variety of distinct merits. For example, a
belief about God can be true or false, it can be probable or
improbable (relative to the epistemic situations of its subject), it can
result from properly or improperly functioning cognitive faculties,
and its subject might be entitled to hold it or might instead hold it
in violation of some moral or epistemic duty. When judging the
signi cance of evil for belief in God, it is crucial that one be clear
about exactly which of these (or other) doxastic merits is at issue.

In this chapter, I will focus on questions about evil that are both
theological and doxastic, and more speci cally alethic—i.e.
questions about whether what we know about evil can be used to
establish the falsity or probable falsity of the belief or proposition
that God exists. Such a focus is natural for an agnostic (like me).1
More generally, it is natural for anyone who is engaged in genuine
inquiry about whether or not God exists. Such inquiry inevitably
raises questions such as: does the evil in the world provide the
resources for proving that God does not exist (e.g. because it can be
shown to be logically incompatible with God’s existence)? If not,
does it nevertheless provide some evidence against God’s existence
(in the sense that it lowers the probability of God’s existing)? If it
does, then how strong is this evidence? And if this evidence is
strong (in the sense that it decreases the ratio of the probability of
theism to the probability of its denial many-fold), is it also
signi cant (in the sense that it makes a large di erence to the
probability of theism being true)? Answers to these questions are no
doubt relevant to many other doxastic problems of evil (and to
a ective and practical problems as well), but alethic problems of
evil are not identical to any of those other problems. For example,
no alethic problem is identical to the problem of whether or not the
evil in the world renders belief in God irrational. After all, a false or
probably false belief can be rational (in more than one distinct sense
of that word); and a true or probably true belief can, of course, be
irrational.

I will further narrow my focus by employing a speci c concept

of God. In particular, I will assume that to assert that God exists or
that ‘theism’ is true is to assert that there exists a supernatural
person who created the natural world and who is perfect in power
(‘omnipotent’), perfect in knowledge (‘omniscient’), and perfect in
moral goodness (‘morally perfect’). This is obviously a narrow sense
of the words ‘God’ and ‘theism’, but it is common in the
philosophical literature.2 The strategy I will use for investigating
alethic problems of evil is also common: I will construct and
evaluate a variety of ‘arguments from evil’ for the conclusion that
God does not exist or that God’s existence is improbable.

MT ARGUMENTS

At rst glance, it may appear easy to construct a convincing

argument from evil against theism. One obvious strategy is to
formulate an argument of this very simple sort:
(1) If God exists, then e does not obtain.
(2) e obtains.
So, (3) God does not exist.
Notice that any argument from evil of this sort will be deductively
valid because it will have a valid argument form, namely, modus
tollens. For this reason, I will call arguments from evil of this sort
‘MT arguments’. Notice also that, so long as the variable e is
replaced with some fact about evil that is known to obtain, an
argument from evil of this sort will be sound if its rst premise is
true and it will be convincing if we have good reason to believe that
its rst premise is true. Since an ‘evil’ should be understood in
discussions of arguments from evil to mean anything that is bad,
including things that are bad because they imply the absence of
something good, there are countless candidates to replace e. For
example, e could be replaced by the fact that evil exists, or that
su ering exists, or that undeserved su ering exists, or that horri c
su ering exists, or that immorality exists, or that heinous
immorality exists, or that some innocent children have been
tortured, or that not every sentient being ourishes, or that some
sentient beings ourish while most do not. Which replacement for e
is chosen may a ect the prospects for showing that the rst premise
of the resulting argument is true. The crucial question is whether
this premise can be shown to be true for at least one suitable
replacement of e.

It is far from obvious that the correct answer to this question is

‘yes’. The rst premise is a conditional statement. Assuming it is a
material conditional, it is true if and only if it is not the case that
both its antecedent is true and its consequent false. In other words,
it is true if and only if it is false that both God exists and e obtains.
But how can the falsity of the proposition ‘God exists and e obtains’
be established? One approach is to argue that the falsity of this
proposition is entailed by the de nition of the title ‘God’. In other
words, one might try to argue that the existence of a being quali ed
to bear the title ‘God’—i.e. the existence of a loving supernatural
person who is omnipotent, omniscient, and morally perfect—is
logically incompatible with e’s obtaining. To argue in this way
would be to o er what philosophers have called a ‘logical argument
from evil’ against theism. All other arguments from evil are called
‘evidential arguments from evil’.

Although logical arguments from evil seemed promising to a

number of philosophers in the 1950s and 1960s (e.g. J. L. Mackie
1955), they are rejected by the vast majority of contemporary
philosophers of religion, for two reasons. First, some serious
attempts have been made to demonstrate that the existence of evil
and several other more speci c facts about evil are logically
compatible with God’s existence. For example, Alvin Plantinga’s
(1974) in uential ‘Free Will Defense’ persuaded many that it is
logically possible that God exists, that God creates people with
morally signi cant free will, and that some of those people make
morally bad choices. If Plantinga is right that this is logically
possible, then God’s existence is logically compatible with the
existence of evil and (more speci cally) with the existence of moral
evil. Plantinga also tried to extend the Free Will Defense to several
other facts about evil. For example, he tried to show that theism is
compatible with the amount of moral evil in the world, and he even
appealed to the possibility of non-human persons who have free will
(e.g. demons) in an attempt to show that theism is compatible with
the existence of ‘natural’ evil (i.e. evil that does not result from
human immorality). The appeal to demons may seem fanciful or
even desperate, but one must keep in mind that Plantinga is
responding to the logical argument from evil. If one’s goal is only to
prove logical compatibility, then it is legitimate to appeal to any
possible state of a airs, no matter how unlikely.

There are, however, many replacements for e that no defense has

shown to be compatible with theism. Thus, the unpopularity of
logical arguments from evil cannot be explained solely by the
success of the Free Will Defense or other defenses. A second, more
fundamental and hence more important reason for the unpopularity
of logical arguments from evil is that, even when some fact about
evil cannot be shown to be compatible with theism, establishing an
incompatibility is no less beyond our abilities (Pike 1963). Granted
a God, being omnipotent and morally perfect, could eliminate any
evils that he wanted to eliminate and would eliminate any evils that
he had no good moral reason to permit; but all that follows from
these claims is that a God would eliminate any evil that he had no
good moral reason to permit. In order for a logical argument from
evil to succeed, it is necessary to show that, for some known fact
about evil, it is logically impossible for God to have a good moral
reason to permit that fact to obtain. This, however, is precisely what
most philosophers nowadays believe cannot be shown.
 But why do they believe this? After all, when one examines the
sorts of ‘excuses’ that human beings have for not eliminating evil,
they all seem to involve imperfect power or imperfect knowledge.
For example, human beings avoid blame for destructive tornadoes
because they lack the power to prevent them. They avoid blame for
not rescuing people lost at sea because they don’t know their
whereabouts. And, depending on the circumstances, they may avoid
blame for evils they know about and have the power to eliminate,
such as visits to the dentist, so long as those evils cause (or are
caused by) greater goods such as healthy teeth. These sorts of
excuses could not work for an omnipotent and omniscient being.
This is obvious in the case of the rst two types because they
explicitly involve a lack of power or knowledge. But it is no less true
about the third case, which implicitly involves a lack of power
because an omnipotent being’s power would not be limited by
causal laws. For example, an omnipotent being would never need to
use unpleasant dental procedures as a causal means of keeping
someone’s teeth healthy. Her merely willing that a person’s teeth be
healthy would su ce to make it so.

Still, it would be a mistake to jump to the conclusion that all

good reasons for not eliminating evil involve some lack of power or
knowledge. To understand why this would be a mistake, it is crucial
to recognize that the inability to produce things that are logically
impossible to produce, or to know statements that are logically
impossible to know, does not count as a lack of power or a lack of
knowledge. In other words, not even an omnipotent and omniscient
being would have more power or more knowledge than it is
logically possible for a being to have. Next, suppose that some good
that is worth my su ering (perhaps it is even a good that bene ts
me) logically implies that I su er (or at least that the objective
chance of my su ering is not zero). For all we know or can prove,
there could be such a good, even if it is beyond our ken. After all,
even some goods we know about logically imply evils. For example,
my showing fortitude in the face of pain logically implies that I feel
pain and thus not even an omnipotent being could produce the good
of my showing such fortitude without permitting the evil of my
feeling pain. Of course, the good of my showing fortitude in the face
of pain is probably not worth my feeling pain, especially since
fortitude of other sorts does not require pain, but for all we know
other goods not only logically imply my su ering but are also worth
my su ering. Such goods would be known to an omniscient being
even if they are unknown to us. Further, if there are such goods,
then not even an omnipotent and omniscient being could produce
them without allowing me to su er and hence even an omnipotent
and omniscient being might have a good moral reason to permit my
su ering.3

One might object here that no good, no matter how great, could
justify allowing such horri c evils as the torturing of innocent
children. (The character Ivan in Fyodor Dostoevsky’s The Brothers
Karamazov is often interpreted to be making this claim.) Indeed,
some philosophers (e.g. D. Z. Phillips 2004) seem to object to the
whole notion of an ‘outweighing’ good. To be sure, this notion
should be rejected if it implies that all value can be measured on a
single numerical scale or if it presupposes a crude consequentialist
understanding of morality. It need not be interpreted in this way,
however. In fact, it is even compatible with the position that no
good, no matter how great, ‘outweighs’ the harm done to an
individual unless it bene ts that individual. This last point is
important because, for all we know, (1) there may exist goods far
more valuable than any we can imagine, (2) these goods may
logically imply the existence or risk of horri c evils, and (3) these
goods may (if there is life after death) include among their
bene ciaries the victims of horri c evils. If this possibility is taken
seriously (and admittedly not all philosophers think it should be
taken seriously or even that it is a possibility), then it is hard to be
con dent that the notion of an outweighing good breaks down in the
face of horri c evil, especially given how imprecise and fallible our
moral intuitions are. Therefore, since I believe this possibility should
be taken seriously, I do not see how it is possible to construct a
convincing logical argument from evil against theism, and for that
reason the rest of the arguments from evil I discuss in this chapter
will be evidential ones.4

Notice that not all MT arguments are logical arguments from

evil. Consider, for example, the following two compound
arguments.5
 (1) No known goods justify God’s permitting horri c
su ering.
So, (2) No goods, whether known or not, justify God’s permitting
horri c su ering.
(3) God’s existence entails that either some goods justify God’s
permitting horri c su ering or horri c su ering does not
exist.
So, (4) If God exists, then horri c su ering does not exist.
(5) Horri c su ering exists.
So, (6) God does not exist.
(1′) Horri c su ering is appallingly bad (and the worse an
evil is, the less likely it is that God would permit it).
(2′) God would be maximally empathetic (and the more
empathetic a being is, the less likely it is that such a being
would permit horri c su ering even if that being had a
morally su cient reason to do so).
(3′) Some goods that God would prima facie want to obtain,
like the ourishing of all sentient beings, are incompatible
with horri c su ering.
So, (4) If God exists, then horri c su ering does not exist.
(5) Horri c su ering exists.
So, (6) God does not exist.
Notice that these two compound arguments share steps 4–6, which
is a standard MT argument from horri c su ering. Neither
argument, however, defends the conditional premise of that MT
argument by trying to show that its antecedent is a logically
su cient condition of its consequent.

In the rst argument, this conditional follows deductively

(necessarily) from steps 2 and 3, but step 2 is supposed to follow
inductively from step 1. In other words, the truth of step 1 is
supposed to make the truth of step 2 probable. This inductive
inference is questionable. It involves reasoning from a sample of
‘known goods’ to a conclusion about all goods whether known or
not. Obviously there is nothing wrong in principle with reasoning
from samples to populations, but such inferences are correct only
when one has good reason to believe that one’s sample is
representative of the population. One problem here is that we know
so little about the population of all goods that it is hard to be
con dent that our sample of known goods is representative of that
population (Alston 1991). A second problem is that any evidence we
have that supports theism will ipso facto be evidence that our
sample is biased and so may undermine the inference in question.

The second argument defends the conditional premise that if

God exists then horri c su ering does not exist by giving reasons
why a God would be unlikely to permit horri c su ering to exist.
However, this strategy also fails because steps 1–3 are not by
themselves good reasons to believe that 4 is true. To begin with,
there might be strong evidence for God’s existence. Any such
evidence would, given the existence of horri c su ering, be strong
evidence for the falsity of the statement that if God exists then
horri c su ering does not exist. But even if no such evidence exists,
giving reasons like the ones described above for believing that a
God would be opposed to horri c su ering does not prove that the
statement ‘if God exists, then horri c su ering does not exist’ is
probably true. Instead, such reasons only show at most that the
consequent of that conditional statement is antecedently very
probable, given (i.e. ‘conditional on’) the truth of its antecedent. In
short, instead of establishing the probability of a conditional, they
establish only a conditional probability. And the former does not
follow from the latter: from the fact that Q is antecedently probable
given P, it does not follow that ‘if P then Q’ is probable.

This is a very easy mistake to make. Suppose, for example, that

Detective Garcia is trying to locate a car and is certain that it is
owned by either Smith or Jones. In her e orts to gure out which of
the two is the owner, she discovers that the car was recently painted
red. She then reasons:

(1) Smith really likes green cars.

So, (2) If Smith owns the car, then it was not recently painted red.
(3) It was recently painted red.
So, (4) Smith does not own the car.
The inference from 1 to 2 appears strong, but it is not. From the fact
that Smith really likes green cars, it follows only that it is
antecedently probable that the car would not have been painted red
given that Smith owns it. It does not follow that the conditional
statement ‘if Smith owns the car, then it was not painted red’ is
probably true. To see why, suppose that Garcia also knows, not only
that Jones really likes yellow cars but also that he was once struck
by a red car and has had a strong aversion to red cars ever since.
Then, even though it is antecedently likely that the car was not
painted red given that Smith owns it, it is even more likely that it
was not painted red given that Smith does not own it. This means
that the fact that the car has recently been painted red is actually
evidence that Smith is the owner! So from the fact that Smith really
likes green cars it does not follow even inductively that, if Smith
owns the car, then it was not recently painted red.

Similarly, from the fact that the non-existence of horri c

su ering is antecedently likely given theism, it does not follow even
inductively that, if God exists, then horri c su ering does not occur.
To draw that conclusion, one would need to show, among other
things, that the non-existence of horri c su ering is less likely on
the assumption that God does not exist than it is on the assumption
that God exists. And showing that is very di cult since, as I will
explain in more detail later, it is far from clear what exactly one is
assuming to be true when one assumes that the statement ‘God
exists’ is false.

BAYESIAN ARGUMENTS FROM EVIL

Problems such as these make me doubt that a convincing MT

argument from evil can be constructed. I’m much more optimistic
about what I call ‘Bayesian’ arguments from evil. (I call them this
because the structure of the reasoning in these arguments can be
analyzed, made precise, and defended by appealing to a theorem of
mathematical probability called ‘Bayes’ theorem’.) I’m more
optimistic about these arguments for at least two reasons. First, they
conclude only that, other evidence held equal, theism is very probably
false and thus they do not presuppose without justi cation that
there is no o setting or even outweighing evidence in support of
theism. Second, they avoid the problem faced by the second
evidential MT argument discussed above by explicitly comparing the
probability of certain facts about evil given theism to the probability
of those facts given some relevant alternative hypothesis.

Bayesian arguments from evil (e.g. Draper 1989) have the

following structure:

(1) We know that e obtains.

(2) e’s obtaining is antecedently many times more probable
given some alternative hypothesis h to theism than it is
given theism.6
So, (3) e’s obtaining is strong evidence favoring h over theism (i.e.
our knowledge that e obtains increases the ratio of the probability of
h to the probability of theism many-fold).
(4) h is at least as plausible as theism (i.e. h is at least as
probable as theism independent of all evidence for and
against theism and h, or at least we have no good reason to
believe otherwise).
So, (5) Other evidence held equal, theism is very probably false.
The inferences in arguments of this sort are not as uncontroversial
as the inference in an MT argument. However, Bayes’ theorem
supports the inference from steps 1 and 2 to step 3 so long as the
claim that e’s obtaining is strong evidence favoring h over theism is
understood to mean that our knowledge that e obtains increases the
ratio of the probability of h to the probability of theism many-fold.
Notice that the only way that a Bayesian argument could have a
false conclusion (step 5) despite having a true sub-conclusion (step
3) is if theism were more probable than the alternative hypothesis h
apart from considerations of evidence. Step 4 of the argument,
however, states that this is not the case, that in fact h is at least as
‘plausible’ as theism. Therefore, step 5 follows from steps 3 and 4: in
the absence of additional evidence besides e that favors theism over
h, theism is very probably false. Notice that it also follows that in the
absence of additional evidence besides e that strongly favors theism
over h, theism is probably false.

I should mention also that the word ‘antecedently’ in the second

premise of Bayesian arguments is crucial. What this means is that
the probabilities being compared are not ‘all things considered’
probabilities. Instead, an abstraction is being made. Premise 2
claims that, independent of the observations and testimony upon which
our knowledge speci cally of e is based, e is much more probable
given h than it is given theism. When assessing the probabilities in
the second premise of a Bayesian argument, however, one should
take into account most of what one knows, including that we live in
a complex physical universe containing a variety of living things,
some of which are conscious beings, some of which are self-aware
beings, and some of which are moral agents. Some philosophers
(like me) believe that at least some of these known facts are
evidence strongly favoring theism over competing hypotheses. Even
if that is so, it is not relevant to the evaluation of premise 2 of the
argument, but instead is taken into account by the clause ‘other
evidence held equal’ in the conclusion of the argument, which is to
say that the conclusion of a Bayesian argument from evil is
compatible with the claim that, on the total evidence, theism is
probably true.

Like MT arguments, Bayesian arguments from evil vary

depending on which facts about evil are used to replace the variable
e. Unlike MT arguments, however, Bayesian arguments also vary
because di erent hypotheses are chosen to replace the variable h.
One natural choice is, of course, atheism. Consider, then, the
following Bayesian argument from evil:

(1) We know that horri c su ering exists.

(2) The existence of horri c su ering is antecedently many
times more probable on the assumption that God does not
exist (atheism) than it is on the assumption that God exists
(theism).
So, (3) The existence of horri c su ering is strong evidence
favoring atheism over theism.
 (4) Atheism is at least as plausible as theism.
So, (5) Other evidence held equal, theism is very probably false.
This argument appears to be very convincing, but it does have one
major aw. Its second premise is not obviously true and it is hard to
see how one could successfully argue for its truth.

Good reasons can, I believe, be given for believing that horri c

su ering is unlikely given theism, but it is di cult to show that
horri c su ering is any less unlikely given atheism. This is di cult
to show because the probability of horri c su ering given atheism
depends on what is likely to be true on the assumption that theism
is false. For example, if theism is false, then naturalism might be
true. Or perhaps panentheism is true. Or maybe pantheism or
polytheism. Indeed, if theism is false, then isn’t there a good chance
that what is true is some non-theistic form of supernaturalism that
can’t even be understood by human beings because of their
cognitive limitations? Some theists might think so. The point is that,
even if we could agree on a list of serious possibilities and restrict
our attention to those (estimating the probability of horri c
su ering given atheism by taking a probability weighted average of
the probability of horri c su ering on each of those possibilities),
the calculation of the probability of horri c su ering given atheism
would be prohibitively di cult.

William Rowe, whose 1996 argument from evil explicitly appeals

to Bayes’ theorem and uses atheism as the alternative to theism,
ingeniously tries to circumvent this problem by using an evidence
statement that is entailed by atheism and hence has a probability of
one given atheism. Consider, for example, this evidence statement:

(R) For all goods g that we know of, g does not justify God in
permitting horri c su ering.

Rowe’s strategy yields the following Bayesian argument:

(1) We know that R.
(2) R is antecedently many times more probable on the
assumption that God does not exist (atheism) than it is on
the assumption that God exists (theism).
So, (3) R reports strong evidence favoring atheism over theism.
(4) Atheism is at least as plausible as theism. So, (5) Other
evidence held equal, theism is very probably false.
Since a good can justify God’s permitting horri c su ering only if
God exists, atheism entails R and hence the probability of R given
atheism is 1, which is, of course, greater than the probability of R
given theism.

This argument faces two problems, however. The rst is that,

while R is obviously more probable on atheism than on theism, it is
not clear that it is much (i.e. many times) more probable. Given
theism and the existence of horri c su ering, perhaps it is not all
that surprising that the goods that justify God’s permitting horri c
su ering are unknown to us. In any case, additional arguments
would be needed to prove otherwise. If it cannot be shown that the
probability of R given atheism is at least one order of magnitude
greater than the probability of R given theism, then R is only
relatively weak evidence for atheism and hence of little signi cance.
The second problem is more serious in the sense that a solution
seems impossible.

This problem was rst clearly identi ed by Richard Otte (2003).

Unfortunately, the technical nature of Otte’s point has resulted in
the importance of his paper going largely unnoticed. I will brie y
reconstruct Otte’s reasoning here. Readers with no background in
con rmation theory may wish to skip the next two paragraphs.

Although Otte himself does not put it exactly this way, the
problem he identi ed is that the evidence statement R very subtly
understates what we know about the relationship of goods we know
of to horri c su ering. This makes it appear that there is a stronger
case against theism based on ‘inscrutable evils’ than there really is.
To see why, compare R with

(R*) For all goods g that we know of, either God exists and g does
not justify God in permitting horri c su ering, or God does
not exist and, if God were to exist, then g would not justify
God in permitting horri c su ering.

R* entails R but is clearly not entailed by R since, unlike R, R* is not

entailed by atheism. Further, our knowledge that R is true is based
(at least in this context), not on any alleged knowledge that God
does not exist but rather on the knowledge that R* is true.7 The
crucial point is that, while the understated evidence R is more
probable on atheism than on theism, it is far from clear that the
fully stated evidence R* is (and even less clear that it is much more
probable on atheism than on theism). The mistake in reasoning here
is essentially the same (though it is much harder to detect and is not
intentional) as the mistake a prosecutor makes when he points out
to a jury that a defendant accused of murder purchased a gun
shortly before the victim was shot to death, neglecting to mention
that the gun in question is a stun gun.

To appreciate this analogy, notice that the evidential argument

from R is based on the fact that R is entailed by atheism and thus
P(R/∼T) = 1 > P(R/T). (Stated in English, this says that the
probability of R given atheism equals one, which is greater than the
probability of R given theism.) But given that theism is true, R can’t
be true unless R* is; while given that atheism is true, R can be true
without R* being true. This means that, given theism and R, R* is
certain, while given atheism and R, R* is not certain: P(R*/T&R) =
1 > P(R*/∼T&R). Thus, relative to R, R* is evidence for theism!
(Compare: the fact that the defendant bought a gun is more likely
on the assumption that he killed the victim, but given that he did
buy a gun, it is more likely that it is a stun gun given that he did not
kill the victim.) It follows that the argument from R could be
repaired only if it could be shown that P(R*/∼T&R) > P(R/T),
from which it would follow that the fully stated evidence R* is more
probable on atheism than on theism—i.e. that P(R*/∼T) >
P(R*/T). [This would follow because R* entails R and so is logically
equivalent to R&R*, which implies that, for any hypothesis h,
P(R*/h) = P(R&R*/h) = P(R/h) x P(R*/R&h).] Indeed, to avoid the
additional objection that we are dealing only with relatively weak
evidence here, what really needs to be shown is that P(R*/∼T&R)
>> P(R/T) and hence that P(R*/∼T) >> P(R*/T), where the
symbol ‘>>’ means ‘is many times greater than’. But showing that
P(R*/∼T&R) is many times greater than or even that it is greater
than P(R/T) is no easy task since, as I pointed out earlier, it is far
from clear what would be true on the assumption that theism is
false, and thus it is very hard to see how any precise calculation of
P(R*/∼T&R) can be made. So if we replace R with R*, then we are
back to the same problem that the argument from R was supposed
to solve. And if we don’t replace R with R*, then we achieve success
only by misleadingly understating the relevant evidence. Either
way, we don’t have a powerful evidential argument from evil
against theism.

I submit, then, that the best way to formulate a Bayesian

argument from evil is to use a speci c alternative hypothesis to
theism instead of the general hypothesis that theism is false. While
doing so will make the third premise of the argument (step 4)
harder to defend, it will make the second premise much easier to
defend. Not any more speci c alternative will do, of course, even if
it renders premise 2 true. For if it is too speci c (e.g. if one simply
builds the facts about evil mentioned in premise 1 into the
hypothesis), then premise 4 will be either false or indefensible.
Thus, choosing the correct hypothesis is crucial. One promising
candidate is the ‘hypothesis of indi erence’ or ‘HI’ for short (Draper
1989), which states that neither the nature nor the condition of
sentient beings on earth is the result of actions performed by
benevolent or malevolent non-human persons. (I will argue later
that this hypothesis is more plausible than theism.) Now let ‘O’
stand for a statement reporting what we know about the pain and
pleasure in the world. Using HI as the alternative hypothesis to
theism and O as the evidence statement yields the following
Bayesian argument from evil:

(1) We know that O.

(2) O is antecedently many times more probable on the
assumption that HI is true than it is on the assumption that
theism is true.
So, (3) O is strong evidence favoring HI over theism.
(4) HI is at least as plausible as theism.
So, (5) Other evidence held equal, theism is very probably false.

This argument is similar in two important ways to the argument

formulated by the character Philo in Part 11 of David Hume’s
Dialogues Concerning Natural Religion (1779).8 First, Philo also
compares theism to a sort of ‘indi erence hypothesis’ (namely, that
the causes of the universe are neither benevolent nor malevolent).
Second, like O, Philo’s evidence statement mentions not just pain,
but also pleasure. This is an important insight on Hume’s part.
Logical arguments from evil can ignore the good in the world (as
Philo emphasizes in Part 10 of the Dialogues), but evidential
arguments from evil cannot. Pointing out that various facts about
pain are evidence against theism is of little signi cance if parallel
facts about pleasure are equally strong evidence for theism. Of
course, there are other goods and evils besides pleasure and pain,
but facts about those goods and evils are part of the evidence that is
being ‘held equal’ by the argument.

Most of my 1989 article is devoted to defending premise 2 of this

argument. Very roughly, I rst make a prima facie case for premise
2 by arguing that we have much more reason on HI than we have
on theism to expect that pain and pleasure will play the biological
roles they do. After all, on HI, the fact that other parts of organic
systems systematically promote survival and reproduction supports
the claim reported by O that pain and pleasure will do the same. But
on the assumption that theism is true, the moral signi cance of pain
and pleasure undermines this support. On theism, it would not be
surprising at all if pain and pleasure played a fundamentally moral
role in the world without also playing the same biological roles that
other parts of organic systems play. Further, even assuming that
God would have a good moral reason to use pain and pleasure to
promote the biological goals of survival and reproduction, what we
know about biologically gratuitous pain and pleasure (e.g. that it
includes horri c su ering and does not include an abundance of
pleasure) is also more likely on HI than it is on theism.

I then turn this prima facie case for premise 2 into an ultima
facie one by arguing for two further claims. The rst is that,
contrary to what the theodicist would have us believe, the known
moral roles played by pain and pleasure in the world do not
signi cantly raise the probability of O given theism. They may raise
the probability of certain individual facts reported by O (e.g. the
fact that su ering sometimes leads to improved moral character),
but only by making other facts reported by O even more surprising
(e.g. the fact that su ering is often demoralizing). The second claim
is that, contrary to what the skeptical theist would have us believe,
the possibility of God having moral reasons unknown to us to permit
O does not undermine my case for premise 2, because God’s having
reasons to permit O that are unknown to us is no more likely
antecedently than God’s having reasons to prevent O that are
unknown to us.

Of course, the skeptical theist could respond that the possibility

of unknown justifying goods is not just supposed to defeat a prima
facie case for premise 2, but instead prove that there is no such case
to be defeated—that O, for example, is not even prima facie more
probable on HI than on theism because it has no discernible
probability on theism. In defense of this response, skeptical theists
might assert that O has a discernible probability on theism only if
the probability that a God would have a good reason to permit O is
discernible. They might then argue that we are in no position to
judge how likely or unlikely it is that a God would have a good
reason to permit O. One problem (from a theistic perspective) with
this response, however, is that it undermines evidential arguments
in support of theism. Another more fundamental problem is that it
ignores the role that background knowledge plays in conferring
di erential probabilities on O given various hypotheses. In addition,
it assumes incorrectly that any probability that would, if known,
a ect a second probability must actually be known in order for that
second probability to be known. If that assumption were true,
skepticism about all non-trivial probability judgments would
follow.9

Turning now to premise 4,10 it is important to realize that

philosophers do not agree on how such a premise should be
interpreted in the context of a Bayesian argument. Some
philosophers believe that plausibility judgments are purely
subjective. We just nd ourselves taking some hypotheses seriously
while dismissing others out of hand. The former we judge worthy of
being tested by evidence, while the latter we ignore. If this is
correct, then a Bayesian argument from evil like the one being
considered may still be sound relative to the epistemic situations of
those who take HI seriously (which I suspect includes most
contemporary philosophers, including many philosophers who are
theists), while it won’t be sound relative to the epistemic situations
of others for whom HI is not a live option.

Other philosophers believe that plausibility judgments are

objective (or at least are subject to substantial objective constraints).
For example, my own view is that the plausibility of a hypothesis
should be equated with its intrinsic (epistemic) probability—the
probability it has simply by virtue of what it asserts and what we
know about the world by means of rational intuition. The intrinsic
probability of a hypothesis depends on its coherence (in the sense of
the degree to which the various statements it entails evidentially
support one another11) and its modesty (in the sense of the degree
to which it is narrow in scope and lacks speci city). If plausibility is
a function of coherence and modesty, then HI is at least as plausible
as theism, because HI is no less coherent than theism, and it is
considerably more modest.

Allow me to explain the latter point. The degree of modesty of a

hypothesis depends on how little that hypothesis purports to tell us
about the world that we do not already know by rational intuition.
Relative to certain practical goals, a lack of modesty is often a
virtue; but relative to the goal of truth, it is a vice. For the more that
a hypothesis asserts that (for all we know by rational intuition)
might be false, the more likely it is (prior to considering evidence) to
assert something that is false, and hence, in the absence of evidence,
the less likely it is to be true. HI is clearly more modest than theism.
It is compatible with metaphysical naturalism (i.e. with the view
that there are no ‘supernatural’ entities, no entities that can a ect
natural entities despite not being natural themselves), as well as
with a variety of supernaturalist hypotheses. Theism, on the other
hand, claims not only that all natural entities possess (proximate or
remote) supernatural causes, but also that they depend for their
existence on a single ultimate supernatural cause, and further that
this cause is a person, and even more speci cally an omnipotent,
omniscient, and morally perfect person. Because of the great
speci city of these claims, it is fairly safe to conclude that HI is
considerably more plausible than theism and it is very safe to
conclude that premise 4 is true—that HI is at least as plausible as
theism. One might object that some theists maintain that God
necessarily exists, but that won’t make theism any less speci c and
thus won’t raise the intrinsic probability of theism because we don’t
know by rational intuition that God’s existence is necessary. In other
words, even if God’s existence is necessary, we can’t just ‘see’ the
necessity of the proposition ‘God exists’ in the way we can see the
necessity of ‘all dogs are dogs’ and ‘2 + 2 = 4’.

OBJECTIONS TO BAYESIAN ARGUMENTS FROM EVIL

In spite of their advantages over MT arguments, Bayesian arguments

from evil have been challenged in several di erent ways. I will
devote the remainder of this chapter to discussing some of the most
common or important of these challenges.

One popular challenge concerns Bayesian arguments that use

naturalism as the alternative hypothesis to theism. According to this
challenge, evil can’t be more probable on naturalism than on theism
because nothing is evil (or good) if naturalism is true. Defenders of
Bayesian arguments have two replies available to them. The rst is
to argue that naturalism does not rule out the existence of genuine
evil— i.e. that some entities are bad, even on the assumption that
naturalism is true. The second is to use an evidence statement that
makes no value judgments; for example, it might simply describe
the su ering in the world without either a rming or denying that
su ering is bad. Such a statement could very well be antecedently
more probable on naturalism than on theism. For while the facts
reported by such an evidence statement might be unsurprising on
naturalism whether or not su ering has negative value, they might
be surprising on theism precisely because su ering would be
objectively bad on the assumption that theism is true.

A second objection to Bayesian arguments is that facts about

su ering must be more likely on theism than on any plausible
alternative to theism because su ering cannot exist unless sentient
beings exist and the existence of sentient beings is more likely given
theism than it is given any plausible alternative to theism. This
objection fails for two reasons. First, it overlooks the fact that,
typically, knowledge of the existence of sentient beings would be
part of the epistemic situation relative to which the probability
judgments in a Bayesian argument from evil are made (and so
would be part of the ‘other evidence held equal’ mentioned in the
conclusion of the argument). Second, even in the case of a Bayesian
argument that included the existence of sentient beings in its
evidence statement, the inference in this objection would be
incorrect. For example, suppose that the evidence statement in some
Bayesian argument were ‘sentient beings exist and sometimes su er
horri cally’. It doesn’t follow from the (alleged) fact that the
existence of sentient beings is more likely on theism than on some
alternative hypothesis h that the existence of sentient beings that
su er horri cally is not more likely or even much more likely on h
than on theism.

A third objection to Bayesian arguments from evil is that they

compare some atheistic hypothesis such as naturalism or HI to
‘generic’ or mere’ theism rather than to a sectarian theistic
hypothesis such as Christian theism or Jewish theism or Muslim
theism. Thus, while they may create epistemic problems for generic
theists (i.e. for theists who do not accept any speci c revealed
religion), they do not create any epistemic problems for Christian or
Jewish or Muslim theists, especially since the evidence statements to
which Bayesian arguments typically appeal are entailed by (and
thus antecedently certain on) these sectarian theistic hypotheses
(Otte 2004). The short answer to this objection is that sectarian
theistic hypotheses entail theism and thus can’t be more probable
than theism.12 Thus, any reason to believe that theism is improbable
is also a reason to believe that Christian or Jewish or Muslim theism
is improbable. Moreover, just because an evidence statement e is
more probable on, say, Christian theism than on, say, naturalism, it
doesn’t follow that there is a good Bayesian argument from e against
naturalism. For building e into a theistic hypothesis (e.g. by adding
Bible stories about su ering and other evils to one’s theistic
hypothesis) will make that hypothesis very speci c (immodest) and
so much less probable intrinsically than a hypothesis such as
naturalism. Further, the fact that the second premise of a Bayesian
argument would be false if theism is replaced with ‘Christian theism’
in the argument is irrelevant. The failure of the revised argument
does not imply that the original argument is not sound.

This is not to deny, however, that other religious (or non-

religious) beliefs may be relevant to the issue of whether the second
premise of a Bayesian argument from evil is true or false. Indeed,
religious or ethical or metaphysical or epistemological beliefs could
potentially raise P(e/T)—i.e. the antecedent probability of the
evidence statement e given theism. To determine if they do, one
should apply this ‘weighted average principle’:

P(e/T) = P(B/T) x P(e/T&B) + P(∼B/T) x P(e/T&∼B).

The idea here is that, if B is, for example, some Christian belief, then
P(e/T) will be an average of P(e/T&B) and P(e/T&∼B). What makes
it an average is the fact that P(B) + P(∼B) = 1. But it is not
necessarily a straight average because P(B/T) and P(∼B/T) may not
each equal Thus, for example, if P(B/T) = and P(∼B/T) = then
P(e/T&B) is given twice as much weight as P(e/T&B) in calculating
the average.

Consider, for example, the belief that there is life after death (L’
for short). Suppose that P(L/T) were very high and also that
P(O/L&T) were much greater than P(O/HI). Then the second
premise of my evidential argument from evil would be false,
because P(O/T) = P(L/T) x P(O/L&T) + P(∼L/T) x P(O/∼L&T).
Of course, while I do believe that P(L/T) is high, I don’t believe for a
moment that P(O/L&T) is much greater than P(O/HI). The
hypothesis that we will survive death explains very little of what we
know about pain and pleasure. The point I want to make here,
however, is that there is a precise method for evaluating the
relevance of other religious and non-religious beliefs to Bayesian
arguments from evil. The same method should be used to evaluate
how successful various theodicies or defenses are in undermining
(the second premise of) a Bayesian argument from evil.

The last objection to Bayesian arguments that I will consider,

one raised by Alvin Plantinga (1996, 2000, 2007), is that Bayesian
arguments from evil are insigni cant, even if they are sound,
because there is overwhelming non-propositional (or propositional)
evidence favoring theism over atheism.13 If it were obvious that
such other evidence exists (or even if it were obvious to all theists
that this other evidence exists), then this objection might be more
convincing. Clearly, however, this is not obvious, even to theists.
Many theists (let alone atheists and agnostics) reject the natural
theologian’s position that there is overwhelmingly strong
propositional evidence for God’s existence, and many theists reject
Plantinga’s position that there is very strong non-propositional
evidence for beliefs about God. Concerning the latter, Plantinga
himself emphasizes that he cannot prove his position; his goal is just
to show that the existence of such evidence is epistemically possible
and that, if (Christian) theism is true, then it is likely that many
theists know that God exists because of such evidence. I nd even
this modest claim hard to accept. Even if God exists, the behavior of
most theists in a variety of circumstances suggests that they lack the
sort of deep con dence in their beliefs about God that would be
necessary for genuine knowledge. Many don’t even claim to have
such knowledge, and many who do claim to have such knowledge
betray their deep uncertainty by undermining that claim with
quali cations (e.g. ‘I know in my heart not my head’) or by
behaving in a variety of other ways (e.g. feeling terror in the face of
imminent death) that would be inappropriate or surprising for
someone who truly had such knowledge.

There are, in my opinion, additional problems with Plantinga’s

sophisticated form of apologetics,14 but I have no space in this
chapter to discuss them. Instead, let’s suppose that I am wrong
about all this. Let’s suppose that it is perfectly reasonable for most
theists to be unmoved by Bayesian arguments from evil because
they possess overwhelming non-propositional evidence in support of
theism. Even supposing all that, it still does not follow that Bayesian
arguments from evil are insigni cant. It does follow that they are of
no interest to Plantinga, because he is interested primarily in the
question of whether what we know about evil threatens the
rationality or (external) warrant of ‘most’ theists or perhaps most
traditional Christian theists. Plantinga’s mistake, however, is to
assume that the signi cance of an argument from evil depends on its
proving that most theistic belief is irrational or unwarranted. In
other words, he assumes incorrectly that the signi cance of alethic
problems of evil derives solely from their relationship to other
doxastic problems of evil. The reason that this is a mistake (a
mistake also made in Draper 1991) is that there are many people
who, like me, don’t believe they already know that God exists (or
that God doesn’t exist), and for that reason believe that it is
appropriate and important to engage, not in apologetics, but in
genuine inquiry designed to determine, to the best of their ability,
whether or not God exists. Included here are agnostics as well as
theists and atheists who have doubts about God’s existence or non-
existence. These skeptical souls have little choice but to do their
best objectively to assess the available evidence. Thus, for them, the
fact that O or some other statement about evil reports strong
evidence favoring HI (or some other serious alternative to theism)
over theism is of great signi cance. For them, a Bayesian argument
from evil, if sound, is very important indeed. Notice also that for
them, indeed for anyone who doubts that God exists, it doesn’t help
to be told by Plantinga that, if God does exist, then some or even
most theists probably know he does!

NOTES

For helpful comments on earlier drafts of this chapter, I am

grateful to Tom Flint, Mike Rea, and William Rowe. I am also
indebted to various members of the philosophy department at
the University of Georgia, where I presented a draft of most of
the paper as part of their Kleiner Lecture Series.
1. No doubt the most important problems of evil are practical,
but since I do not believe in God, those problems are not, for
me, theological and thus are not relevant to this volume.

2. For a defense of this choice, see Swinburne 2004, ch. 5.

3. I am indebted to Nelson Pike for most of the points in this and

the preceding paragraph.

4. The previous seven paragraphs, as well as some of the points I

make in the remainder of this section and the next, also appear
in Draper 2007.

5. The rst of these arguments is similar to William Rowe’s

(1979, 1986) early evidential arguments from evil, and the
second is similar to J. L. Schellenberg’s (2000) evidential
argument from evil. I do not wish to claim, however, that
these two arguments do justice to Rowe’s and Schellenberg’s
sophisticated arguments from evil. Although I hope that my
criticisms of these two arguments will suggest strategies that
could be used successfully to refute Rowe’s and Schellenberg’s
arguments, proving that is beyond the scope of this chapter.

6. One might wonder why I require of a Bayesian argument from

evil that e’s obtaining be many times more probable on the
alternative hypothesis than on theism. After all, if this phrase
were deleted from premise 2, it would still follow from the
premises of a Bayesian argument from evil that, other
evidence held equal, theism is probably false (though it would
 no longer follow that, other evidence held equal, theism is very
probably false). My reason for restricting the class of Bayesian
arguments in this way is that very many facts are slightly more
or less probable on various alternatives to theism than they are
on theism. So Bayesian arguments from evil that don’t meet
this higher standard are of little or no signi cance.

7. In claiming that we know that R* is true, I assume here that

none of the attempts that have been made to explain horri c
su ering in terms of theism has succeeded. If this assumption
is incorrect, then horri c su ering should be replaced with
some other sort of evil or, following Rowe, some particular
evil for which the corresponding assumption is correct.

8. See Draper 2004 for a description of important di erences

between Philo’s evidential argument from evil and my own.

9. For a detailed defense of skeptical theism, see Bergmann’s Ch.

17 in this volume. For a detailed defense of my argument from
evil against skeptical theism, see Draper 1996.

10. I do not defend or even explicitly state premise 4 of the

argument in my 1989 article. It is implicit in my requirement
in that paper that HI be a ‘serious’ alternative to theism.

11. Consider the following two statements: (1) ‘ball #1 is red and
ball #2 is red’ and (2) ‘ball #1 is red and ball #2 is green’.
These two statements are equally speci c, but the rst is
 intrinsically more probable than the second because it is more
coherent.

12. For a much longer answer, see Draper 2004.

13. Plantinga appears to reject the reasoning in Bayesian

arguments from evil, but many of the examples he uses to
challenge that reasoning involve alternative hypotheses that,
unlike HI, are far less plausible than the hypothesis to which
they are being compared. And the rest of his examples involve
hypotheses that, unlike theism, are obviously supported by
overwhelming propositional or non-propositional evidence.
Since a Bayesian argument from evil could be sound even if
there were such overwhelming counterevidence in support of
theism, I take the point of these examples to be to challenge
the signi cance instead of the soundness of Bayesian
arguments from evil.

14. See e.g. Sennett 1998. Sennett argues convincingly that, in

order for a class of beliefs (e.g. memory beliefs or beliefs about
God) to be directly justi ed, giving beliefs of that sort prima
facie credence in the appropriate circumstances must be
unavoidable, practically speaking. If this appealing criterion of
‘proper basicality’ is correct, then beliefs about God, unlike
memory beliefs, are not directly justi ed or warranted by non-
propositional evidence.

REFERENCES
ALSTON, W.(1991). ‘The Inductive Argument from Evil and the
Human Cognitive Condition’, Philosophical Perspectives 5: 29–67.

DRAPER, P. (1989). ‘Pain and Pleasure: An Evidential Problem for

Theists’, Noûs 23: 331–50.

——— (1991). ‘Evil and the Proper Basicality of Belief in God’, Faith
and Philosophy 8:135–47.

——— (1996). ‘The Skeptical Theist’, in Daniel Howard-Snyder

(ed.), The Evidential Argument from Evil. Bloomington: Indiana
University Press, 175–92.

——— (2004). ‘More Pain and Pleasure: A Reply to Otte’, in Peter

van Inwagen (ed.), Christian Faith and the Problem of Evil. Grand
Rapids: Eerdmans, 41–54.

——— (2007). ‘The Argument from Evil’, in Paul Copan and Chad
Meister (eds.), Philosophy

of Religion: Classic and Contemporary Issues. Oxford: Blackwell.

MACKIE, J.L. (1955). ‘Evil and Omnipotence’, Mind 64: 200–12.

OTTE, R. (2003). ‘Rowe’s Probabilistic Argument from Evil’, Faith and

Philosophy 19:147–71.

——— (2004). ‘Probability and Draper’s Evidential Argument from

Evil’, in Peter van Inwagen (ed.), Christian Faith and the Problem
of Evil. Grand Rapids: Eerdmans, 26–40.

PHILLIPS, D. Z. (2004). The Problem of Evil and the Problem of God.

London: SCM.
PIKE, N.(1963). ‘Hume on Evil’, Philosophical Review 72: 180–97.

PLANTINGA, A. (1974). The Nature of Necessity. Oxford: Clarendon.

——— (1996). ‘On Being Evidentially Challenged’, in Daniel

Howard-Snyder (ed.), The Evidential Argument from Evil.
Bloomington: Indiana University Press, 244–61.

——— (2000). Warranted Christian Belief. Oxford: Oxford University

Press.

——— (2007). ‘Objections to Draper’s Argument from Evil’ in Paul

Draper (ed.), God or Blind Nature? Philosophers Debate the
Evidence, <www.in dels.org/library/modern/debates/great-
debate.html>.

ROWE, W.L. (1979). ‘The Problem of Evil and Some Varieties of

Atheism’ American Philosophical Quarterly 16: 335–41.

——— (1986). ‘The Empirical Argument from Evil’ in R. Audi and

W. J. Wainwright (eds.), Rationality, Religious Belief, and Moral
Commitment. Ithaca: Cornell University Press.

——— (1996). ‘The Evidential Argument from Evil: A Second Look’,

in D. Howard-Snyder (ed.), The Evidential Argument From Evil.
Bloomington: Indiana University Press, 262–85.

SCHELLENBERG, J.L. (2000). ‘Stalemate and Strategy: Rethinking the

Evidential Argument from Evil, American Philosophical Quarterly
37: 405–19.
SENNETT, J.F. (1998). ‘Direct Justi cation and Universal Sanction’,
Journal of Philosophical Research 23: 257–87.

SWINBURNE, R. (2004). The Existence of God, 2nd edn. Oxford:

Clarendon.
 CHAPTER 16

THEODICY

MICHAEL J. MURRAY

INTRODUCTION

DURING a recent visit to Germany, Pope Benedict XVI visited the Nazi
death camp at Auschwitz. While surveying the memorial to the
nearly 1.5 million victims, he found himself at a loss for words of
explanation or consolation: ‘In this place, words fail. In the end,
there can only be dread silence—a silence which is itself a heartfelt
cry to God … How could you tolerate all this?’ Washington Post
columnist Richard Cohen commented as follows:

Religious people can wrestle with the Pope’s remarks. What does it
mean that God was silent? That he approved? That he liked what he
saw? That he didn’t give a damn? You tell me. And what does it
mean that he could ‘tolerate all this’? That the Nazis were OK by
him? That even the murder of Catholic clergy was no cause of
intercession? I am at a loss to explain this. I cannot believe in such a
God. (Cohen 2006: A10)
Theists and atheists alike seem to agree that evil does indeed count
against the existence of God. And it is not merely the vast quantities
of evil that stagger belief. In addition, it seems that the types and
distribution of evil t poorly with the claim that the world is
providentially directed by an all-good, all-powerful creator. While
we might admit that a comprehensive divine plan for the universe
could be expected to include some evil, how can we be expected to
accept that senseless torture, degrading sexual abuse, catastrophic
tsunamis, and so on are also components of such a plan? And what
is more, how can theists explain the fact that the distribution of evil
seems almost entirely random? Vice and su ering on the one hand,
and virtue and happiness on the other, do not seem proportional in
this life, a fact acknowledged by religious believers and atheists
alike. In the words of the Hebrew prophet Jeremiah: ‘You are
always righteous, O LORD, when I bring a case before you. Yet I
would speak with you about your justice: Why does the way of the
wicked prosper? Why do all the faithless live at ease?’ (Jer. 12: 3).
Needless to say, the pattern of evil we nd in the world does not
exactly t our initial expectations of what a world would look like if
theism were true.

Since the dawn of religious thought, religious believers have

wrestled with how to explain God’s permission of evil. During the
last three centuries, theistic attempts to explain evil have been
labeled ‘theodicies’. The word springs from a title of a book written
in 1710 by the German Lutheran philosopher Gottfried Leibniz
(1988). Leibniz coined the word ‘theodicy’ from two Greek words
which mean, respectively, ‘God’ and ‘justice’, and the book aims to
o er a comprehensive explanation of how it is that God’s justice can
be vindicated in the face of the apparent quantity, types, and
distribution of evil that we nd around us.

From Leibniz’s time until the mid-1970s, the word ‘theodicy’ was
used to describe attempts of this sort to explain God’s permission of
evil. Since the mid-1970s however, the word has taken on a more
re ned sense among philosophers of religion, a change that can be
attributed to Alvin Plantinga’s book God, Freedom and Evil (1974). In
this work, Plantinga distinguishes between two types of
explanations of evil that theists might construct. The rst type is
o ered in response to arguments that the coexistence of God and
evil is impossible. Explanations of this sort, which Plantinga calls
‘defenses’, need only show the logical compatibility of God and evil.
The second type aims to provide plausible and perhaps even likely-
to-be-true explanations of evil, explanations that show that the
existence of evil is not unlikely given the existence of God (or
perhaps given the existence of God and some additional plausible
and/or likely-to-be-true claims). Plantinga labeled explanations of
this latter sort ‘theodicies’ (ibid. 27–8). Plantinga’s distinction has
left a lasting mark on the eld and indeed, in the contemporary
literature, philosophers of religion use the term ‘theodicy’ in this
narrower sense, and it is in this sense that it will be addressed in
this chapter.
 RECENT DETERRENTS TO THEODICY

While philosophers and theologians have constructed and defended

explanations for God’s permission of evil through the centuries, two
recent developments in philosophy of religion have made the
enterprise of theodicy less appealing to some. First, some
philosophers have argued that there is something distasteful,
prideful, or arrogant about any attempt to explain how or why God
would permit evil. Those who nd it distasteful think that such
explanations serve either to blind us to the genuine horror that evil
represents, or perhaps to console us (or, worse, induce a Stoic
resolve) in a way that de ates our enthusiasm for opposing it (Tilley
1991).

These critics may be right that attempting to explain evil

sometimes has this e ect on us. However, it is also true that it
should not have this e ect. The reason for this is the (obvious) fact
that attempts to explain evil, even if they succeed, do not make evil
any less real or urgent. The patient whose hand is amputated due to
an incurable gangrenous infection might understand that the loss
was required in order to save his life. Yet explaining why such an
evil was unavoidable does not serve to diminish his su ering or his
loss. By attempting to explain evil, the theist is not, or ought not be,
attempting to ‘explain it away’ (Stump forthcoming).

Likewise, critics are right that explanations of evil are sometimes

o ered in a spirit of pride and arrogance. One need only consult the
news media after a tragic ood, famine, or epidemic to nd
religious believers who are quick to catalogue the sins of the victims
that are to blame. This sort of ponti cating often betrays a
disgraceful spirit of moral superiority and indi erence to human
su ering. Yet theists who shy away from any attempt to explain evil
are themselves forced to confront the fact that sacred writings in the
major theistic traditions often provide the very sort of explanation
these critics nd unpalatable. Although some sacred texts, the
Hebrew Bible’s book of Job for example, adopt a more skeptical
stance toward the human endeavor to understand evil, the very
same Hebrew Bible explains the plagues visited on Egypt in advance
of the Exodus as arising from Pharaoh’s ‘hardness of heart’.
Similarly, for Christians, the Passion of Christ is portrayed as an evil
necessary for bringing about the greater good of restoring humanity
to communion with God.

The second challenge arises from the increasing popularity of a

view with the unfortunate label of ‘skeptical theism’. Skeptical
theists note that explanations of God’s permission of evil typically
aim to show how certain types of evil are unavoidable preconditions
for, or by-products of, outweighing goods. As a result, insofar as we
want to explain the evil in our world, we must be well positioned to
know what goods God wants to bring about in the world, and how
the evils that we nd are connected with getting them. Therein lies
the problem. For according to skeptical theists, we labor under two
liabilities that make such explanations either very di cult or
impossible. First, given the immensity of divine goodness and the
 nitude of our human cognitive and moral faculties, it seems likely
that there are some (and perhaps many) types of good with which
we are not in any way acquainted. If we do not even know what the
relevant goods are that evils are supposed to secure, our attempts to
make judgments about whether or not God has good reason for
permitting evil are futile. To be clear, one can defend skeptical
theism without shying away from engaging in the task of theodicy.
Yet there is nonetheless a natural tendency to think that the stance
of the skeptical theist makes the prospects for a successful theodicy
look grim.

Second, even if we believe ourselves to be acquainted with the

relevant goods, there is good reason to think that we would be
unable to grasp whether or not particular evils play a role in
bringing about such goods. To know that they do, I would have to
see how the evils I encounter might be connected with outweighing
goods in ways that render the evil justi able. But how could I know
what sorts of ultimate good ends might be accomplished by the
permission of this or that token instance or type of evil? Some evils
might be necessary conditions for outweighing goods not realized
for hundreds or thousands of years. Absent omniscience, I simply
cannot have any good grasp of which events are necessary
conditions for which other events distant in time and space. This
should leave us with a healthy dose of uncertainty about our ability
to make judgments about the connections between evils and
outweighing goods.1
 There is nothing about skeptical theism that requires that we
abandon the task of theodicy. However, there are two reasons that
skeptical theism might serve to diminish our enthusiasm for it. First,
theists who are inclined to rely on skeptical theism as a way of
resisting the seemingly potent argument for atheism that arises from
evil might demur when it comes to the task of constructing
theodicies insofar as they think that the only reason to construct and
defend such explanations is to turn back this argument. Second, if
the skeptical theist is right, we are, for all we know, in a very poor
position to reconstruct the reasons that God might have for
permitting evil. As a result, attempts to piece together such
explanations might leave us either empty-handed or in serious error.
That realization too can be a deterrent to the task of theodicy.

However, the specter of skeptical theism should not have the

deterrent e ect it sometimes seems to have. As we saw earlier, evil
raises concerns for the theist that have nothing to do with the
argument for atheism from evil. Thus, even if skeptical theism
undermines that argument, explaining evil can still be a useful
enterprise. Further, though the considerations raised by skeptical
theists might lead us to doubt that we can explain God’s permission
of evil, theists will need to balance this against the fact that the
major theistic traditions are often quite forthright in o ering
explanations for evil. In this way, these theistic traditions are
committed to the idea that one can and does know some of the
actual reasons that God permits some evils.
 STANDARDS FOR A SUCCESSFUL THEODICY

God is justi ed in permitting an evil only if he has a morally

su cient reason for doing so, and theodicies are attempts to spell
out such reasons. According to most theists, a reason counts as
morally su cient when the evil permitted is connected with a
greater good in a way that meets the following three conditions:2

(A) The Necessity Condition: The good secured by the permission

of the evil, E, would not have been secured without
permitting either E or some other evils morally equivalent to
or worse than E.

(B) The Outweighing Condition: The good secured by the

permission of the evil is su ciently outweighing.

(C) The Rights Condition: It is within the rights of the one

permitting the evil to permit it at all.

Let’s consider these conditions in turn. No one thinks a surgeon is

evil in causing you to endure the pain and su ering of surgery when
that surgery is a necessary condition for saving your life. However,
you would be rightly indignant if a surgeon tells you, after the
surgery, that you could just as easily have been cured by taking a
pill that is equally e ective and has no side-e ects. If the pain and
su ering of surgery were not necessary to secure the good result
(saving your life), we would conclude that this surgeon is a greedy
moral monster. Likewise, if God permits an evil to occur, it must be
because the good that is in fact connected with it would not have
been secured at a ‘lower price’ (i.e. with less evil).

Note that the Necessity Condition requires that the goods in

question would not have been obtained without allowing either the
evil in question or some other evil that is as bad or worse. This is
important when thinking about theodicy because it means that evils
can be connected with outweighing goods in ways that will
sometimes be far from transparent. Consider this thought
experiment: imagine that all creatable worlds can be rank ordered
with respect to goodness. Further imagine that on this ordering, two
worlds tie for the title of best world. Finally, imagine that these
worlds are identical except that World 1 contains Smith dying in a
tragic auto accident, while World 2 contains Jones dying in a tragic
car accident. If God were to arbitrarily choose to create World 1,
those looking for an explanation for the evil of Smith’s death would
likely never imagine that the reason for it is that it (or an equally
bad evil) is a necessary by-product of creating the best world. But
this would indeed be the explanation. Because of considerations like
this, theists and atheists alike are forced to admit that there might
be evils that God permits for very good reasons, but where those
reasons will be opaque to those of us who aren’t omniscient.

Not all theists accept the Necessity Condition as a requirement

for theodicy. While critics of this condition cite various motivations
for their opposition, the most important criticism of it is found in
the work of defenders of open theism. Open theism is a cluster of
theological doctrines that includes the denial that God does or can
have detailed foreknowledge concerning certain types of future
events. Some open theists argue that such foreknowledge is
impossible because God is bound in the present time in a way that
precludes such knowledge. Others argue that it is impossible
because no being, God included, can have knowledge concerning
future events that are not determined by current events, including
free choices and perhaps certain natural events as well. As a result,
open theists argue that at least some future events are unpredictable
and, as a result, that there is nothing God can do to prevent their
particular occurrence (Hasker 2004).3 God permits some evil, then,
not because it is a foreseen and necessary condition for some greater
good, but rather because God ‘didn’t see it coming’. Because of this,
God could not, in creating, have any way to ensure or know that the
goods secured by permitting the evils the world contains would not
have been secured at the cost of less evil, and so the Necessity
Condition need not be met after all.

However, open theists do not adopt this unusual explanation for

evil merely because they take it to follow from certain ways of
thinking about divine foreknowledge or God’s relationship to time.
Rather, open theists argue that this explanation of evil is more
theologically palatable. Many of them nd it hard to believe that
God would create a world knowing that it would contain such evils
as the Holocaust or the tsumani disaster in Bande Aceh. But if God
lacks foreknowledge with respect to such evils, he could not have
prevented them and is thus not culpable for allowing them. Open
theists regard explaining some evils in this way as theologically
superior to explanations that suppose that God knew that such
events were going to occur, and simply chose not to intervene to
prevent them.

Open theists are right to claim that if God lacks foreknowledge

there will be some evils that God cannot prevent. However, it is not
obvious that this will solve the problems they take it to solve, for
three reasons. First, it is not at all clear that God’s inability to know
that certain events will occur renders God any less responsible for
their occurrence. On open theism, God does not, and cannot, know
that Adam would sin if created by God. But God knows that Adam
can sin. Indeed, on open theism, when God creates, he knowingly
leaves open the possibility that Adam will sin, that Holocausts (or
worse) will occur, and that tsunamis will wipe out hundreds of
thousands (or more). Since there is no good reason to think that God
will be able to calculate the probabilities of such evils occurring
before creating, God might often be utterly in the dark about the
likelihood that events like these will occur. This might leave God
blissfully ignorant of the actual evils that the future contains; but it
also casts God as objectionably reckless. If the evils that occur in our
world result from this sort of divine recklessness, God is hardly o
the hook for the evils that subsequently occur.

Second, since God cannot foresee the future, he leaves the door
open, when creating, to the possibility that that world will
ultimately turn out to be lled with free creatures who willfully
despise God and scorn their neighbors. As a result, open theists are
not entitled to claim—though they often do—that God can create
with su cient con dence that things will not ultimately come to
ruin. In addition, given God’s lack of foresight, there are two
possible outcomes that might arise in a world with free creatures,
either of which should provide su cient reason for God to refrain
from creating at all. First, if the world might come to ruin as
described, then it might come to have evils that are not outweighed
by the goods the world contains. Second, if God cannot foresee how
creatures will freely choose, those creatures might choose to do
things that it is not within God’s rights to permit. In either case, it
would be impermissible for God to risk creating at all.

Finally, even if it is true that there are some evils that God
cannot prevent because they cannot be foreseen, there are other
evils that God can foresee and yet still permits, even on open
theism. Because of this, the open theist’s explanation of evil has less
vindicating power than one might have initially thought. Elsewhere,
Michael Rea and I have expressed this worry as follows:

Why is it repugnant to think that God knew millions of years ago

that many would su er at the hands of Hitler but chose to permit
that su ering anyway, but not repugnant to think that God knew in
(say) 1941 that many (more) would su er at the hands of Hitler but
chose to permit that su ering anyway? In short, the Openist view,
like its rivals, is committed to acknowledging the existence of
preventable su ering that was both foreseen and permitted by God.
Given this, it is hard to see why the limitations on God’s foresight
imposed by the Openist view should be regarded as especially
advantageous. (Murray and Rea 2007)4

Open theism thus seems to provide less help in reconciling theism

and the existence of evil in the world than we might have initially
supposed.

The Outweighing Condition stipulates that the greater good in

question is ‘su ciently outweighing’. Why not rather stipulate that
the greater good merely be ‘outweighing’? The answer is that
gaining an extra measure of overall goodness might sometimes
require allowing a substantial additional quantity of evil, and that
substantial quantity might well make the overall greater balance of
good over evil not worth the price. This is important since it shows
us why explanations for evil should seek to explain not only how
evil is a necessary condition for some outweighing good, but also
why it would not be better to have neither the evil nor the good that
it spawns. For example, creating creatures with free choice may
yield a world in which the good that results ultimately outweighs
the moral evil. But given the vast quantity of permitted moral evil
required to get those goods, some might conclude that a world with
no free creatures is, all things considered, to be preferred.

The Rights Condition is necessary since there may be occasions

on which one being can permit evils in the service of securing
greater goods, but where the absence of a right to do so would
render such permission immoral. If a stranger wants to take the life
of my son in order to use his two kidneys to save the lives of her
two children, I have no right to give my child to this stranger, even
if we agree that saving two lives is better than saving one.5

Does the Rights Condition place any substantive constraints on

theodicy? Some philosophers have argued that it does. In particular,
many theists argue that God cannot properly permit an evil unless
the following condition is satis ed: when there are innocent victims
of evil, those evils must not only be necessary conditions for
outweighing goods generally, they must also be necessary conditions
for outweighing goods for the victim. Eleonore Stump (1985: 58)
puts the point this way, ‘if a good God allows evil, it can only be
because the evil in question produces a bene t for the su erer and
one that God could not produce without the su ering’.6 We might
characterize the underlying principle here as follows:

(P) For any innocent person S, God has the right to allow S to
su er only if there’s some outweighing good that S wouldn’t
enjoy if God didn’t allow S so to su er.

The motivation behind this requirement is similar to the motivation

behind Kant’s categorical imperative. If God were to allow someone
to be victimized by evil only for the sake of securing global goods or
goods for other persons, then the victim would have been used as a
mere means. But using persons merely as means is morally
impermissible for God or anyone else. As a result, an adequate
theodicy would have to show not only that the evils that occur are
necessary for outweighing goods, but also how the victims of the
evil are directly and properly compensated for their role in the
overall scheme.

What should we think of (P)? While it initially seems to be a

fairly non-controversial application of Kant’s principle, critics have
noted some crucial di culties. First, if (P) were true, it would seem
to provide us with a powerful argument against some basic
principles of ordinary morality. Ordinary morality dictates that
when I see an innocent su erer it is morally good for me to
intervene to stop the su ering. However, if (P) is true, in stopping
the su ering I am ultimately depriving the su erer of a greater
good. Thus, by intervening to help the victim, I ultimately cause her
harm. Any principle that wars with ordinary morality in this way
must surely be false (Jordan 2004).

Others have argued instead that the principle applies only under
conditions that make it inapplicable to God. For example, one
person (X) can justly allow another person (Y) to su er even when
the su ering does not yield a net bene t for Y, under the following
conditions: (a) X is in a position of lawful authority over Y and any
others who stand to gain or lose from Y’s su ering; (b) X is
responsible for the welfare of Y and these others; (c) the good to be
gained by allowing Y to su er substantially outweighs the su ering
experienced by Y; and (d) there is no other way to obtain the goods
produced by allowing Y’s su ering without permitting some
situation in which the overall balance of goods and evils is worse.
For example, the state might be permitted to quarantine a patient
with a virulent and incurable disease, perhaps without the bene t of
even being able to receive medical treatment, in order to protect
other citizens from contracting the disease. Since God has lawful
authority over us that exceeds the authority of the state, and since
God is responsible for the well-being of the entirety of his creation,
God too might be in a rightful position to allow there to be
uncompensated victims of su ering if that were necessary for
securing certain outweighing goods (van Inwagen 1995: 121).7

In addition to satisfying the three conditions set out above, there

are epistemic conditions that an explanation must satisfy to count as
a theodicy. Above we characterized a theodicy as an explanation
aimed at showing why the existence of evil is not unlikely given the
existence of God. In order to show this, a theodicy will have to
undermine the claim—a claim that seems obvious to us when we
rst think about evil—that the existence of evil just does seem
unlikely if indeed there is a God who is wholly good,
knowledgeable, and powerful. What does a theodicy need to look
like in order to play this role?

David Lewis claims that we need to develop ‘hypotheses that are

at least somewhat plausible, at least to the [theist]’ (Lewis 1993:
152) which show how evil does or can provide a necessary means
for certain goods. However, the standard of requiring ‘somewhat
plausible’ hypotheses is too high. We can see this by considering an
example. My colleague tells me he is going to Miami today. Later
today I get a call from a friend in the Midwest telling me that he
saw my colleague having dinner that evening in a swanky Detroit
restaurant. Should I conclude that my colleague lied to me? Not
necessarily. Maybe he took a ight that connected through Detroit
and, due to bad weather, was left stuck there for the night. Is that
explanation plausible? No. It’s possible. But note: I don’t really have
any reason to believe it at all. And, of course, I don’t have any
reason to disbelieve it either. Still, the explanation (a) would be a
good explanation if it’s true and (b) is not ruled out or even unlikely
given the other things I believe. Because of this, an explanation of
this sort is su cient to preserve my belief in the integrity of my
colleague even in the face of evidence that might be taken to show
that he is a liar.

Theodicies should be judged su cient when they meet a similar

standard. Even when the theist is unaware of plausible explanations
for evil, there might be a variety of reasons that are true for all she
knows, not rendered unlikely given the other things she believes,
and are such that if they were true, they would constitute good
explanations for evil (that is, they would be consistent with theism
and would explain why the types of evil in question would be
necessary for securing outweighing goods). The theodicies we will
consider in this chapter should and do aim to meet this standard.

The Punishment Theodicy

Christians, Jews, Muslims, theistic Hindus, and numerous other
theistic religious believers have a rmed that some evil is to be
explained as a result of divine punishment for human wrongdoing.

Successful theodicies must show that the evils they treat are
connected to securing outweighing goods. Thus in this case we
ought to ask: is it reasonable to think that evils of divine
punishment secure greater goods? Answering that question depends
on what punishment is supposed to be good for. Defenders of the
punishment theodicy have argued that punishment can be good for
one (or more) of four things: rehabilitation, deterrence, societal
protection, and retribution.

The rst three goods of punishment describe good consequences

for the wrongdoer or other human agents. In the case of
rehabilitation, the result is that the wrongdoer herself learns the
wrongness of her action and no longer performs the bad action. The
third-century Christian thinker Origen of Alexandria summarized
this bene t of punishment along such lines: ‘But our chastisement
by Providence has been that of children by a teacher or father. God
does not take vengeance, which is the requital of evil for evil, but he
chastises for the bene t of the chastised’ (Bettenson 1956: 257). The
goods of deterrence or societal protection instead bene t those around
the wrongdoer. In the case of deterrence, the punishment in icted
on the wrongdoer either acts as a disincentive for others tempted to
commit evil actions or leads others to reform their behavior.
Protection of society can be secured if the punishment makes the
wrongdoer unable to carry out further wrong acts. Such punishment
either brings about a physical or psychological disability or, in
capital punishment, the death of the wrongdoer.

The fourth and most controversial good of punishment is the

good of retribution. Many theistic traditions defend the notion that
when someone commits a wrong they merit a punishment which
exacts a cost that goes above and beyond mere recompense. If I steal
two hundred dollars from the bank and am caught, I am expected to
repay what I stole. But merely having to give back the money is not
enough. I committed a wrong, and this obligates me to do
something more than merely o er recompense—I must, for
example, pay a ne or spend time in jail. According to retributivists,
this additional cost is due simply as the penalty I have earned for
the wrong I have done. Exacting the additional penalty constitutes
retribution. St Thomas Aquinas argues that retribution is an
important aspect of divine punishment because it is a necessary
condition for maintaining justice in the universe: ‘Nevertheless the
order of justice belongs to the order of the universe; and this
requires that penalty should be dealt out to sinners. And so God is
the author of the evil which is penalty, but not of the evil which is
fault’ (St Thomas 1981: I q. 49 a. 2). Since divine punishment could
potentially serve to bring about any of these goods, some evil might
be justi ed as divine punishment.

Natural Consequence Theodicy

The Punishment Theodicy shows how some evil (the punishment)
can be an indirect result of morally bad choices. Sometimes,
however, bad moral choices lead to bad consequences directly. If I
choose to spend my life indulging my every desire, seeking out
sensual pleasure at every turn, and having no concern for the well-
being of others, I may end up fat, lazy, and alone. Those
consequences would be bad, but they are not the result of divine
punishments. Instead, these evils are the natural consequences of
immoral choosing. Leibniz put the point this way:

The most usual aim of punishment is amendment; but it is not the

sole aim, nor that which God always intends…Original sin, which
disposes men towards evil, is not merely a penalty for the rst sin; it
is a natural consequence thereof. …It is like drunkenness, which is a
penalty for excess in drinking and is at the same time a natural
consequence that easily leads to new sins. (Leibniz 1988: 200)

It is reasonable to think that a world designed by God would be one

in which choosing badly would sometimes turn out to be bad for us.
God might be able to use the bad consequences that arise from bad
choosing as a tool for helping us to learn how to live lives of moral
uprightness in loving communion with God and our fellow human
beings. Recognizing the poverty of a life lived in immorality and out
of communion with God and our fellows might be the only way of
moving us to change our ways freely. In this way, allowing
wrongdoing to have bad natural consequences brings about an
outweighing good. Peter van Inwagen defends this view as follows.

God’s plan of Atonement …is to make us dissatis ed with our state

of separation from Him; and not by miraculously altering our values
or by subjecting us to illusion or by causing us su ering that has no
natural connection with our separation, but simply by allowing us to
‘live with’ the natural consequences of this separation, and by
making it as di cult as possible for us to delude ourselves about the
kind of world we live in: a hideous world, much of whose
hideousness is quite plainly traceable to the inability of human
beings to govern themselves or to order their own lives. (van
Inwagen 1995: 110)

Punishment and Natural Consequence theodicies can only go so far,

however. First, it seems clear that many evils that occur cannot be
explained as divine punishment or natural consequences for moral
wrongdoing, most notably evils experienced by those who are
innocent, such as infants or non-human animals. Second, on the
punishment theodicy, the evils that result from punishment are
necessary in order to restore justice in the cosmic order or to
generate certain consequences for the wrongdoer or other human
agents. Such things might be good, but their goodness only makes
sense because there was some prior wrongdoing on the part of free
creatures. If those free creatures never did wrong, there would be no
need for punishment in the rst place. As a result, the punishment
theodicy needs to be supplemented with another theodicy that
explains why God would permit moral evil in the rst place.

The Free Will Theodicy

Philosophers addressing the topic of theodicy typically divide the

types of evil that our world contains into two broad categories:
moral evil and natural evil. Moral evil is evil that results from free
creatures using their freedom in morally blameworthy ways. Natural
evil is evil that does not directly involve blameworthy creaturely
action. The most commonly invoked theodicy for moral evil is the
Free Will Theodicy. Theists and atheists alike largely agree that
creating a universe would count as a good, and that it would further
count as good if the universe God creates were to contain creatures
capable of making free choices. Such creatures can enjoy the very
great good of making free and autonomous choices. Furthermore,
such creatures are capable of producing moral good in the world,
engaging in relationships of love and friendship, displaying genuine
charity and courage, and so on. Yet free creatures of that sort
necessarily have the ability to choose to do evil. And if those
creatures are genuinely free in making their choices, they cannot be
determined to choose only the good.

Now let’s imagine that God is faced with the prospect of creating
a universe. Wanting to ll the creation with the greatest types of
good, God decides to create a world containing a number of
creatures with free choice. Can God create a world with such freely
choosing creatures who never choose to do wrong? Maybe or maybe
not. We just don’t know. When God considers all the possible
universes he can create with free creatures in them, it just might
turn out that in every one of them, at least one of the creatures (or
perhaps even every one of the creatures) in them chooses to do
wrong. And if that’s right, then God cannot create a world that
contains the good of free choice but that also has no evil in it. In
this case, putting up with moral evil is the price God in fact must
pay in order to get a universe with the very great good of creatures
with free choice.8

This theodicy provides us with an explanation of why a universe

with free creatures might contain morally evil choices. However,
critics have pointed out that this does not, all by itself, explain why
those evil choices must issue in (further) evil results. There is, after
all, a distinction between the moral evil of making a morally bad
choice, and the moral evil that results from that bad choice. It might
be reasonable to hold that the reality of free choice makes it
inevitable that there will be some bad moral choices. But couldn’t
God allow free choices, bad choices among them, without allowing
those bad choices to have further bad consequences? It might be
good for me to have the ability to make a free choice to run out of
the restaurant before I pay my bill. But couldn’t God safeguard the
restaurant owner from harm by miraculously making money appear
on the table in an amount equal to the cost of my meal (with a
generous tip to boot)? It might be good for me to be able to choose
to run over your mailbox in anger. But couldn’t God make the
mailbox post magically elastic so that shortly after I drive away—
content that I smashed it—the mailbox springs back upright without
a scratch?

Some philosophers have argued that while it is a good thing to

allow creatures to have free choice, it is a bad thing to allow those
free choices to cause harm or injury to others. Instead, they argue,
God should put us all in a virtual ‘playpen’ in which choices can be
made without any real harm to others being caused. Good choices
could be made, and the good consequences that follow from them
allowed. But bad choices, while not prevented altogether, would be
prevented from causing additional damage. Couldn’t God simply
block such negative consequences? (Lewis 1993).

The rst question to ask here is: What would it look like if God
were to do so? I intend to pull the trigger to shoot you but suddenly
nd that my nger is paralyzed, or… I intend to steal the car but
when I rear back to throw the brick at the car window to gain entry,
I suddenly fall asleep. Would this su ce? In one sense it would.
Were the world so con gured, I would not be able to bring about
any evil beyond my evil choices. Unfortunately, the result of such an
arrangement is that, before long, I would not be able to make evil
choices either. The reason for this is that my experience will make it
clear to me that doing evil is in fact impossible. We can see this by
considering an example. When I was 5 years old, I and a few friends
decided that we would jump from a concrete wall and y. After
twenty minutes of consistent failure and sore fannies, we could see
that ying was not in our future. To the best of my knowledge, no
one in that group has since been tempted to y o a concrete wall.
Indeed, I suspect none of my kindergarten companions could now
even form the intention to y o the wall. They know by their
experience that doing so is as impossible as leaping to the moon or
swallowing the ocean. If God were simply to block the evil
consequences as the virtual playpen scenario advocates, choosing
evil would be no more possible for us than choosing to y is for us
in the actual world.9

As a result, for there to be a playpen world that preserves our

ability to choose and intend to do evil, God would need to block the
consequences of our evil choices while at the same time giving us the
illusion that the evil action we intended was indeed carried out.
There is real moral evil in such a world, since the choice itself is a
morally evil one. But there is less overall evil since the evil
consequences are never realized and, perhaps best of all, no one
except the evildoer is ever harmed by those evil choices. Wouldn’t
this souped-up virtual playpen be preferable to the actual world?

There are good reasons to think that the playpen so described is

impossible, as I will show below. However, if the critic thinks that
such a world were possible, that critic would have to confront this
question: how do we know that we are not in fact in such a world?
If we are, the prospect of such a playpen does not count as an
objection to the Free Will theodicy but rather an articulation of it!
Leaving that possibility aside, one might object that the playpen
world casts God objectionably in the role of a deceiver since he
would be leading the murderer or the car thief to believe that they
have murdered someone or stolen a car, when in fact they have
merely elected to do so (and been subjected to an illusion that they
have actually done so). But if such deception is the only way to
allow the possibility of evil-choosing-freedom without allowing
harm to others, perhaps the deception is itself an evil necessary for
this outweighing good.

There are good reasons, however, to think that virtual playpens

would quickly degenerate into semi-solipsistic, causally isolated
worlds. This is the case because it would be impossible for free
creatures to remain in genuine contact with one another after an
extended period of making evil choices with these illusory
consequences. Were I to be in the playpen, I would be unable to
apologize to someone for harming them, or to receive apologies
from those who (think that they) harmed me. I would be unable to
see someone wearing a watch or driving a car that I think I stole
from them. And so on. In the playpen I would have to become
increasingly blind to the way the world really is. In order to keep all
these illusions working, I would be withdrawn deeper and deeper
into a solipsistic world until I became disconnected from the real
world entirely. But such an arrangement also serves to eliminate the
possibility of engaging in relationships of love and friendship with
others, thus undermining what is perhaps the central good that
freedom makes possible.

Thus, even if playpen worlds were free of these liabilities, there

are good reasons to think such a world would be worse than the sort
of world we actually nd ourselves in. While such a world would be
better in the sense that the evil consequences of our immoral
choices would be thwarted, something else is missing from such a
world—and it is part, perhaps the largest part, of what makes
freedom valuable in the rst place. What is good about free choice
is not simply that I have the ability to choose among a list of items
on a mental menu. Rather, what is good is that I have the ability to
exercise my powers in my environment by means of my choices. I
have the capacity to choose to show love to my spouse or my child
and to do it.I have the ability to choose to sacri ce for the sake of my
friend, and to do it. Being able to choose is a great good. Being able
to e ect states of a airs in the world in virtue of our choices is even
better.

The Free Will Theodicy thus seems to provide at least a possible

explanation for the fact that God allows a world in which creatures
can make evil moral choices and in which those choices can
sometimes issue in bad consequences. But like Punishment
Theodicies, Free Will Theodicies are not comprehensive. Even if
these considerations su ced to explain moral evil, they don’t seem
to o er us any explanations for natural evil.
The Natural Law Theodicy

Free Will Theodicies focus on explaining evil as a consequence or

result of creaturely free choice. Yet this is not the only way that
freedom might be connected with the permission of evil. A number
of conditions must be in place for free creatures to be able to
exercise their freedom. We have, in fact, seen an example of this
already. Above I told the story of my youthful attempts at ying.
Those attempts failed, of course, and as a result I no longer attempt
to y, and, in fact, I doubt I could even form the intention to y
because of those experiences. Because the world would not
cooperate with my intentions back then, I am no longer free to try
to act in that way.

What this shows us is that in order for us to be free in a certain

domain of action, the world needs to cooperate with our choice-
making. If the world proves wholly unresponsive to certain choices I
make (to y unaided, for example), I will lose the freedom to choose
such things. What is required for the world to ‘cooperate’ with my
choices in the needed way? At least one thing that is required is that
the environment around me be governed by regular, orderly laws of
nature. The reason for this is that in most cases, I act by moving my
body in certain ways and, in turn, a ecting the world in certain
ways. When I want to tell my daughter a story, for example, I move
my mouth in ways that make sounds that communicate a story, and
so on. If the environment around me were not regular and lawlike, I
couldn’t do such things, since I wouldn’t know how to intend to
bring them about.

We can see this by considering a chaotic world in which there

are no stable and reliable laws of nature. In this world, vibrating my
vocal cords would sometimes cause air to move in such a way that I
make sounds, and sometimes it wouldn’t. If the world were chaotic
in this way, I would never come to discover a connection between
vibrating my vocal cords and making sounds that allow me to tell a
story. If I had a desire to tell a story, I wouldn’t have any idea how
to go about doing it, since the chaotic world has failed to cooperate
in ways that allow me to do it reliably. As a result, it seems that I
could have desires to do all sorts of things, but that I would never
actually choose to do them because I would have no idea how to do
so (Mawson 2004; Reichenbach 1982: 103–4).

Any world in which there are going to be free creatures capable

of carrying out free actions with consequences beyond their own
skin must then be a world that operates according to regular,
orderly laws of nature. And this can lead to problems. The very
same laws of momentum that allow me to use a hammer to drive a
nail, can cause that hammer to smash my thumb. The very same
laws that allow me to tell stories by vibrating air with my vocal
cords, allow tornadoes to knock down houses. And so on. In a world
governed by regular, orderly natural laws, it is possible for these
laws to conspire to intersect with the interests of creatures to cause
them harm. When they do so, natural evil will be the result.
 There are two serious objections that Natural Law Theodicies
need to confront. First, one might wonder why God did not create a
world with better laws that yield less natural evil. After all, would
the world have been any worse if the laws were set up that did not
allow viruses to occur? Second, aren’t there plenty of cases of
natural evil that could be eliminated without undermining the
regularity of the laws of nature to such an extent that our freedom
would be disabled? Would preventing one major hurricane
undermine the possibility of my exercising my free will? If not,
shouldn’t God prevent one or two (or ten) more hurricanes? Let’s
consider these in turn.

Could the laws of nature have been con gured to yield a world
that has a substantially better overall balance of good than our
world? To show that such a better world is possible, we would need
to describe a regular, lawlike world that (a) contains goodness of
the sorts (either the same sorts or equivalent or better sorts) and
amounts found in the actual world and which (b) contains
substantially less natural evil than the actual world. There are two
problems with trying to o er such a description. First, there is good
reason to think that there is not much wiggle room in the way the
laws and constants of the world are structured. One fairly recent
discovery of scientists is that the cosmos seems to be balanced on a
razor’s edge in such a way that were the laws and constants that
govern its activity slightly di erent, the cosmos would be unable to
sustain intelligent life (Barr 2003: 118–38). This provides us with
some good reason to suppose that if the universe is going to be
capable of supporting life, it will have to be governed by laws and
constants similar to those we nd in the actual world.10 Second,
even if a better set of laws could be speci ed, it is doubtful that we
could know this. Knowing such a thing would require knowing how
changes we propose to certain laws and constants would impact not
only the natural evils we are trying to prevent, but other laws of
nature and the goods and evils that arise from their mutual
interactions. It is unreasonable to think we could unscramble such
things and thus unreasonable to think that the laws might be
changed to yield a better world with less natural evil.

The second objection is more formidable. If lawlike regularity in

the world exists for the purpose of allowing free creatures to use
their free choice, then any natural evils that could be eliminated
without eliminating that good result would be gratuitous. And yet it
seems that there are many such evils. Even if God could not prevent
all such evils by systematically altering the laws that hold in our
universe, he could at least eliminate many by miraculous
intervention. The evils of kidney stones or ingrown toenails seem
good candidates for such elimination.

The theist might respond that God already does miraculously

intervene to prevent some such evils. This answer is not su cient,
however, since the critic wants to know why even more such evils
are not prevented. The only answer available to the theist is that
natural evils serve as necessary conditions for some good ends. One
of those is the good of allowing free creatures to exercise their
freedom. But there must be others. Thus, while the regularity and
lawlikeness that allows for natural evil serves to make the good of
free choice possible, it makes other goods available as well, and it is
those goods that explain why some of the apparently gratuitous
natural evils we nd around us are not gratuitous after all. What
this shows us is that the Natural Law Theodicy will not be able to
provide a comprehensive explanation for natural evil.

Soul-Making Theodicies

The theodicies considered so far all regard evil as a consequence of

free choice or of necessary preconditions for free choice. Other
theodicies treat evil as a necessary condition for goods of other
sorts. For example, many theistic traditions regard the earthly life as
an arena in which people make choices for the purpose of
cultivating moral and spiritual growth. If the world were lled with
perpetual pleasure and satisfaction we would never experience the
growth that can come only from real su ering, hardship, and defeat.
As a result, some theists propose that God allows for evil in the
world to allow us to cultivate virtues of outweighing goodness that
could not otherwise be cultivated.

It is easy to think of some such virtues. I could not become

charitable unless there were cases of someone in need. I could not
become courageous unless there were real dangers to be confronted.
And so on. More than that, I could not become a lover or a friend
unless the possibility of losing that love or friendship made it
possible for both me and the beloved or the friend to lose something
of great value. All these cases highlight the fact that one of the
important goods in our world is that it provides an arena for soul-
making or character building. And this important good requires that
the universe contain some evil.

This theodicy, pioneered by the second-century Christian thinker

Irenaeus (Irenaeus 2001) and resurrected in the twentieth century
by philosopher John Hick, stipulates that four conditions must be in
place for soul-making to occur. First, there must be creatures
capable of choosing between good and evil. Second, those creatures
must be placed in an environment that allows free choices to be
carried out. Third, the environment must contain challenges to one’s
character of a sort that allows for both virtuous and vicious
responses. And nally, creatures must have su cient opportunities
to respond to make character building possible (Hick 1977).

Soul-making theodicies must confront a couple of important

objections, the more important of which is that many sorts of moral
and spiritual growth envisioned in soul-making scenarios require
only that there be apparent evils in the world. For me to develop the
virtues of charity and courage there need not be any actual need or
actual peril, it only needs to seem to me that there is. Once again,
such a world as the virtual playpen described earlier would seem to
be su cient; as long as it seems to me as if I lived in a world lled
with evil, soul-making can occur, and without the cost of my having
to confront any actual su ering or danger. Wouldn’t such a world be
preferable? Some theists have argued that it would not. Daniel-
Howard Snyder thus argues as follows:

However, if God were to set up a world in which there was only

illusory evil to which we could respond in the formation of our
character, something of immense value would be missing. No one
would in fact help anybody else; and no one would be helped. No
one would in fact be compassionate and sympathetic to another;
and no one would receive compassion and sympathy…. No one
would in fact generously give of their time, their talents or their
money to the poor; and no one would receive generosity from
another. In short, if every opportunity for a virtuous response were
directed at illusory evils, each of us would live in our own little
‘world,’ worlds devoid of any genuine interaction and personal
relationships.

(Howard-Snyder 1999: 99)

While some evils might be avoided in such a world, the cost of
avoiding them would be the elimination of many of those aspects of
the world that we take to be the most valuable: our interactions of
love and friendship with others.

Theodicies of Animal Su ering

Some philosophers have recently argued that one distinctive sort of
evil is especially troublesome and di cult to explain: non-human
animal su ering (or ‘animal su ering’ for short). Animal su ering is
problematic not only because there are more animals than humans,
and because animals have been around for so much longer than
humans, but because, if Darwinian naturalism is right, animal
su ering, pain, and death, seem to be among the very instruments
of creation. In light of the evolutionary carnage, philosopher Philip
Kitcher (2005: 268) nds it incredible that theists can sustain belief
in an all-wise, benevolent creator:

[Were we to imagine] a human analogue peering down over a

miniaturized version of this arrangement—peering down over his
creation—it’s hard to equip the face with a kindly expression.
Conversely, it’s natural to adapt Alfonso X’s famous remark about
the convolutions of Ptolemaic astronomy: had a benevolent creator
proposed to use evolution under natural selection as a means for
attaining his purposes, we could have given him useful advice.

Animal su ering is all the more vexing when we realize that the
theodicies considered so far do not o er much help in explaining it.
Even if some of the explanations take us some way towards
explaining the reality of animal su ering (some animal su ering is
caused by the evil free choices of humans, for example), most of it is
still entirely unexplained. In general, there are two facts about
animal su ering that seem to make it especially resistant to
explanation using traditional theodicies. First, most animal su ering
predates the advent of human beings. As a result, any theodicy that
aims to explain evil as a consequence or precondition of human free
choice, soul-making, and the like is going to struggle to explain
animal su ering. Second, there does not seem to be any reason on
the earlier described theodicies that God would have to create
animals capable of experiencing pain and su ering in the rst place.
If a world with natural regularities is one in which organisms might
sometimes become innocent victims of the forces of nature, it might
be better, all things considered, not to out t non-human animals
with the tools that allow them to experience pain and su ering in
the rst place.

The problem raised by animal su ering was of some concern to

philosophers and theologians in the late nineteenth century, but
interest in the topic seems to have waned from that time until quite
recently. If there is to be a successful theodicy of animal su ering, it
will have to follow one of the following outlines.

First, some theists have tried to account for the reality of animal
pain and su ering not as a result of the misuse of human free choice
but rather as a result of the misuse of the free choice of non-human
free creatures. There is a long tradition in western theism of
supposing the existence of creatures (demons, for example) with
free choice whose existence predates human existence. If such
beings did or do exist, they might perhaps have had powers
su cient to steer evolutionary history in a way that allowed for
non-human animals to have the capacity to experience pain and
su ering, as well as to in ict actual pain and su ering on them
(Boyd 2001).

Second, some theodicies claim that, for all we know, animals do

not have the necessary mental equipment to experience pain and
su ering, or to experience it in a way that is morally relevant.
Descartes, for example, argued that animals were merely complex
machines displaying behaviors that make it appear that they are in
pain. Some recent philosophers have argued that pain and su ering
are possible only for organisms that are able to have a ‘second-
order’ awareness of their rst-order mental states. Human beings
have such advanced mental capacities. However, second-order
awareness requires a good deal of neural and cognitive complexity,
and it is not clear that animals have what is needed to experience
such second-order awareness (Murray and Ross 2006).

Third, some theodicies claim that nervous systems that produce

pain and su ering are necessary for embodied organisms to be able
to avoid harm and preserve their bodily integrity. While there are
many ways in which our nervous systems could signal that our
bodily integrity is in jeopardy, there is some experimental evidence
that only signals that are ‘qualitatively unpleasant’ will cause
organisms to behave in ways that are necessary to avoid injury
(Eccles 1980; Murray 2008: ch. 4).

Finally, some theodicies of animal su ering contend that animals

with minds capable of experiencing pain and su ering are necessary
evolutionary precursors for organisms with the special cognitive
skills of human beings—skills that, for example, make possible free
choice and moral responsibility. Of course, explanations of this sort
seem to make animals mere by-products on the evolutionary step-
ladder leading to human beings, making one wonder whether there
is some good explanation for why God had to employ such
precursors in the rst place. One answer might be found in the
claim, defended most famously by the French philosopher Nicholas
Malebranche, that a world is better insofar as the complexity in that
world results from the regular, orderly operation of divinely
ordained laws of nature (Malebranche 1958–84: ix. 1085). If one
nds such a position plausible, one might argue that the only way to
get such beings as us into creation via lawlike processes is to allow
our universe to unfold in much the way that it did: evolving from an
initial singularity to later stages in which there are highly
concentrated pockets of privileged order (such as is found on our
own planet, and perhaps others as well). Animals would then be
necessary stages in the development of this lawlike universe, leading
to such organisms as us as well as perhaps organisms of even greater
complexity and worth in the future.

NOTES

1. For a robust defense of skeptical theism see Michael Bergmann

2001: 278–96, as well as his ‘Skeptical Theism and the
 Problem of Evil’, Ch. 17 in this volume. For a recent critique of
skeptical theism see Almeida and Oppy 2003.

2. Adapted from Swinburne 1998.

3. God could, of course, prevent any token instance of an

unpredictable event by simply annihilating the entirety of
creation. The point here is that there is nothing God could do
to prevent speci c unpredictable events from occurring while
leaving the created order intact.

4. A similar point has been made by Edward Wierenga (1997).

5. The example is a modi cation of one used by Swinburne

(1998: 11). Of course one might not agree that saving the lives
of two persons through killing a third, innocent person is an
outweighing good. But none of this matters for the general
point that there are conditions where one is not morally
permitted to allow the occurrence of preventable evils the
permission of which is necessary to secure outweighing goods.

6. Emphasis mine. Similar views have been defended by Adams,

1999: 29–31;William Rowe, ‘The Empirical Argument from
Evil’, in Audi and Wainwright (eds.) 1986: 244–5; Eleonore
Stump, ‘Providence and the Problem of Evil’, in Flint (ed.)
1990: 65–8; and Michael Tooley, ‘The Argument from Evil’, in
Tomberlin (ed.) 1991: 110–11.

7. Finally, Sorensen and Boorse 1988 argue that the principle is

false on other grounds. Consider the following: a boulder is
 bouncing down an embankment towards me. If I duck, the
boulder will miss me, but hit you (who happens,
unfortunately, to be standing right behind me). If I do not
duck, I will be crushed and you will be spared. Surely, they
claim, it is morally permissible for me to duck. But in that
case, (B) is false. Some might object, of course, that authentic
neighbor love entails that one not duck after all. If so, this
objection will miss the mark.

8. Such an explanation for evil is worked out in detail as a

defense (rather than a theodicy) in Plantinga 1972.

9. Both volume editors have remarked that in such a world one

could nonetheless continue to fantasize about visiting harm on
others, or at least hope that harm would be visited on them.
While neither of these involve full edged intentional action,
they are still reasonably seen as being produced or sustained
by our free choices. I take the rst to be unlikely since such an
environment lacks necessary conditions for allowing me to
form the belief that I can in fact cause such a thing as harm in
the rst place. Nothing on the account I have described here
would preclude the second. However, it is not clear that mere
wishing ill on others counts as free or voluntary.

10. One might object that, while it might not be possible to have a
world that sustains life and preserves the other’s goods
dependent on lawlike regularity without allowing the
possibility of natural evil that causes injury, injury and injury
 avoidance need not involve the qualitatively unpleasant
sensations of pain. This objection is addressed in detail in
Murray 2008: ch.4.

REFERENCES

ADAMS, MARILYN MCCORD (1999). Horrendous Evils and the Goodness of

God. Ithaca: Cornell University Press.

ALMEIDA, MICHAEL, and OPPY, GRAHAM (2003). ‘Skeptical Theism and

Evidential Arguments from Evil, Australasian Journal of
Philosophy 81: 496–516.

AUDI, ROBERT, and WAINWRIGHT, WILLIAM J. (eds.) 1986. Rationality,

Religious Belief, and Moral Commitment. Ithaca: Cornell University
Press.

BARR, STEPHEN M. (2003). Modern Physics and Ancient Faith. Notre

Dame: Notre Dame University Press.

BERGMANN, MICHAEL (2001). ‘Skeptical Theism and Rowe’s New

Evidential Argument from Evil’, Nous 35: 278–96.

BETTENSON H. (1956). The Early Christian Fathers. Oxford: Oxford

University Press.

BOYD, GREGORY A. (2001). Satan and the Problem of Evil: Constructing

a Trinitarian Warfare Theodicy. Downers Grove: Inter-Varsity
Press.
COHEN, RICHARD (2006). ‘Whose Silence’, Lancaster Intelligencer
Journal, 6 June 2006, A10.

ECCLES, SIR JOHN (1990). The Human Psyche. New York: Springer-
Verlag.

FLINT, T. (ed.) (1990). Christian Philosophy. Notre Dame: University

of Notre Dame Press.

HASKER, WILLIAM (2004). Providence, Evil, and the Openness of God.

New York: Routledge.

HICK, JOHN (1977). Evil and the God of Love. New York: Macmillan.

HOWARD-SNYDER, DANIEL (1999). ‘God, Evil, and Su ering’, in Michael

Murray (ed.), Reason for the Hope Within. Grand Rapids:
Eerdmans.

IRENAEUS (2001). ‘Against Heresies’, in The Apostolic Fathers with

Justin Martyr and Irenaeus, The Ante-Nicene Fathers. Grand
Rapids: Eerdmans.

JORDAN, JEFF (2004). ‘Divine Love and Human Su ering’,

International Journal for Philosophy of Religion 56: 169–78.

KITCHER, PHILIP (2005). ‘The Many-Sided Con ict Between Science

and Religion’, in William E. Mann (ed.) The Blackwell Guide to
the Philosophy of Religion. Malden: Black-well.

LEIBNIZ, GOTTFRIED (1988). Theodicy, trans. E. M. Huggard. La Salle:

Open Court.
LEWIS, DAVID (1993). ‘Evil for Freedom’s Sake’, Philosophical Papers
22:152.

MALEBRANCHE (1958–84). ‘Abregé du traité de la nature et de la

grace’, in A. Robinet (ed.), Œuvres complètes de Malebranche, 20
vols. Paris: J. Vrin.

MAWSON, T.J.(2004). ‘The Possibility of a Free-Will Defense for the

Problem of Natural Evil’, Religious Studies 40: 23–42.

MURRAY, MICHAEL J. (2008). Nature Red in Tooth and Claw. Oxford:

Oxford University Press.

———and ROSS, GLENN (2006). ‘Neo-Cartesian Theodicies of Animal

Su ering’, Faith and Philosophy 23: 169–90.

———and REA, MICHAEL (2007). Introduction to Philosophy of

Religion. Cambridge: Cambridge University Press.

PLANTINGA, ALVIN (1974). God, Freedom and Evil. Grand Rapids:

Eerdmans.

REICHENBACH, BRUCE (1982). Evil and a Good God. New York:

Fordham University Press.

SORENSON, ROY, and BOORSE, CHRISTOPHER (1988). ‘Ducking Harm’,

Journal of Philosophy 85: 115–34.

STUMP, ELEONORE (1985). ‘The Problem of Evil’, Faith and Philosophy

2.

———(forthcoming), Wandering in the Darkness.

SWINBURNE, RICHARD (1998). Providence and the Problem of Evil.
Oxford: Oxford University Press.

THOMAS, AQUINAS, St (1981). Summa theologiae. Wheaton: Christian

Classics.

TILLEY, TERRENCE (1991). The Evils of Theodicy. Washington:

Georgetown University Press.

TOMBERLIN, J. (ed.) (1991). Philosophical Perspectives 5. Philosophy of

Religion. Atascadero: Ridgeview.

WIERENGA, EDWARD (1997). ‘Review of: The Openness of God by Clark

Pinnock et al.’, Faith and Philosophy 14: 248–52.
 CHAPTER 17

SKEPTICAL THEISM AND THE PROBLEM OF

EVIL

MICHAEL BERGMANN

IN philosophy of religion, some moves—like appealing to evil to

support atheism or appealing to the appearance of design in nature
to support theism—are very natural. They occur easily to non-
philosophers in their re ective moments. Others (e.g. the
ontological argument) are moves only a philosopher would think of.
The skeptical theist’s signature move is a very natural one. To see
this, consider what would happen if you were addressing some
freshman college students in an introductory philosophy class and
you presented them with the following argument from evil:

(1) There are some evils that are such that humans can’t think of
any God-justifying reason for permitting them.1

(2) So probably there aren’t any God-justifying reasons for

permitting those evils.

(3) If God existed, he wouldn’t permit these evils if there were no

God-justifying reason for permitting them.

(4) Therefore, probably God does not exist.

If you asked them what they think of the argument, it’s almost a
certainty that someone in the class would point out that the
inference from (1) to (2) doesn’t seem persuasive: the fact that
humans can’t think of any God-justifying reason for permitting an
evil, doesn’t make it likely that there are no such reasons; this is
because if God existed, God’s mind would be far greater than our
minds, so it wouldn’t be surprising if God had reasons we weren’t
able to think of. This very natural sort of response is precisely the
move the skeptical theist is known for.

Some say the term ‘skeptical theism’ is a bad name for the view
under consideration here. The main complaint is that one needn’t be
a theist to object to the above argument in the way skeptical theists
do.2I agree that one needn’t be a theist to object to the above
argument in the way skeptical theists do. But I don’t think that
makes ‘skeptical theism’ a bad name for the view. Skeptical theism
has both a skeptical component and a theistic component. The
theistic component is just theism, the view that there exists an
omnipotent, omniscient, wholly good, eternal person—i.e. a perfect
being of the sort endorsed by the western monotheisms. The
skeptical component advocates skepticism about the realm of
potentially God-justifying reasons—a degree of skepticism that leads
to a denial of the cogency of such noseeum inferences as the one
above from (1) to (2). (Wykstra 1996: 126 calls this a ‘noseeum
inference’ because it says that since we don’t see ’um, they probably
ain’t there.) And although non-theists won’t endorse skeptical
theism given its theistic component, many think that non-theists
should—and some do—endorse its skeptical component, which is
why they can agree with the objection in the previous paragraph.
Moreover, it makes perfect sense that those who rst made popular
this sort of move in response to the above argument from evil were
called ‘skeptical theists’: they were, after all, theists; and their
advocacy of skepticism about certain matters relevant to God’s ways
was a striking feature of their view. It was only natural, then, to call
the view they espoused ‘skeptical theism’.

For our purposes here, what’s most interesting about skeptical

theism is its skeptical component. When skeptical theists use that
skeptical component in responding to arguments from evil, they
think it is reasonable for their non-theistic interlocutors to accept
that skeptical component, even if they don’t expect them to accept
their theism. It is that skeptical component that will be the focus of
this chapter. In the rst section, I will explain more precisely what
the skeptical theist’s skepticism amounts to and how it is used in
response to various sorts of arguments from evil. Then, in sect. II, I
will consider and respond to objections to skeptical theism. One
thing we’ll nd is that just as there are non-theists who accept the
skeptical theist’s skepticism, so also there are theists who reject it.

I. THE SKEPTICAL THEIST’S SKEPTICISM

The skeptical theist’s skepticism applies to the realm of God-

justifying reasons. (What exactly is required for something to count
as a God-justifying reason? Here’s a very natural proposal that can
serve as a rst approximation: a good state of a airs G—which
might just be the prevention of some bad state of a airs E*—is a
God-justifying reason for permitting an evil E if and only if (1) G’s
goodness outweighs E’s badness and (2) G couldn’t be obtained
without permitting E or something as bad or worse.3) Skepticism
about the realm of God-justifying reasons leads many theists as well
as some non-theists to reject noseeum inferences such as the one
from (1) to (2), which Howard-Snyder (1996a: 291) calls ‘the
inference from inscrutable to pointless evil’. In this section, I want
to address three questions: what exactly is involved in this
skepticism? What motivates it? To which arguments from evil does
the skeptical theist’s skepticism apply?

A. What’s Involved in the Skeptical Theist’s Skepticism and

What Motivates It?

The skeptical theist’s skepticism4 is, I believe, best explained as an

endorsement of some skeptical theses, among which these three are
prominent:

(ST1) We have no good reason for thinking that the possible

goods we know of are representative of the possible goods
there are.

(ST2) We have no good reason for thinking that the possible evils
we know of are representative of the possible evils there
are.
 (ST3) We have no good reason for thinking that the entailment
relations we know of between possible goods and the
permission of possible evils are representative of the
entailment relations there are between possible goods and
the permission of possible evils.5

Three clari catory remarks are in order.

First, as William Rowe (1996: 264) emphasizes, possible goods

are abstracta—good states of a airs that could obtain. Thus, if we
set aside concerns about God being a necessary being if he exists at
all, atheists can agree that the beati c vision is a possible good,
despite the fact that they think it isn’t an actual good since it entails
God’s existence. Likewise, possible evils are bad states of a airs that
could obtain. And, of course, among the possible goods and evils are
actual goods and evils as well as merely possible ones.

Second, one might wonder why there is a focus in (ST1)–(ST3)

on possible goods and evils instead of on actual goods and evils (i.e.
possible good and evil states of a airs that obtain). In the case of
evils, this isn’t di cult to understand. God might permit an evil E in
order to prevent a worse evil E* which will obtain if E isn’t
permitted. Here there is clearly no need for E* to be an actual evil.
As for goods, it’s true that in order for God’s aim of obtaining G to
be of use in actually justifying his permission of E, G must eventually
be actual. But it doesn’t need to be currently actual. It may currently
be merely possible and it may become actual only as a result of
permitting E, perhaps long after E is permitted. Moreover, if one’s
goal is simply to respond to arguments from evil such as the one
mentioned at the beginning of this chapter (a goal an agnostic and
theist might share), there’s no need to defend the claim that God
does exist or that there in fact is a God-justifying reason for
permitting the evils mentioned. It’s enough that we lack any good
reason or justifying ground for thinking it’s likely that there isn’t
such a God-justifying reason. Hence, considerations having to do
with possible goods that we have no good reason to think are
unlikely to be actual (now or in the future) are relevant in
addressing such arguments—even if those goods are in fact merely
possible.

Third, a sample of Xs can be representative of all Xs relative to

one property but not another. For example, a sample of humans can
be representative of all humans relative to the property of having a
lung while at the same time not being representative of all humans
relative to the property of being a Russian. To say a sample of Xs is
representative of all Xs relative to a property F is just to say that if
n/m of the Xs in the sample have property F, then approximately
n/m of all Xs have F. In (ST1)–(ST3), what we are interested in is
whether our sample of possible goods, possible evils, and entailment
relations between them (i.e. the possible goods, evils, and relevant
entailments we know of) is representative of all possible goods,
possible evils, and entailment relations there are relative to the
property of guring in a (potentially) God-justifying reason for
permitting the inscrutable evils we see around us.6 Although that
property is not explicitly mentioned in (ST1)–(ST3), it is
representativeness relative to that property that (ST1)–(ST3) are
speaking of.

Thus, the skeptical theist’s skepticism a rms certain limitations

to our knowledge with respect to the realms of value and modality.
The claim isn’t that we know nothing about those realms. I can
confess to being in the dark about which of two proposed courses of
action will have the best overall consequences without thereby
admitting complete skepticism about value. And I can confess that I
don’t know whether simple mathematical truths entail Goldbach’s
conjecture without admitting to complete modal ignorance.
Likewise, endorsing the limitations mentioned in (ST1)–(ST3) isn’t
an acknowledgment of complete skepticism about value and
modality. As we’ll see below in sect. II, objectors to skeptical theism
often argue that the skeptical theists’ skepticism commits them to
further unpalatable skepticism. But we should recognize up front
that the skeptical theist intends to a rm only a modest form of
skepticism.

What, exactly, is the upshot of (ST1)–(ST3)? Suppose we’ve

thought long and hard about what God-justifying reason there might
be for permitting the following horri c evils, which are commonly
used as examples in the literature:

(E1) the evil of a fawn trapped in a forest re and undergoing

several days of terrible agony before dying.
 (E2) the evil of a 5-year-old girl being raped, beaten, and
murdered by strangulation.

Such thinking will typically involve considering various possible

goods and evils and the conditions of their realization—e.g. whether
permitting (E1) and (E2) is necessary for obtaining some
outweighing possible good or for avoiding the obtaining of some
worse possible evil. Now suppose we fail to come up with anything
that we think is a God-justifying reason for permitting, say, (E2).
That is, suppose that both of the following are true:
(a) none of the possible goods we know of that outweigh (E2)
stand in entailment relations we know of to (E2) such that
obtaining those goods would justify permitting (E2);
(b) none of the possible evils we know of that are worse than
(E2) stand in entailment relations we know of to (E2) such
that preventing the obtaining of those evils would justify
permitting (E2).
If we recognize the truth of (ST3), then it seems we can’t infer from
(a) and (b) that it’s false or even unlikely that permitting (E2) (or
something as bad or worse) is required—by entailment relations we
don’t know of—for the obtaining of outweighing possible goods we
know of or the prevention of worse possible evils we know of. We
are simply in the dark about whether there are such entailment
relations between the possible goods and evils we know of.7 And if
we recognize the truth of (ST1) and (ST2), then it seems we can’t
infer from (a) and (b) thatit’s false or even unlikely that permitting
(E2) (or something as bad) is required for the obtaining of
outweighing possible goods we don’t know of or for the prevention
of worse possible evils we don’t know of. We are simply in the dark
about whether there are goods and evils we don’t know of that
could feature in a God-justifying reason for permitting (E2). Thus,
(ST3) keeps us from using (a) and (a) to conclude that it’s false or
even unlikely that:
(c) the possible goods or evils we know of feature in a God-
justifying reason for permitting (E2).
And (ST1) and (ST2) together keep us from using (a) and (b) to
conclude that it’s false or even unlikely that:
(d) the possible goods or evils we don’t know of feature in a God-
justifying reason for permitting E2.
In short, (ST1)–(ST3) tell us that we can’t use our failure to think of
a God-justifying reason for permitting the horrendous evil (E2) to
conclude that it’s unlikely that there is such a reason—either among
known goods and evils or among unknown goods and evils.8

Analogies are often used to support and drive home the skeptical
theist’s point. We can’t use our failure to see any insects in the
garage (when taking a look from the street) to conclude that it’s
unlikely that there are any insects in the garage. We can’t use our
failure to discover any rational agents on other planets to conclude
that it’s unlikely that there are some on some other planet. We can’t
(if we’re chess novices) use our failure to detect a good reason for a
particular chess move made by a world champion chess player to
conclude that it’s unlikely that there is any good reason for that
chess move. Likewise, say skeptical theists, we can’t use our failure
to discern any God-justifying reason for permitting (E2) to conclude
that it’s unlikely that there is any God-justifying reason for
permitting (E2). There’s nothing unreasonable or excessive about
the skepticism involved in the cases of the insects, extraterrestrial
life, or chess champion. Skepticism in those cases doesn’t seem to
force us to accept other more extreme and unpalatable sorts of
skepticism. Likewise, says the skeptical theist, there’s nothing
unreasonable or excessive about the skepticism involved in the case
of God-justifying reasons for permitting (E2).

Are these good analogies? Are we really that ignorant about

possible goods, possible evils, and possible entailment relations
between them? Notice that (ST1)–(ST3) don’t deny there are many
possible goods, evils, and entailment relations between them that
we know of. So the claim isn’t that we know very little about these
things. Rather, the claim is just that we have no good reason to
think that what we know of these things is representative of what
there is to know about them. We’ve no reason to deny that what we
know about possible goods, evils, and entailments between them is
a very small (percentage-wise) and unrepresentative sample of these
things—unrepresentative with respect to the property of guring in a
(potentially) God-justifying reason for permitting the evils we see around
us. But if we have no such reason, then we are seriously in the dark
about whether the possible goods, evils, and entailments between
them are likely to contain the makings of a potentially God-
justifying reason to permit (E2). And this, says the skeptical theist,
makes the analogies in the previous paragraph seem like good ones.

(ST1) and (ST2) suggest that we don’t have good reason to deny
that there is, among the unknown goods and evils, a God-justifying
reason for permitting (E2). (ST3), on the other hand, suggests that
we don’t have good reason to deny that there is, among the known
goods and evils, a God-justifying reason for permitting (E2). There’s
another skeptical thesis, the import of which is similar to (ST3)’s:

(ST4) We have no good reason for thinking that the total moral
value or disvalue we perceive in certain complex states of
a airs accurately re ects the total moral value or disvalue
they really have.9

The question raised here is: in comparing some of the very complex
goods and evils we know of that are unrelated to the concerns of
everyday life, why think we are able to grasp them su ciently to
make the value comparisons needed to determine whether securing
or preventing them could justify the permission of the evils around
us? If we can’t grasp them su ciently to make such value
comparisons, then our failure to think of a God-justifying reason for
permitting some evil might be due to our failure to recognize that
some good we know of outweighs (or that some evil we know of is
worse than) the evil in question. Less emphasis is placed on (ST4) in
the literature and it’s not needed to make the skeptical theist’s
point. But it’s worth mentioning (ST4) as an additional
consideration that supports the lesson taught by (ST3)—namely,
that there may be a God-justifying reason for permitting (E1) and
(E2) among the goods and evils we know of. Using van Inwagen’s
terminology, (ST4) expresses skepticism about our grasp of the
intrinsic value (or, as Alston puts it, the nature) of at least some of
the goods and evils we know of while (ST3) expresses skepticism
about our grasp of the extrinsic value (or, as Alston puts it, the
conditions of realization) of the goods and evils we know of.10

Three further clari catory points are worth mentioning here.

First, the skepticism encouraged by (ST1)–(ST4) seems to be focused
on our ability to make informed judgments about how
considerations of consequences would (if God existed) factor into
God’s decisions about what is the best thing to do. (ST1)–(ST4) have
to do with our knowledge and understanding of the realm of
possible goods and evils—including our knowledge and
understanding of the entailment and comparative value relations
that hold between possible goods and evils. An appreciation of these
relations is important when considering what the consequences are
of bringing about or preventing a good or of preventing or
permitting an evil. However, the fact that (ST1)–(ST4) are relevant
to considering such consequences doesn’t in any way take for
granted the truth of consequentialist ethical theories. For it may be
that consequentialist ethical theories are false (because, say, we
have absolute duties that bind us regardless of the consequences),
but that very often, moral agents should be guided in their moral
deliberations by considerations of consequences. This is because
very often the right thing to do is to try to bring about what seems
best for those we love, so long as doing so involves no violation of
duties. Thus, non-consequentialist ethical theories have no trouble
allowing for considerations of consequences to play a role in moral
decision-making.

The second clari catory point is that there’s a di erence

between (a) describing a potentially God-justifying reason X and
then announcing that we have no good reason to think it’s unlikely
that X itself is the God-justifying reason God has for permitting some
evil such as (E1) or (E2) and (b) simply pointing out that, in light of
(ST1)–(ST4), we have no good reason to think it unlikely that there
is some God-justifying reason God has for permitting such an evil.
Although some skeptical theists (e.g. Alston 1991 and van Inwagen
1991) sometimes aim to do (a), it’s a mistake to think (as some
philosophers seem to) that (b) by itself—without any e ort to do (a)
—is insu cient for defending the skeptical theist’s skepticism. So
long as we have reason to endorse (ST1)–(ST4) and we can see that
they do in fact imply that we aren’t justi ed in thinking it’s unlikely
that there are God-justifying reasons for permitting the horri c evils
we see around us, doing (a) is unnecessary for making the skeptical
theist’s case.

Third, it’s important to realize that the skeptical theist’s

skepticism does nothing to show that theism is likely to be true or
reasonable to believe. But, so far as I know, no theist has claimed
otherwise. Instead, skeptical theists claim that the skeptical theist’s
skepticism undermines certain arguments from evil for atheism,
showing that such arguments don’t make it reasonable to reject
theism. Which arguments from evil does it undermine? Let’s
consider that question next.

B. To Which Arguments from Evil Does the Skeptical Theist’s

Skepticism Apply?

There are many di erent arguments from evil. For some it’s pretty
clear that the skeptical theist’s skepticism applies whereas for others
it’s controversial whether it applies.11 (Whether there are some
arguments from evil to which it clearly doesn’t apply is a question I
won’t address in this chapter.) Let’s begin by looking at three
arguments—all by William Rowe—to which the skeptical theist’s
skepticism seems quite clearly to apply.

It’s easy to see how (ST1)–(ST4) apply to the following

argument, similar to the one given in the opening paragraph of this
chapter:

(A1) We can’t think of any God-justifying reasons for permitting

(E1) and (E2).12

(A2) So probably there aren’t any God-justifying reasons for

permitting (E1) and (E2).

(A3) If an omnipotent, omniscient, perfectly good being existed,

it wouldn’t permit any evils unless it had a God-justifying
 reason for permitting them.

(A4) Therefore, probably there is no omnipotent, omniscient,

perfectly good being.

This is basically how Rowe’s 1979 argument from evil goes. The
skeptical theist’s skepticism straightforwardly challenges the
inference from (A1) to (A2). The fact that we can’t think of any
God-justifying reasons for permitting (E1) and (E2) doesn’t make it
probable that there aren’t any—any more than the fact that we can’t
see any insects in the garage (from our vantage point standing in the
street) makes it probable that there aren’t any insects in the garage.

Rowe proposed a slightly di erent argument from evil in later

papers (see his 1988 and 1991):

(P) No good we know of justi es an omnipotent, omniscient,

perfectly good being in permitting (E1) and (E2);
therefore [it’s probable that],
(Q) no good at all justi es an omnipotent, omniscient, perfectly
good being in permitting (E1) and (E2);
therefore [it’s probable that],

(∼G) there is no omnipotent, omniscient, perfectly good being.13

This attempt by Rowe to improve the 1979 argument has been

viewed by skeptical theists as a change for the worse. For now we
have two problems. First, the inference from (P) to the likelihood of
(Q) seems faulty given (ST1) and (ST2). Even if we could be sure
that no known goods gure in God-justifying reasons for (E1) and
(E2), this tells us nothing about whether there are unknown goods
and evils that could gure in God-justifying reasons for (E1) and
(E2) (since—as (ST1) and (ST2) tell us—we’ve no reason to think
the goods and evils we know of are representative of the goods and
evils there are). But second, (ST3) and (ST4) imply that we have no
good reason to believe (P). As I noted in sect. IA, (ST3) and (ST4)
emphasize that we have an inadequate grasp of the extrinsic and
intrinsic value of some known goods. In light of this, we are simply
in the dark about whether there is a God-justifying reason for (E1)
and (E2) among the goods we know of. For both these reasons, this
second argument from Rowe also fails by the skeptical theist’s
lights.14

A third argument by Rowe, proposed speci cally to avoid the

skeptical theist’s skepticism, nevertheless also falls prey to it.15
Instead of arguing from (P) to (Q) and then from (Q) to (∼G), as in
the previous argument, this time Rowe argues directly from (P) to
(∼G). We needn’t examine any further how this argument goes in
order to see that (ST3) and (ST4) imply (as I noted in the previous
paragraph) that we have no good reason to believe (P). This is
enough by itself to make this argument unacceptable from the
skeptical theist’s perspective.16

Let’s turn next to two arguments whose proponents view them as

not being targeted by the skeptical theist’s skepticism, but which
are, arguably, targeted by it nonetheless. The rst of these, touched
on by Rowe and developed at length by John Schellenberg, is the
argument from divine hiddenness:

(B1) If God exists and is unsurpassably loving, then for any

human subject H and time t, if H is at t capable of relating
personally to God, H at t believes that God exists, unless H
is culpably in a contrary position at t.17

(B2) There is a human subject H and a time t such that H is at t

capable of relating personally to God, H is not culpably in a
contrary position at t, and yet H at t fails to believe that
God exists.18

(B3) Therefore, God does not exist.

Does the skeptical theist’s skepticism raise any di culties for this
argument? Here’s a way in which it might: (B1) is false if there are
God-justifying reasons to permit a period of divine hiddenness
(that’s what I’ll call a period of time during which a human, who is
capable of relating personally to God and is not culpably in a
contrary position, fails to believe in God). For if there were such
reasons, then, if God existed, he would permit periods of divine
hiddenness, contrary to what (B1) says. After all, God, being
perfectly loving, would want what is best for his creatures. So long
as it isn’t intrinsically wrong to permit a period of divine hiddenness
regardless of the bene ts it might produce (and there seems to be no
good reason for thinking this is the case), God would do so if doing
so would bring about a greater good or prevent a worse evil. But
(ST1)–(ST4) suggest that we’re simply in the dark about whether
(and how likely it is that) there are any God-justifying reasons for
permitting a period of divine hiddenness. Thus, since we know that
the existence of a (potentially) God-justifying reason for permitting
divine hiddenness entails the falsity of (B1), and we are in the dark
about the truth and the likelihood of the claim that there exists such
a reason, it follows that we are likewise in the dark about the truth
and likelihood of (B1).

The second argument from evil I want to look at whose

vulnerability to the skeptical theist’s skepticism is controversial is
proposed by Paul Draper. Let ‘T’ be theism. Let ‘HI’ be the
Hypothesis of Indi erence—i.e. the hypothesis that neither the
nature nor the condition of sentient beings on earth is due to the
malevolent or benevolent actions of any non-human person. And let
‘O’ report what we know about the kinds, amounts, and distribution
of pain and pleasure in the world. Using these abbreviations, Draper
argues as follows:

(C1) O is known to be true.

(C2) HI is at least as probable intrinsically as T.

(C3) Pr(O/HI) >! Pr(O/T).19

(C4) Therefore, other evidence held equal, T is very probably

false.20

How might the skeptical theist’s skepticism apply to this argument?

Well, in order sensibly to assert (C3), we’d have to have some idea
what Pr(O/T) is—at the very least, we’d have to know that it’s not
quite high, since if Pr(O/T) is high, (C3) is false. But it seems that
Pr(O/T) depends on what the likelihood is of there being a God-
justifying reason for permitting the evil state of a airs described in
O (i.e. if the latter is quite high, so is the former; and if the latter is
quite low, so is the former). And, according to (ST1)–(ST4), we’re in
the dark about whether—and how likely it is—that there is such a
reason. This means we’re in the dark about whether that likelihood
is quite high. Thus, since (1) we know that the high probability of
the existence of a (potentially) God-justifying reason for permitting
the evils described in O entails that Pr(O/T) is high (making (C3)
false) and (2) we are in the dark about whether it’s highly likely
that there exists such a God-justifying reason, it follows that we are
likewise in the dark about the truth and likelihood of (C3).

As I’ve already indicated, Schellenberg and Draper object to the

charge that the skeptical theist’s skepticism applies to their
arguments from evil.21 As I understand them, they each o er this
sort of reply: ‘It’s true that we know that my premise implies that
it’s unlikely for there to be a God-justifying reason for permitting
the evil I’m focusing on. But I’ve given reasons for my premise.
From all this it follows that these reasons for my premise double as
reasons for thinking it’s unlikely for there to be a God-justifying
reason for permitting the evil I’m focusing on. Hence we have a
reason for thinking it’s unlikely for there to be such a reason, even if
(ST1)–(ST4) are true.’22
 Is this reply adequate? Can the reasons Draper and Schellenberg
give for their premises successfully be used to show that it’s unlikely
for there to be a God-justifying reason for permitting the evils in
question? The evils they focus on are these:

(E3) the evil of there being a period of time during which a

human who is capable of relating personally to God and is
not culpably in a contrary position, fails to believe in God.

(E4) the evil of there being the distribution of pain and pleasure
that we know there is in the world.

It’s clear that there are evils worse than (E3) (e.g. that human never
experiencing the beati c vision; if God were forced to in ict one or
the other on a beloved creature, he would in ict (E3)). And it’s clear
that there are goods that outweigh (E3) (e.g. that human
experiencing the beati c vision; if God wanted one of his creatures
to experience the beati c vision and could do so only by permitting
that creature to undergo (E3), he would permit it). The same points
apply to (E4). But could an omnipotent being be forced to permit
(E3) or (E4) or something as bad, in order to obtain some
outweighing good? (ST3) suggests we are seriously in the dark
about the answer to this question. Insofar as we have no reasons for
thinking the entailments we know of between possible goods and
evils are representative of the entailments there are between goods
and evils, we simply aren’t in a position to comment in an informed
way about how likely it is that an omnipotent being would be forced
to permit (E3) or (E4) or something as bad, in order to obtain some
outweighing good. What do Schellenberg and Draper have to say
about this? What reasons do they give for thinking it’s false or
unlikely that God would permit (E3) or (E4) that might possibly
double as reasons to think it is unlikely or false that any
outweighing possible good entails the permission of (E3) or (E4) or
something as bad?

As I read them, they point, ultimately, to two main reasons for

thinking it’s false or unlikely that God (if he exists) would permit
(E3) or (E4): God’s perfect love and God’s in nite resourcefulness.23
But I don’t quite see how either of these is at all relevant to whether
any outweighing possible good entails the permission of (E3) or (E4)
or something as bad. Suppose that God, if he exists, deeply loves all
humans and places an exceedingly high value on having a
relationship with them. What does that imply about whether some
outweighing possible good entails the permission of (E3) or (E4) or
something as bad? Nothing much as far as I can see. Whether some
outweighing possible good entails the permission of (E3) or (E4) or
something as bad is a necessary truth that doesn’t seem to be
relevant to whether God loves us deeply and wants a relationship
with us. Likewise, suppose—as also seems plausible—that God, if he
exists, has the in nite resourcefulness implied by omniscience and
omnipotence. Does that suggest that no outweighing possible good
entails the permission of (E3) or (E4) or something as bad? Again, I
can’t see how. It’s widely accepted that omnipotence doesn’t imply
the ability to actualize what is metaphysically impossible. This
means that if some outweighing possible good does entail the
permission of (E3) (or something as bad), God wouldn’t be able to
do anything about that.

Can one sensibly appeal to God’s perfect love and in nite

resourcefulness to support the claim that God is unlikely to permit
(E3) or (E4) and then use that result to support the further claim
that no outweighing possible good entails the permission of (E3) or
(E4) or something as bad? No. Once we see that God’s perfect love
and in nite resourcefulness don’t support the conclusion that no
outweighing possible good entails the permission of (E3) or (E4) or
something as bad, we thereby see that there’s also no reason to
think they support the claim that God is unlikely to permit (E3) and
(E4).

There is, however, a way in which Draper’s argument has an

advantage over Schellenberg’s. For, given that Draper’s argument
focuses on comparing probabilities, he can o er the following reply
(suggested by the work of Mark Bernstein24): ‘I can grant the
skeptical theist’s point that I’m in the dark about Pr(O/T). But
consider what we should do when deciding between inconsistent
claims X and Y if we know that Pr(P/X) is high and we’re
completely in the dark about Pr(P/Y)—where P is some piece of
evidence we have. For example, suppose we know that P, where
that is the claim that the ball recently pulled out of one of two
nearby urns is white. And suppose that X is the claim that the ball
was randomly pulled out of the left urn and Y is the claim that the
ball was randomly pulled out of the right urn. Let’s say you know
that the strong majority of the balls in the left urn are white and,
therefore, that Pr(P/X) is high. And let’s say you have no idea what
Pr(P/Y) is—you have no idea whether all the balls in the right urn
are white or none of them are or some fraction between 0 and 1 are
white. Now, if you are o ered a bet of one million dollars to identify
correctly which urn the white ball was drawn from, the reasonable
thing for you to do is to say X is true (i.e. that the ball came from
the left urn). This shows that when you are comparing a known high
probability with an unknown probability, it’s reasonable to think
that the unknown probability is lower than the known high
probability. Thus, (C3) from Draper’s argument is true even if we
don’t know what the value of Pr(O/T) is.’

This reply is e ective if we can reasonably use the Principle of

Indi erence to conclude that, if the likelihood of Pr(O/T) is
unknown, then the likelihood that Pr(O/T) is less than 0.5 is 0.5,
and the likelihood that it is higher than 0.9 is 0.1, and the
likelihood that it is between 0.11 and 0.12 is about 0.01, and so on.
But suppose it’s not reasonable to use the Principle of Indi erence in
that way in this case. Suppose instead it’s reasonable to conclude,
based on (ST1)–(ST4), that we have no idea what the likelihood is
that Pr(O/T) is, say, higher than 0.99. Suppose the likelihood that
Pr(O/T) is higher than 0.99 might be 0 or 1 or anything in between;
contrary to what the Principle of Indi erence suggests, we just don’t
know how likely it is that Pr(O/T) is higher than 0.99. Then the
reply in the previous paragraph—which seems to assume that
Pr(O/T) is more likely to be lower than a high probability than it is
to be higher than that high probability—fails. From the failure of
this reply (given the suppositions noted in this paragraph) we may
conclude that if the skeptical theist’s skepticism can reasonably be
combined with rejecting the use to which the Principle of
Indi erence was put in the reply inspired by Bernstein in the
previous paragraph, then the skeptical theist’s skepticism causes
trouble for Draper’s argument.25 I won’t undertake here to discuss
the antecedent of that conditional except to say that I myself don’t
nd it implausible.26

II. OBJECTIONS TO THE SKEPTICAL THEIST’S SKEPTICISM

I will consider two sorts of objections to the skeptical theist’s

skepticism. The rst argues (1) that for some horri c evils there
appear not to be any God-justifying reasons to permit them and (2)
that this fact counts as a prima facie reason for thinking there are no
God-justifying reasons to permit them—a prima facie reason that is
not overridden by the considerations the skeptical theist highlights.
The second sort of objection comes in several versions. The basic
idea of each version is that by endorsing the skeptical theist’s
skepticism, one is committed to some other unpalatable form of
skepticism (such as skepticism about the past or the external world
or moral reasoning). The obvious implication is that, given that we
should reject the unpalatable skepticism, we should reject the
skeptical theist’s skepticism too. In what follows, I’ll consider and
respond to both sorts of objection.

A. Objection One: The Appearance of No God-Justifying

Reasons

Swinburne’s objection to the skeptical theist’s skepticism depends on

three points. First he argues that, in the absence of any God-
justifying reason we can think of for permitting horri c evils such as
(E1) and (E2), it appears that there is no God-justifying reason for
permitting them. Second, he notes that the Principle of Credulity
tells us that, other things being equal, it’s rational to believe that
things are as they appear. Third, although the skeptical theist is
right that things may be better than they appear (since there may be
some unknown greater good that is secured by permitting the evil in
question and that itself brings about no greater evil), we should also
recognize that things might be worse than they appear (since there
may be some unknown greater evil that is produced by the evil in
question and that doesn’t itself bring about any greater good); and
since it’s just as likely that things are worse than they appear as it is
that they’re better than they appear, we should conclude that things
are probably as bad as they appear. Based on the rst two points,
Swinburne concludes that if we can’t think of any God-justifying
reason for permitting (E1) and (E2), then, other things being equal,
this makes it reasonable for us to conclude that probably there is no
such reason.27 And based on the third point, Swinburne concludes
that although the skeptical theist is right to point out that God may
have reasons we are ignorant of, this does nothing to change the
conclusion derived from the rst two points.28

The main problem with Swinburne’s objection to the skeptical

theist’s skepticism is that the skeptical theist thinks there is good
reason not to grant his rst point that it appears that there is no God-
justifying reason for permitting evils such as (E1) and (E2).
According to (ST1)–(ST4), it doesn’t appear that there is no God-
justifying reason for permitting for (E1) and (E2). Nor does it appear
that there is such a reason. Nor does it appear likely that there is.
Nor does it appear likely that there isn’t. Rather, we just don’t know
how likely it is that there is a God-justifying reason for permitting
evils such as (E1) and (E2). To see this, notice the di erence
between saying ‘it appears that there are no Fs’ and saying ‘it
doesn’t appear that there are Fs’. If, while standing in a friend’s
garage that is normal in size and uncluttered, you look around and
can’t see an automobile in it, then it’s reasonable for you to
conclude both that it doesn’t appear that there is an automobile in the
garage and that it appears that there is no automobile in the garage. But
suppose instead you are standing on the street and, upon looking
into a rather cluttered garage, you fail to see any eas. Then,
although it’s rational for you to conclude that it doesn’t appear (to
you upon looking from the street) that there are eas in the garage, it’s
not rational to conclude that it appears that there are no eas in the
garage. According to (ST1)–(ST4), we’ve no good reason to deny that
God-justifying reasons for permitting (E1) and (E2) sought for from
our vantage point (apart from divine revelation) are like eas in a
cluttered garage viewed from the street. Hence, we can’t reasonably
conclude, as Swinburne does, that it appears that there is no God-
justifying reason for permitting evils such as (E1) and (E2). At best we
can reasonably conclude that it doesn’t appear (to us) that there is a
God-justifying reason for permitting (E1) and (E2). But of course that
can’t be used in conjunction with Swinburne’s Principle of Credulity
to get him the conclusion he’s after in responding to the skeptical
theist’s skepticism. For if it could, then we could use a Swinburne-
like response to conclude, on the basis of our failure to see any eas
in the garage when looking from the street, that there probably
aren’t any in the garage.

Thus, Swinburne misconstrues the skeptical theist’s response. He

thinks the skeptical theist’s aim is to show that the likelihood of
some evil or other on theism might for all we know be higher than
it initially appears. And he replies that similar remarks show that it
might for all we know be lower than it initially appears. And he
replies that similar remarks show that it might for all we know be
higher than it initially appears. Since, according to him, it’s just as
likely to be higher than it initially appears as it is to be lower than it
initially appears, it’s reasonable to go with initial appearances. But
in fact, the skeptical theist’s response is that we aren’t justi ed in
thinking the probability judgment initially appears the way
Swinburne says it appears. Clear thinking and re ection on (ST1)–
(ST4) reveal that there’s no particular value or range (short of the
range between 0 and 1) that the probability in question appears to
be.29

Swinburne’s (1998: 23) response to the remarks in the previous

paragraph is as follows:30

And if our understanding of possible reasons why anyone might

allow su ering to occur provides us with no reason for supposing
that a good God might allow certain su ering, we ought to believe
that there is no God—unless we have a contrary reason. Just re ect
on some of the horrors that we read about in our newspapers and
history books: the prolonged cruelty of parents to lonely children,
the torture of the innocent, the long-drawn-out acute physical pain
of some disease, and so on. If we cannot see all that as a reason for
believing that there is no all-good and all-powerful being, when we
cannot think of any reason why such a being should allow it all to
happen, there really is something deeply wrong with us. We have
lost our sensitivity to the good.

In short, if we don’t conclude, upon our failure to think of any God-

justifying reason for permitting evils like (E1) and (E2), that (other
things being equal) there probably is no God, we have lost our
sensitivity to the good.31 I nd this extremely unpersuasive. What I
grant is that if we can’t see that a good God would, other things
being equal, want to prevent horri c evils, then it seems we have
lost our sensitivity to the good. When we see or learn of utterly
horri c su ering, the sensible and appropriate response is to be
extremely upset that it has occurred and, with deep feeling, to think
‘There had better be a good reason for God, if he exists, to permit
that su ering; if there isn’t, then there is no perfectly good God.’

But although it’s extremely common, it’s not reasonable, given

(ST1)–(ST4), to go on from there to think that it’s unlikely that there
is a God-justifying reason for permitting such su ering (and, hence,
unlikely that God exists). For the only basis we have for that
conclusion is our own inability to think of any such reason. Suppose
I’m considering possible goods, possible evils, and possible
entailment relations between them with the aim of discovering a
(potentially) God-justifying reason for permitting (E1) and (E2). If I
accept (ST1)–(ST4), then my failure, upon engaging in such a
search, to discover any God-justifying reason for permitting (E1)
and (E2) won’t lead me to conclude that there is no such reason. In
refraining from drawing that conclusion, I’ll be just as rational as
the person who refrains from concluding there are no eas in the
garage on the basis of a failure to see any when looking from the
street. Does the fact that I refrain under these conditions from
concluding that there are no God-justifying reasons for permitting
(E1) and (E2) (or even that it’s unlikely that there are any such
reasons) demonstrate that I’ve lost my sensitivity to the good?
Hardly. It demonstrates only that I don’t want to jump
inappropriately to unfounded conclusions (about eas in the garage
or about God justifying reasons). But if it’s reasonable for me to
refrain from concluding that it’s false or unlikely that there are God-
justifying reasons for permitting (E1) and (E2) (on the basis of my
failure to discover any such reasons), then it’s also reasonable for
me to refrain from concluding on that basis that it’s false or unlikely
that God exists. Such reasonable thinking does nothing to suggest
any insensitivity to the good.

B. Objection Two: Commitment to Unpalatable Skepticism

This second type of objection to skeptical theism seems to be the

most common, though (as I’ve already noted) it comes in di erent
forms, depending on what sort of unpalatable skepticism it focuses
on. I’ll brie y consider four forms of this objection.

B.1. Skepticism about Certain Theistic Arguments

The rst charge of skeptical commitment is the charge that if one

endorses the skeptical theist’s skepticism, one can’t consistently
endorse certain arguments for theism.32 In particular, some
arguments for God’s existence based on identifying something as an
all-things-considered good—even in light of its consequences—will
be undermined by the skeptical theist’s skepticism. So, for example,
if the order one sees in the natural world or the joy one witnesses in
people’s lives is identi ed as reason to think that there is a good
being who is the cause of such things, one is failing to take into
account the lessons of (ST1)–(ST4). Given our cognitive limitations,
we simply don’t know what evils might be entailed by those good
things and this prevents us from being able to conclude that they
are all-things-considered goods that an omnibenevolent being would
bring about.

The skeptical theist’s response to this charge should, I think, be

to accept it. We aren’t able to determine whether something is an
all-things-considered good simply by noticing how good it is since
we don’t know what it might bring in its wake. Of course, perhaps
there are other ways of learning something is an all-things-
considered good. Maybe divine revelation is such a way. But given
that theists don’t seem to be able to arrange for divine revelation to
be passed on to non-theists, this way of learning something is an all-
things-considered good won’t be of much help in o ering a theistic
argument that will be persuasive to a non-theist. We needn’t
conclude from this that the skeptical theist’s skepticism is
inconsistent with every way of arguing for the existence of a good
God (just as we needn’t conclude it is inconsistent with every
atheistic argument). But there’s no doubt that some theistic
arguments collapse under pressure from (ST1)–(ST4).

B.2. Skeptical Theism No Matter How Much Evil There Is

The second charge of skeptical commitment is the charge that

skeptical theists are forced to admit that, no matter how much
su ering or evil we witness, we cannot reasonably conclude that
God wouldn’t permit it. As Rowe puts it, even ‘if human life were
nothing more than a series of agonizing moments from birth to death,
[the skeptical theists’] position would still require them to say that
we cannot reasonably infer that it is even likely that God does not
exist’.33 But since this is an absurd conclusion, Rowe argues, the
skeptical theist’s skepticism, which forces her to this absurd
conclusion, must also be mistaken.

This charge does not stick. It’s true that, given (ST1)–(ST4), we
can’t determine merely by trying to consider the consequences of goods
and evils whether a certain amount or kind of su ering is such that
there couldn’t be a God-justifying reason to permit it. But there are
other ways of determining this that don’t rely on considerations of
consequences. Tooley has proposed, as a premise in one of his
arguments from evil, the principle that God would permit horri c
su ering only for the bene t of the su erer.34 I don’t nd that
particular moral principle plausible.35 But there are others like it
that seem more promising. Swinburne (1998: 229–36) argues that a
perfectly good God would not permit su ering unless the su erer’s
life is on the whole a good one (notice that this is a weaker
requirement than Tooley’s according to which the reason the
su ering is permitted must be to bene t the su erer). It’s true that
Swinburne is no friend of skeptical theism, but I see no reason why
those endorsing the skeptical theist’s skepticism couldn’t
consistently accept this principle Swinburne proposes (since we can
see the truth of such general principles even if we can’t see what all
the consequences of the goods and evils we know of are). And by
accepting this principle, skeptical theists would have reason to say
that a good God would not permit a human life to be literally
nothing more than a series of agonizing moments from birth to death.

B.3. Skepticism about the Past and the External World

According to a third charge, those who endorse the skeptical theist’s

skepticism are committed to skepticism about the external world
and the past. The idea is basically that, given (ST1)–(ST4), I have no
idea whether there is a God-justifying reason to permit or to arrange
for the following bad states of a airs to obtain:

(E5) My being a bodiless victim of an evil demon who deceives

me into thinking there’s an external physical world when in
fact there is not.

(E6) My being deceived by an evil demon into believing that I

and the physical universe have been around for years when
in fact I and the physical universe came into existence ve
minutes ago (me with false memories, the physical universe
with the misleading appearance of being old).

Given this, the objector concludes that if I endorse the skeptical

theist’s skepticism, I should likewise endorse skepticism about the
external world and about the past.
The skeptical theist’s reply is to note that our way of knowing that
(E5) and (E6) aren’t actual is not by considering possible goods,
possible evils, and entailments between them—seeing that these
provide no God-justifying reason to permit the obtaining of (E5) and
(E6) and concluding that, since God exists, (E5) and (E6) must not
be actual.36 Not at all. Rather, we have some independent way of
knowing that (E5) and (E6) aren’t actual and we can conclude, from
the fact that they aren’t actual, that if God exists, he has no good
reason to arrange for them to be actual. This way of knowing
something about God’s reasons is consistent with (ST1)–(ST4).
(What is the independent way—i.e. independent of considering
possible God-justifying reasons for permitting (E5) and (E6)—in
which we know (E5) and (E6) aren’t actual? Epistemologists o er
many di erent answers to this question. This isn’t the place to
explore these answers. But it’s widely held, by theists and non-
theists, that we have some independent way of knowing that (E5)
and (E6) aren’t actual.37)

Why can’t the proponent of the argument from evil make the
same sort of move? Why can’t she say that we know independently
that there’s no God-justifying reason for permitting evils such as
(E1) and (E2)—not by surveying possible goods, possible evils, and
entailments between them but in some other way? That’s certainly a
strategy worth considering. But what we need is some plausible
suggestion of what that independent way of knowing might be. And
in the case of the arguments from evil we’ve been considering, no
such suggestion is forthcoming. It’s not plausible to claim that we
know independently that such a supremely loving and resourceful
being as God is likely to prevent evils such as (E1) and (E2) or (E3)
and (E4) mentioned above in connection with Draper and
Schellenberg). What we seem to know independently is that a
perfect being de nitely wouldn’t permit (E1)–(E4) without a God-
justifying reason for doing so. But this doesn’t enable us to know
independently that God is likely to prevent (E1)–(E4), not unless we
have some independent way of knowing that it is unlikely for there to
be a God-justifying reason for permitting those evils. But plausible
suggestions of independent ways of knowing that—ways that don’t
rely on our failure to think of any such reasons upon considering
possible goods, possible evils, and entailments between them—are
in short supply.

B.4. Skepticism about Morality

Perhaps the most common and in uential charge of skeptical

commitment lodged against skeptical theists is the one that says
that, by endorsing the skeptical theist’s skepticism, we are forced
into an appalling sort of skepticism about the morality of various
actions. For example, the skeptical theist’s skepticism tells us that
we have no good reason to think that the horri c rape and murder
of a small child won’t bring about some outweighing greater good.
Given this, why should we think it’s good to prevent such horri c
su ering if we are easily able to do so? According to this sort of
objector, considerations like these suggest that consistency requires
the skeptical theist to be skeptical about whether it’s right to
prevent such horri c su ering when we easily can. But skepticism
about such moral issues as these is both appalling and implausible.
Hence, the skeptical theist’s skepticism, which supposedly leads to
this unpalatable moral skepticism, should be rejected.38

By way of response, those (theists and non-theists) who accept

the skeptical theist’s skepticism can o er the following proposals
about how we make moral decisions. First, it is very often important
in making moral decisions that we consider the consequences of our
actions—the good and the harm that we think will result from our
choices. We can acknowledge this while at the same time
recognizing that we may have some duties that constrain our
behavior independently of the consequences of our actions. (So, as
already noted at the end of sect. I.A, recognizing the importance of
considering consequences doesn’t commit us to a consequentialist
moral theory.) Second, in cases where it is important for us to be
guided by considerations of possible good and bad consequences of
our actions, we aren’t morally bound to do what in fact has the
overall best consequences (since we typically can’t determine that).
What is relevant are the likely consequences we have some reason
to be con dent about after a reasonable amount of time and e ort
aimed at identifying the expected results of our behavior. If, after
such consideration, a particular course of action seems clearly to
maximize the good (or minimize the bad) among the consequences
we’re able to identify and we non-culpably and reasonably take
ourselves to have no overriding consequence-independent obligation
to refrain from that action, then that action is a morally appropriate
one for us to perform. Third, God’s moral decision-making can be
viewed as analogous to our own as it was just described. God too
will seek to bring about the best consequences except in cases where
what morality requires is not dependent on consequences. And, in
those cases where consequences of an action matter, God too will
put the right amount of e ort and time into determining what the
best consequences are. (Of course, in God’s case, this might require
no time and not much e ort; and, unlike in our case, what God
thinks is the action with the best consequences is the action with the
best consequences.39) Fourth, when considering whether to permit
someone to su er in order to bring about some outweighing good, it
matters tremendously what one’s relationship is to the one
permitted to su er. It may be morally appropriate for me to allow or
even bring about certain minor sorts of su ering in my own child
for her good whereas similar treatment of some stranger’s child
would be morally inappropriate. Likewise, it may be morally
appropriate for your loving and omniscient creator to permit you to
experience preventable horri c su ering in order to achieve some
good whereas it wouldn’t be morally appropriate for another human
to do so.

In light of the four considerations from the previous paragraph,

the ‘moral skepticism’ objection to skeptical theism seems to lose its
force. The fact that we’re in the dark about whether there are
reasons that would justify a perfect being in permitting easily
preventable horri c su ering doesn’t give us a reason to doubt that
we ought to prevent easily preventable horri c su ering when, even
after taking a reasonable amount of time and e ort, we can think of
no outweighing goods that will be achieved by our permitting it. For
we are reasonable and moral to base our decision on the likely
consequences we know of and ignore the far-o ones we’re ignorant
of (only the most committed actual-consequence consequentialist
would think otherwise). Moreover, in the case of preventing easily
preventable horri c su ering, we know we have a prima facie duty
to prevent great harm to others when this is easy for us to do and
that this gives us a strong prima facie reason to prevent the
su ering. If after a reasonable amount of time and e ort we can’t
think of any negative consequence of such su ering-prevention that
might outweigh its obvious goodness, then we ought to prevent the
su ering we have strong prima facie reason to prevent—even
though we don’t know the long-term consequences of such
prevention (again, only a diehard actual-consequence
consequentialist would think otherwise). And even if we knew that
the overall consequences that would result from permitting the
easily preventable horri c su ering would be good, it’s not at all
clear that we have the sort of relationship with the su erer (e.g. we
aren’t the su erer’s loving creator) that makes it appropriate to
permit the person’s horri c su ering for the sake of some greater
good.

Derk Pereboom has responded on behalf of the ‘moral

skepticism’ objection by arguing that skeptical theists have a reason
for thinking they shouldn’t intervene to prevent easily preventable
horri c su ering. Pereboom makes his case by comparing two
scenarios.40 In the rst, Jack (a nurse) knows that morphine will
ease the su ering of the patients in the clinic in which he assists
doctors. But he’s noticed that in his experience, the doctors never
give morphine to the bone cancer patients though they give it to
other patients. He has no idea why this hasn’t been done (at least
not when he’s been watching) or whether they’ve given morphine to
bone cancer patients in the past. On a day when the doctors are
unable to make it into work, he has the opportunity to relieve the
su ering of bone cancer patients. But (says Pereboom) Jack clearly
has some signi cant moral reason not to give the patients morphine.
In the second scenario, Sue is a doctor who has (rationally) become
a skeptical theist. She sees that God has for millennia let people
su er from a disease X. But a cure has just been developed and she
has an opportunity to administer it. It seems (says Pereboom) that
just as Jack has a signi cant moral reason not to give the patients
morphine, so also Sue has a parallel reason not to administer the
cure for disease X.

The problem with this response by Pereboom is that the most we

should suspect (based on noticing that the disease has progressed in
humans unchecked for millennia) is that a person with a
relationship to humans such as the one God has with them has a
reason not to administer the cure. In Jack’s case, he knows that he
has to the patients a relationship su ciently like the relationship
the doctors have—i.e. a human caregiver—so he knows that if the
doctors have a right to withhold the morphine for some greater
good, he does too. But it’s clear to Sue that she doesn’t have a
relationship to other humans that is like that of a loving creator to
his creatures. She thinks she does not have the right to let them
su er terribly for their moral development, say, even though God
does have that right in light of his relationship with them. So
although Jack has a good reason to hesitate rather than relieve the
su ering of the bone cancer patients, Sue does not have such a
reason to hesitate rather than administer the cure for disease X.
Thus, the ‘moral skepticism’ objection to skeptical theism—like the
other objections I’ve considered in sect. II—does not succeed.

NOTES

My thanks to Paul Draper, Tom Flint, Patrick Kain, Trenton

Merricks, Michael Rea, William Rowe, and Chris Tucker for
helpful comments on earlier drafts.

1. A God-justifying reason for permitting an evil E is, as you

might guess, a reason for permitting E that would justify God,
if God existed, in permitting E.

2. Draper (unpublished) and Howard-Snyder (forthcoming) both

express this sort of concern, though Draper confesses there to
being the one to introduce the terminology in print in his
1996.
3. Unfortunately, that rst approximation needs some tweaking.
For starters, as Plantinga 1967:120 points out, G’s possessing
the two features noted won’t guarantee that the aim of
obtaining G can be used to justify God in permitting E.
Suppose, for example, that G is a conjunctive good (G*&E). It
may be that because G* is so good, the goodness of (G*&E),
outweighs the badness of E. And clearly (G*&E) entails E. So
the conjunctive good (G*&E) has the two features noted
above. But this doesn’t guarantee that the aim of obtaining
(G*&E) would justify God in permitting E. For suppose that G*
doesn’t entail E and that G*’s goodness is greater than the
goodness of the conjunctive good (G*&E). In that case, it
seems God could have done better by obtaining G* instead of
the conjunctive good (G*&E) and that he could do so without
permitting E. Rowe (1979: 10) and Plantinga (1967: 121)
suggest ways to handle this di culty. The basic idea is to say
that if G is to justify God in permitting E, then, in addition to
satisfying the two conditions identi ed above in the text, it
must also be the case that (3) there is no distinct good G* that
is as good or better than G and could be obtained without
permitting E (or something as bad or worse).

Another complication that needs to be mentioned has to do

with the Molinist view that there are counterfactuals of
freedom that are true of individual creaturely essences even
before those essences are instantiated as free creatures.
According to this view, the truth of these counterfactuals of
freedom is contingent. Moreover, on one plausible notion of
freedom, God (if he exists) wouldn’t have control over their
truth, despite their contingency. So—in addition to necessary
truths having to do with entailment relations between possible
goods and possible evils—there might be these contingent
truths that place additional constraints on what God can bring
about. (For an account of how such truths might place
constraints on what God can do, see Plantinga 1974: ch. 9. For
a defense of the view that there are true counterfactuals of
freedom, see Flint 1998.) To deal with these concerns about
counterfactuals of freedom that might place limitations on
what God (if he exists) is able to bring about, we need to say
something about how to interpret the ‘couldn’t’ in clause (2)
and the ‘could’ in clause (3). To say ‘G couldn’t be obtained
without permitting E’ is to say that God (if he exists) is not
able to bring about G without permitting E—either because G’s
obtaining entails the permission of E or because G’s obtaining
together with the counterfactuals of freedom that are contingently
true of individual creaturely essences entails the permission of E.
Likewise, to say ‘G* could be obtained without permitting E’ is
to say that God (if he exists) is able to obtain G* without
permitting E—because it’s the case both that G*’s obtaining
doesn’t entail the permission of E and that G’s obtaining
together with the counterfactuals of freedom that are contingently
 true of individual creaturely essences doesn’t entail the
permission of E.

For simplicity’s sake, I will continue to focus in the main

text only on clauses (1) and (2) mentioned there. But clause
(3) should also be understood as being required and the
interpretation of clauses (2) and (3) mentioned at the end of
the previous paragraph in this note should also be assumed.

4. Probably the best place to start in getting a feel for the

skeptical theist’s position is William Alston’s 1991. Peter van
Inwagen’s 1991 is also a good resource for this purpose.
Wykstra’s 1984 is a classic for contemporary skeptical theists
(and their opponents), though he focuses on formulating a
general epistemic principle (for when we are entitled to claim
‘it appears that p’) instead of on explaining and defending the
rationale behind the skeptical theist’s views as Alston does.

5. These are from Bergmann 2001.

6. The inscrutable evils we see around us are those that many

thoughtful atheists and theists agree are ones for which we
can’t think of a God-justifying reason.

7. As I mentioned in n. 3, the Molinist view that there are true

counterfactuals of freedom reminds us that—in addition to
necessary truths having to do with entailment relations
between possible goods and possible evils—there may be some
contingent truths that place additional constraints on what
 God can bring about. It’s worth noting here that if that
Molinist view is true, we’ve no reason to think that what we in
fact know about these contingent truths for all creaturely
essences—instantiated and uninstantiated—is representative of
what there is to know about them (again, I have in mind
representativeness relative to the property of their guring in
a potentially God-justifying reason for permitting the
inscrutable evils we see around us).

8. It may not be true, as a general rule, that an inference from a

sample is justi ed only if the person making the inference
explicitly believes that the sample is representative. But
justi cation for such inferences does require that it’s false that
the person making the inference does or should disbelieve or
(due to uncertainty) withhold the proposition that the sample
in question is representative (in the relevant respect). And
those who recognize the truth of (ST1)–(ST3) should, it seems,
withhold (due to uncertainty) the propositions about the
representativeness of the samples there mentioned.

9. This is how John Beaudoin formulates this skeptical thesis,

which is re ected in the work of Alston (1996: 324), Howard-
Snyder (1996a: 302–3), and van Inwagen (1991: 161–2). See
Beaudoin 2005: 50.

10. See Alston 1996: 325 and van Inwagen 1991:162–3.

11. To say it applies to an argument isn’t to say it’s right. Rather,

it’s to say that if the skeptical theist’s skepticism is right, then
 these arguments fail.

12. (E1) and (E2) are the evils of the horri c deaths of the fawn
and the 5-year-old girl, both mentioned earlier in this chapter.

13. This summary of the argument given in his 1988 and 1991 is
from Rowe 1996: 262–3.

14. And by Rowe’s lights too, it seems. See ibid. 267 where he says
tha the thinks’ this argument is, at best, a weak argument’ and
he proposes ‘to abandon this argument altogether’.

15. This argument is proposed in Rowe 1996.

16. See Alston 1996: 323–5; Bergmann 2001: 294 n. 9; Howard-

Snyder 1996 a: 295; and Plantinga 1998: 534 for this
objection to P. See Plantinga 1998 and Bergmann 2001 for
further objections to this third argument by Rowe; and see
Rowe 1998, 2001 for his replies.

17. In Schellenberg (2002: 51) this proposition is called ‘P2’.

18. To culpably be in a contrary position involves one’s own free

choice: ‘whether it is the free choice to ignore a God we are
aware of, or to take steps to remove that awareness, and so to
remove ourselves from that place where we are in a position to
relate personally to God’ (ibid. 42–3).

19. Pr(X/Y) >! Pr(X/Z) says that the probability of X given Y is

much greater than the probability of X given Z.

20. He gives the argument this formulation in Draper 2004: 45 n.

6. He defends something like this argument in Draper 1989.
21. See Schellenberg 1996 and Draper 1996, both of which make it
clear that they would like to avoid con ict with the skeptical
theist’s skepticism—presumably because they think there is
some merit to that skepticism.

22. Draper has made this sort of point to me in email

correspondence in July 2006. See also Schellenberg 1996:
456–9.

23. See Draper 1989: 17–18 and Schellenberg 2002: 42–52.

24. See Bernste in 1998: 155–6.

25. I should note that Bernstein himself makes no explicit appeal to

the Principle of Indi erence in laying out his objection, but I
think that is what’s lurking behind his reasoning.

26. Consider the following comment by Hájek (2003: sect. 3.1) on

how philosophers think about applications of the Principle of
Indi erence: ‘This brings us to one of the chief points of
controversy regarding the classical interpretation. Critics
accuse the principle of indi erence of extracting information
from ignorance. Proponents reply that it rather codi es the
way in which such ignorance should be epistemically managed
—for anything other than an equal assignment of probabilities
would represent the possession of some knowledge. Critics
counter-reply that in a state of complete ignorance, it is better
to assign vague probabilities (perhaps vague over the entire
[0, 1] interval), or to eschew the assignment of probabilities
 altogether.’ One further possibility to consider is that—because
of di erences in the amount or degree of ignorance in
particular cases—the Principle of Indi erence may be sensibly
applied in some cases of ignorance whereas eschewing the
assignment of probabilities altogether may be appropriate in
other cases (and the latter approach may be best in the cases
discussed by skeptical theists).

27. As becomes clear in his 1998, Swinburne himself thinks we can

think of God-justifying reasons for permitting evils such as
(E1) and (E2) so he doesn’t endorse the atheist’s argument.

28. These three points come out in Swinburne 1998: 20–8.

29. Remarks similar to those I’vemade in these last few paragraphs

apply to what William Hasker 2004—another theist who
rejects skeptical theism—says.

30. In correspondence, Swinburne pointed me to these remarks of

his when I mentioned to him the points raised in the previous
paragraph.

31. Gale 1996: 214 says something similar.

32. See e.g. Wilks 2004: 317–18.

33. Rowe 2001: 298.

34. From this he concludes that God would not permit animals to
su er lonely, horri c deaths since, he thinks, they cannot
bene t from them. See Tooley 1991: 111. Stump 1985, 1990
also endorses a principle like this, though, unlike Tooley, she
 maintains it while defending theism against arguments from
evil rather than using it to argue for atheism.

35. See van Inwagen 1988: 121–2 and Swinburne 1998: 223–36 for
some reasons to doubt it.

36. See Beaudoin 2005: 45 and Bergmann 2001: 295 n. 27.

37. My own account of how we know (E5) and (E6) aren’t actual is
discussed in Bergmann 2006: 206–11.

38. This charge is developed in a number of places. See Almeida

and Oppy 2003; Pereboom 2004; Hasker 2004; Russell 1996;
and Tooley 1991. For replies to this line of reasoning see
Bergmann 2001; Bergmann and Rea 2005; and Howard-Snyder
forthcoming.

39. Or at least this is so if we assume the falsity of the open theist’s

conception of God.

40. See Pereboom 2004: 164–5.

REFERENCES

ALMEIDA, MICHAEL, and OPPY, GRAHAM (2003). ‘Sceptical Theism and

Evidential Arguments from Evil’. Australasian Journal of
Philosophy 81: 496–516.

ALSTON, WILLIAM (1991). ‘The Inductive Argument from Evil and the
Human Cognitive Condition’, Philosophical Perspectives 5: 29–67;
 repr. in Howard-Snyder 1996b: 97–125. Page references are to
the reprint.

——— (1996). ‘Some (Temporarily) Final Thoughts on Evidential

Arguments from Evil’, in Howard-Snyder 1996b: 311–32.

BEAUDOIN, JOHN (2005). ‘Skepticism and the Skeptical Theist’, Faith

and Philosophy 22: 42–56.

BERGMANN, MICHAEL (2001). ‘Skeptical Theism and Rowe’s New

Evidential Argument from Evil’, Noûs 35: 278–96.

——— (2006). Justi cation without Awareness. Oxford: Oxford

University Press.

——— and REA, MICHAEL (2005). ‘In Defence of Sceptical Theism: A

Reply to Almeida and Oppy’, American Philosophical Quarterly
83: 241–51.

BERNSTEIN, MARK (1998). ‘Explaining Evil’, Religious Studies 34:151–

63.

DRAPER, PAUL (1989). ‘Pain and Pleasure: An Evidential Problem for

Theists’ Noûs 23: 331–50; repr. in Howard-Snyder 1996b: 12–29.
Page references are to the reprint.

——— (1996). ‘The Skeptical Theist’, in Howard-Snyder 1996b:

175–92.

DRAPER, PAUL (2004). ‘More Pain and Pleasure: A Reply to Otte’ in

Peter van Inwagen (ed.), Christian Faith and the Problem of Evil.
Grand Rapids: William B. Eerdmans, 41–54.
——— (unpublished). ‘Partisanship and Inquiry in Philosophy of
Religion’, presented as an invited symposium paper at the
Central Division Meeting of the American Philosophical
Association in Chicago, April 2006.

FLINT, THOMAS (1998). Divine Providence: The Molinist Account. Ithaca:

Cornell University Press.

GALE, RICHARD (1996). ‘Some Di culties in Theistic Treatments of

Evil’, in Howard-Snyder 1996b: 206–18.

HÁJEK, ALAN (2003). ‘Interpretations of Probability’, in Edward N.

Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer
Edition),
<http://plato.stanford.edu/archives/sum2003/entries/probabili
ty-interpret/>, accessed 22 May 2008.

HASKER, WILLIAM (2004). ‘The Sceptical Solution to the Problem of

Evil’, in William Hasker, Providence, Evil, and the Openness of
God. London: Routledge, 43–57.

HOWARD-SNYDER, DANIEL (ed.), (1996a). ‘The Argument from

Inscrutable Evil’ in Howard-Snyder 1996b: 286–310.

——— (1996b). The Evidential Argument from Evil. Bloomington:

Indiana University Press.

——— (forthcoming). ‘Does Skepticism about Arguments from Evil

Lead to Moral Skepticism?’ Oxford Studies in Philosophy of
Religion.
PEREBOOM, DERK (2004). ‘The Problem of Evil’, in William Mann
(ed.), The Blackwell Guide to the Philosophy of Religion. Oxford:
Blackwell, 148–70.

PLANTINGA, ALVIN (1967). God and Other Minds. Ithaca, NY: Cornell
University Press.

——— (1974). The Nature of Necessity. Oxford: Oxford University

Press.

——— (1998). ‘Degenerate Evidence and Rowe’s New Evidential

Argument from Evil’ Noûs 32: 532–44.

ROWE, WILLIAM (1979). ‘The Problem of Evil and Some Varieties of

Atheism’, American Philosophical Quarterly 16: 225–41; repr. in
Howard-Snyder 1996b: 1–11. Page references are to the reprint.

——— (1988). ‘Evil and Theodicy’, Philosophical Topics. 16:119–32.

——— (1991). ‘Ruminations about Evil’ Philosophical Perspectives 5:

69–88.

——— (1996). ‘The Evidential Argument from Evil: A Second Look’,

in Howard-Snyder 1996b: 262–85.

——— (1998). ‘Reply to Plantinga’, Noûs 32: 545–52.

——— (2001). ‘Skeptical Theism: A Response to Bergmann’, Nous

35: 297–303.

RUSSELL, BRUCE (1996). ‘Defenseless’ in Howard-Snyder 1996b: 193–

205.
SCHELLENBERG, JOHN (1996). ‘Response to Howard-Snyder’, Canadian
Journal of Philosophy 26: 455–62.

——— (2002). ‘What the Hiddenness of God Reveals: A

Collaborative Discussion’, in Daniel Howard-Snyder and Paul
Moser (eds.), Divine Hiddenness: New Essays. New York:
Cambridge University Press, 33–61.

STUMP, ELEONORE (1985). ‘The Problem of Evil’, Faith and Philosophy

2: 392–423.

——— 1990). ‘Providence and the Problem of Evil’, in Thomas Flint

(ed.), Christian Philosophy. Notre Dame: University of Notre
Dame Press.

SWINBURNE, RICHARD (1998). Providence and the Problem of Evil.

Oxford: Oxford University Press.

TOOLEY, MICHAEL (1991). ‘The Argument from Evil’ Philosophical

Perspectives 5: 89–134.

VAN INWAGEN, PETER (1988). ‘The Magnitude, Duration, and

Distribution of Evil’, Philosophical Topics 16: 161–87. Reprinted
in Peter van Inwagen, God, Knowledge, and Mystery. Ithaca:
Cornell University Press, 96–122. Page references are to the
reprint.

——— (1991). ‘The Problem of Evil, the Problem of Air, and the
Problem of Silence’ Philosophical Perspectives 5: 135–65; repr. in
Howard-Snyder 1996b: 151–74. Page references are to the
reprint.
WILKS, IAN (2004). ‘The Structure of the Contemporary Debate on the
Problem of Evil’, Religious Studies 40: 307–21.

WYKSTRA, STEPHEN (1984). ‘The Human Obstacle to Evidential

Arguments from Su ering: On Avoiding the Evils of
“Appearance” International Journal for Philosophy of Religion 16:
73–94.

——— (1996). ‘Rowe’s Noseeum Arguments from Evil’, in Howard-

Snyder 1996b: 126–50.
 PART IV

TOPICS IN CHRISTIAN PHILOSOPHICAL

THEOLOGY
 CHAPTER 18

THE TRINITY

MICHAEL C. REA

ONE of the central mysteries of the Christian faith concerns the tri-
unity of God. According to traditional Christian doctrine, God is
three persons who are somehow consubstantial—one in substance.
The persons are the Father, the Son, and the Holy Spirit. Each
person possesses all the traditional divine attributes—omnipotence,
omniscience, perfect goodness, eternality, and so on. And yet (in the
words of the Athanasian Creed), ‘they are not three eternals, but
there is one eternal …there are not three almighties, but there is one
almighty …there are not three Gods, but there is one God.’1 But
what does all of this really mean? And how could it possibly be
true?

In addressing these two very general questions, there are several

more speci c issues on which we might try to focus. One is
interpretative: how were the central terms in the doctrine—terms
such as person, substance, and consubstantial— understood when the
doctrine was rst formulated, and how have they evolved
throughout the history of the doctrine? Another is more
straightforwardly philosophical: how could three distinct persons (in
any reasonable sense of that term) be consubstantial in a way that
would make them countable as one God? These questions are not
wholly distinct from one another, and there are many others in the
neighborhood that are also worth pursuing. But the one on which I
will focus is the second. More exactly, I’ll explore various attempts
to show that the central statements in the doctrine, under some
intelligible and orthodox interpretation, do not imply a
contradiction. Along the way, I’ll touch a bit on the rst question,
but for the most part it will be set aside.2

The question that concerns us here is commonly referred to by

philosophers as ‘the logical problem of the Trinity’ and by
theologians as ‘the threeness-oneness problem’. The goal of the rst
section of this chapter is to explain in some detail just what the
problem is supposed to be. I’ll begin by stating the central theses of
the doctrine of the Trinity. Then I will formulate the logical problem
of the Trinity and lay out the constraints that a solution must satisfy
in order to preserve an orthodox understanding of the doctrine. In
the next section, I will brie y sketch a few of the most important
solutions to the problem. Finally, in sect. 3, I’ll present my own view
of the Trinity and argue that it has both better historical pedigree
and better prospects for solving the problem of the Trinity than the
rival views presented in sect. 2.

1. THE PROBLEM OF THE TRINITY

The two creeds to which contemporary Christians typically look for
‘o cial’ expressions of Trinitarian doctrine are the Nicene-
Constantinopolitan Creed of 380/1 and the Athanasian Creed
(c.500). The fourth and fth centuries witnessed a great deal of
philosophical re ection and controversy over the doctrine of the
Trinity, and the language of these two creeds is in large part a
product of that discussion.3 The former creed is a revised and
expanded version of the Creed of Nicaea, which was produced by
the First Nicene Council in AD 325. (Nowadays the Nicene-
Constantinopolitan Creed is typically called just the ‘Nicene Creed’.
I’ll follow this usage.) Like its predecessor, the Nicene Creed was
written in Greek and subsequently translated into Latin. It includes
the following words:

We believe in one God, the Father, almighty, maker of heaven and

earth, of all things visible and invisible; And in one Lord Jesus
Christ, the only-begotten Son of God, begotten from the Father
before all ages, light from light, true God from true God, begotten
not made, of one substance [homoousion] with the Father …And in
the Holy Spirit, the Lord and life-giver, Who proceeds from the
Father, Who with the Father and the Son is together worshipped and
together glori ed, Who spoke through the prophets.

(Schwartz 1914: 244–50; quoted in Kelly 1972: 297–8)

I have highlighted the Greek term ‘homoousion’ because that
term—the term we translate as ‘consubstantial’ or ‘of the same
essence’ or ‘of one substance’—was at the center of some of the most
important fourth-century debates about the doctrine of the Trinity.

The Athanasian Creed—named after, but not written by, the

famous fourth-century defender of Nicene orthodoxy, Athanasius—
was written in Latin, and includes the following statements:

Now this is the catholic faith, that we worship one God in Trinity,
and the Trinity in unity, without either confusing the persons or
dividing the substance. For the Father’s person is one, the Son’s
another, the Holy Spirit’s another; but the Godhead of the Father,
the Son, and the Holy Spirit is one, their glory equal, their majesty
co-eternal. Such as the Father is, such is the Son, such also the Holy
Spirit…. Thus the Father is God, the Son God, and the Holy Spirit
God; and yet there are not three Gods, but there is one God. Thus
the Father is Lord, the Son Lord, the Holy Spirit Lord; and yet there
are not three Lords but there is one Lord…. Because just as we are
obliged by Christian truth to acknowledge each person separately
both God and Lord, so we are forbidden by the catholic religion to
speak of three Gods or Lords. (Kelly 1964: 17–20, repr. in Leith
1972: 705)
The Athanasian Creed is widely regarded as manifesting a bias
toward ‘Latin’ theories of the Trinity (see sect. 2 below).

In the two passages just quoted, we have the three central tenets
of the doctrine:4

(T1) There is exactly one God, the Father almighty.5

 (T2) Father, Son, and Holy Spirit are not identical.

(T3) Father, Son, and Holy Spirit are consubstantial.

Each creed includes each of these tenets, if not explicitly, then by

implication. The Nicene Creed a rms (T1) in its rst line (though
the ‘exactly’ is only implied). It also explicitly a rms the Father–
Son components of (T2) and (T3). The distinctness and
consubstantiality of the Spirit are implicit in the remark that the
Holy Spirit is ‘together worshipped and together glori ed’ with the
Father and the Son. Moreover, they are explicitly a rmed in the
Synodical Letter written in 382 by the same council that produced
the Nicene Creed.6 The Athanasian Creed enjoins us not to confound
the persons or divide the substance, thus committing to (T2) and
(T3). It also a rms (T1): we worship one God, and the Father is
God. To be sure, it says a lot more than this (for example, it says
that the Son is God too); but the more that it says is pretty clearly
just a listing of some of the logical consequences of the conjunction
of (T1) and (T3).

The logical problem of the Trinity is just the fact that (T1)–(T3)
appear to be mutually inconsistent. There are various ways of trying
to demonstrate the inconsistency, but the one I favor focuses on the
(apparent) meaning of consubstantiality. To say that x and y are
consubstantial, or of the same substance is, it seems, just to say that
x and y share a common nature—i.e. they are members of one and
the same kind. To say that two divine beings are consubstantial,
then, would be to say that the two beings in question are identical
with respect to their divinity: neither is subordinate to the other; they
are not divine in di erent ways; and if one is a God, then the other
one is too. (Note, by the way, that ‘God’ functions in (T1) above as a
kind-term, like man, and not as a name, like Fred. Thus, though it
looks a bit odd, it makes perfect sense to speak of the Father as a
God. If ‘God’ were functioning as a name, then (T1) would be saying
something very much like ‘There is exactly one YHWH,’ which isn’t
so much a monotheistic claim as a rather strange way of asserting
the existence of YHWH.)

Given all this, the logical problem of the Trinity can be expressed
as follows:

(LPT1) There is exactly one God, the Father Almighty. (From

(T1))

(LPT2) The Father is a God. (From (LPT1))

(LPT3) The Son is consubstantial with but not identical to the

Father. (From (T2) and (T3))

(LPT4) If there are x and y such that x is a God, x is not identical

to y, and y is consubstantial with x, then it is not the case
that there is exactly one God. (Premise)

(LPT5) Therefore: It is not the case that there is exactly one God.
(From (LPT2), (LPT3), (LPT4))

***Contradiction
The only way out of the contradiction is either to give up one of the
tenets of the doctrine of the Trinity or to give up (LPT4).

At this juncture, some will wonder why Christians don’t just give
up on one of the tenets of the doctrine of the Trinity. After all, the
doctrine isn’t explicitly taught in the Christian Scriptures, and the
precise language in which it is expressed wasn’t settled until the
fourth century.7 So why not just abandon (say) (T2) or (T3)? What
reasons are there for accepting traditional Trinitarian doctrine?

Some have argued that the doctrine of the Trinity (or something
like it) can be established via a priori argument.8 But the main
reason Christians take themselves to be committed to (T1)–(T3) is
that they seem to be implied both by Christian practice and by
central claims in the Christian Scriptures.9 For example, both the
Old and New Testaments make it clear that there is only one being
who deserves worship and who deserves such titles as ‘God
Almighty’ or ‘the one true God’; and Jesus refers to this being as ‘our
heavenly Father’.10 Hence (T1). Moreover, though Jesus says things
such as ‘I and the Father are one’, it is clear that, from the point of
view of the New Testament, Jesus (the Son) and the Father are
distinct.11 Jesus prays to the Father; claims to submit to the Father’s
will; is blessed by the Father; and so on. Likewise, the Holy Spirit is
distinct from the Father and Son: the Spirit is sent by the Son and is
said to intercede for us with the Father. Hence (T2). And yet the
New Testament advocates worshipping Jesus (the Son) and the Holy
Spirit;12 we nd Jesus saying such things as, ‘Anyone who has seen
me has seen the Father’; and we nd the apostle Peter saying (of
someone who has lied to the Holy Spirit), ‘You have not lied to men,
but to God.’13 In short, there is pressure to say that the Son and
Spirit are divine—and not in some derivative, or degenerate sense,
but truly divine, like the Father. The only clear way to say this
without contradicting (T1), however, is to say that the Son and the
Spirit are consubstantial with the Father: the divinity of the Father,
which is the ‘substance’ of the Father (more on this later), is no
di erent from the divinity of the Son. Hence (T3).14

Thus we return to the challenge of unpacking the notion of

consubstantiality in a way that enables us to reject (LPT4) without
incoherence or heterodoxy. The next two sections will be devoted to
this task; but rst let me say a bit about the boundaries of
orthodoxy. Broadly speaking, there are three main ‘errors’ that the
church has condemned with respect to the doctrine of the Trinity:
subordinationism, modalism, and polytheism. The task, then, is to
explicate the doctrine in a way that avoids these errors. I’ll discuss
each in turn.

Subordinationism is the view that neither the Son nor the Spirit
is truly and fully divine. Either they are not divine at all, or their
divinity is somehow subordinate to that of the Father. They are gods
of a sort, but lesser gods. Subordinationism is ruled out by such
language as ‘true God from true God’ or, more explicitly, ‘the Father
is God, the Son is God, and the Holy Spirit is God’.
 Modalism is the view that the Father, Son, and Holy Spirit are
merely di erent aspects or manifestations of God—di erent modes of
appearance by which God makes himself known. If modalism were
true, then the terms ‘Father’, ‘Son’, and ‘Holy Spirit would be
analogous to terms such as ‘Superman’ and ‘Clark Kent’. The
substance called ‘Superman’ is strictly identical to the substance
called ‘Clark Kent’. But there is, nevertheless, a distinction to be
drawn. The Superman-disguise is di erent from the Clark-Kent-
disguise; and so it makes perfect sense to say that Superman and
Clark Kent are di erent manifestations of Kal El (the Kryptonian who
is both Superman and Clark Kent), or di erent modes in which Kal
El appears. If modalism is true, then precisely the same sort of thing
can be said about the terms ‘God the Father’, ‘God the Son’, and
‘God the Holy Spirit’. Insofar as they are distinct, Father, Son, and
Holy Spirit do not fall into the category of substance; rather, they
fall into the category of ‘aspect’ or ‘property’.

Polytheism is harder to characterize. According to standard

dictionary de nitions, ‘polytheism’ is the view that there are many
gods, and a ‘god’ is any divine being. Given that Father, Son, and
Holy Spirit are divine, these de nitions plus (T2) imply polytheism.
Father, Son, and Holy Spirit are three divine beings; so they are
three gods; so polytheism is true. If this is right, and if (T1) and (T3)
are together meant to rule out polytheism, then the doctrine of the
Trinity is incoherent. But, of course, it can’t be a simple matter of
dictionary-de nition that sinks the doctrine of the Trinity. The
standard de nitions require nuance. But how shall we modify them?
I suggest the following:

x is a god =df x is a divine substance.

polytheism = the view that there is more than one divine

substance.

I won’t insist on these de nitions here, though. For present

purposes, I’ll leave the terms o cially unde ned and invite readers
simply to consult their own intuitions about polytheism in making
decisions about whether various models of the Trinity have
managed to avoid it.

2. SOLVING THE PROBLEM

In this section, I want to lay out some of the main strategies for
solving the problem of the Trinity. Since the end of the nineteenth
century, it has been common to divide the landscape of views into
two camps: Latin (or Western) Trinitarianism (LT) and Greek (or
Eastern) Trinitarianism (GT).15 Those who divide the territory this
way tell roughly the following story about their classi catory
scheme: the Latin tradition traces its historical roots through the
western church. It is epitomized in the work of such theologians as
St Augustine and St Thomas Aquinas; it takes the unity of the
Godhead as given and seeks to explain the plurality in God (rather
than vice versa); and those in the tradition tend to gravitate toward
psychological analogies. The Greek tradition, on the other hand,
traces its roots through the eastern church; it is epitomized in the
work of the Cappadocian Fathers; it takes the plurality of the divine
persons as given and seeks to explain their unity; and those in the
tradition tend to favor social analogies.16 (GT is commonly
identi ed with ‘social trinitarianism’, discussed below.)

In recent years, the standard way of dividing the territory has

come under heavy attack, and I am inclined to reject it as well.17
Moreover, as we’ll see shortly, there is a lot more at stake in making
a decision about the standard story than the viability of a mere
heuristic device for classifying views about the Trinity. For purposes
of a handbook chapter, however, it seems prudent to start by
presupposing the standard classi catory scheme, present some of
the main views that fall under each heading, and only later subject
the standard scheme to criticism. So that is how I shall proceed.

In the two main parts of this section, I’ll present some of the
more well-known LT and GT models of the Trinity. Though each of
these models is intended to guide us toward a solution to the logical
problem of the Trinity, we’ll see that all of them fall short.
Moreover, I’ll argue that the most popular contemporary view,
social trinitarianism, depends heavily for its plausibility upon one of
the central claims involved in the LT-GT classi catory scheme:
namely, the claim that the Greek tradition, starting with the
Cappadocian Fathers, favored ‘social analogies’ as ways of
explicating the doctrine of the Trinity. This done, I’ll go on in sect. 3
to present my own view of the Trinity—a version of the so-called
‘Relative Identity’ solution (which I’ll also save for discussion in
sect. 3) that transcends the alleged LT-GT divide. I’ll also present
alternative readings of Augustine and the Cappadocian Fathers that
identify their views as ancestors of mine. Ultimately, my conclusion
will be that, of all the models considered herein, the view defended
in sect. 3 has the best claim to being orthodox and in accord with
the views of the earliest defenders of the Creed of Nicaea.

2.1 Latin Trinitarianism

In On the Trinity, St Augustine provides several di erent analogies,

or ‘images’ of the Trinity. According to one analogy, Father, Son,
and Holy Spirit are like the mind, its understanding of itself, and its
love for itself.18 Another—his preferred analogy—compares Father,
Son, and Holy Spirit with the mind’s memory of itself, the mind’s
understanding of itself, and the act(s) of will whereby the mind
obtains self-understanding from its own memory of itself and
delights in and makes use of what it remembers and understands.19
The usual glosses on these psychological models emphasize two
points (not always discussed together, or connected very clearly):
(1) The Son and the Spirit in the rst analogy, and the Father as
well in the second, are being compared to distinct faculties of the
mind, or to distinct ways in which the mind operates. (2) Both
analogies indicate that the di erence between at least two of the
persons is fundamentally a relational di erence: the persons are at
least roughly analogous to di erent ways in which a subject might
be related to itself, or di erent ways in which the mind relates to
objects of thought in general.20 As noted above, psychological
models of the Trinity are common in the work of thinkers falling
under the LT classi cation. So too is the idea that di erences among
persons are akin to di erences among re exive relations. Indeed,
according to Richard Cross (2002a: 287), the ‘vast consensus in the
West’ is that ‘the only distinguishing features among the persons are
their relations—that, in the standard terminology, they are
subsistent relations’. The view that the persons are subsistent
relations is a development of the Augustinian models,21 and it is
most notably associated with the name of St Thomas Aquinas.
According to Aquinas,

Distinction in God arises only through relations of origin…. But a

relation in God is not like an accident inherent in a subject, but is
the divine essence itself. So it is subsistent just as the divine essence
is subsistent. Just as, therefore, the Godhead is God, so the divine
paternity is God the Father, who is a divine person. Therefore,
‘divine person’ signi es a relation as subsistent. (Summa Theologiae I
q. 29 a. 4 c, quoted and translated in Cross 2002: 286)

What exactly it would mean for a relation to subsist, however, is

open to interpretation.

The usual objection to the views of Augustine and Aquinas

(based on the characterizations just o ered) is that they slip into
modalism. Granted, Augustine and Aquinas both succeed in drawing
distinctions among the persons of the Trinity. But in doing so, they
seem to locate the persons in the wrong category. On some
interpretations, the Augustinian view suggests that the persons are
to be identi ed with cognitive faculties, or mental modes of
operation, both of which are clearly mere aspects of a mind. But
even on somewhat more careful interpretations, the Augustinian
view at least suggests that the persons are to be identi ed with
relations. And the Thomistic view is explicit on that score. But
relations are commonly construed as properties of a certain kind
(polyadic properties, to be speci c) rather than as substances.22 But
if neither Father, Son, nor Holy Spirit is a substance—if they are
mere properties—then modalism is true.

Aquinas, at least, tries to locate the persons in the right category

by calling them ‘subsistent’ relations. But contemporary writers
typically respond to this idea with ba ement. What, really, could it
mean for a relation to subsist? Not having an answer, they then
proceed to object that Aquinas has simply identi ed the divine
persons with relations, which brings us back to modalism.23

Similar problems beset contemporary psychological analogies,

most of which can also be classi ed under the LT heading. For
example, Thomas V. Morris suggests (without a rming) that the
Trinity might be modeled on the di erent personalities of a patient
su ering from multiple personality disorder.24 Trenton Merricks, on
the other hand, suggests that the persons might be thought of on
analogy with the distinct ‘centers of consciousness’ that seem to be
associated with the two hemispheres of a human brain. (In
experimental situations, commissurotomy patients—people who
have undergone a surgical procedure that severs the bundle of
nerves that allows the two hemispheres of the brain to communicate
with one another—show behavior that seems to indicate that their
consciousness is divided, as if there is a separate stream of thought
associated with each hemisphere.) There is some pressure to classify
these models under the GT heading instead; for, as we’ll shortly see,
the standard lore about GT is that it is wedded to ‘social models’,
which, allegedly in contrast to Latin views, regard the persons of the
Trinity as distinct psychological subjects. My own inclination,
though, is to think that neither of them compares the persons of the
Trinity with distinct psychological subjects. Rather, each compares
the Trinity as a whole with a fragmented psychological subject. The
personalities of someone with multiple personalities are not
substances; they are aspects of a substance. Plausibly, the same is
true of the distinct ‘centers of consciousness’ that are elicited as a
result of commissurotomy.25 Thus, again, modalism looms.

The nal version of LT that I want to consider is Brian Leftow’s

(2004).26 Leftow regards his view as an instance of LT because it
‘begin[s] from the oneness of God’ (ibid. 304). On his view, the
persons of the Trinity might be thought of as analogous to a time
traveler who appears thrice located at a single time. He o ers us the
example of Jane, a Rockette who is scheduled to dance in a chorus
line but, at the last minute, discovers that two of her partners have
failed to show up. Jane goes on stage and dances her part, then later
enters a time machine (twice) so that she can (twice) go on stage
with herself and dance the leftmost and rightmost parts as well.
According to Leftow, there is a very clear sense in which this part of
the chorus line contains three of something; and yet there is just one
substance (Jane) in that part.

Is the view adequate? It is hard to tell, because Leftow’s

presentation is imprecise at a crucial juncture. Consider Jane’s part
of the chorus-line, and suppose we use the labels, ‘L’, ‘M’, and ‘R’ to
refer to the occupants of the three positions in that part of the line.
Now, let us ask what the relations are between L, M, and R. Upon
re ection, we face a problem: it is not at all clear what we mean by
the occupants of the three positions. One might think that there is just
one occupant, Jane, multiply located in three positions. In that case,
‘L’, ‘M’, and ‘R’ are all just names for Jane; and the relation between
L, M, and R is identity:

Case 1: L = M = R = Jane
Alternatively, we might think that what the names ‘L’, ‘M’, and ‘R’
refer to are three distinct events in the life of Jane. In that case,
there are, after all, three Rockettes, and each Rockette is an event:

Case 2: L = the leftmost dancing event; M = the middle dancing

event; R = the rightmost dancing event; and L ≠ M ≠ R

Which case does Leftow have in mind? Unfortunately, it looks as if

he has both in mind. He writes:

If (as I believe) Jane has no temporal parts, then not just a temporal
part of Jane, but Jane as a whole, appears at each point in the
chorus line, and what the line contains many of are segments or
episodes of Jane’s life-events. This may sound odd. After all,
Rockettes dance. Events do not. But what you see are many
dancings of one substance. What makes the line a line is the fact
that these many events go on in it, in a particular set of relations.
Each Rockette is Jane. But in these many events, Jane is there many
times over. (ibid. 308, emphasis mine)

The quoted passage says that ‘what there are many of’ is events; and
it speaks of Rockettes in the plural. Thus, Case 2 looks like the
correct interpretation. On the other hand, each Rockette is Jane.
Thus, Case 1 looks like the correct interpretation. But it is
impossible for both interpretations to be correct. So it is hard to
know what to make of what Leftow is saying here.
 Elsewhere, however, he says a bit more. ‘Perhaps’, he says, ‘the
triune Persons are event-based persons founded on a generating
substance, God’ (2007: 373–4). An event-based person is, roughly, a
person whose existence is constituted by the occurrence of an event:
what it is for the person to exist is for that event to occur in a
particular substance (ibid. 367–8). In the case of Jane, then, L, M,
and R are presumably supposed to be analogous to event-based
persons. Jane (the generating substance) exists in each of the three,
as Leftow says at the end of the quoted passage above; but she is not
strictly identical to any of the three. Likewise, God exists in each of
the event-based persons that together constitute the Trinity.

If we take Leftow (2007) as de nitive, then we have clarity on

the relation between L, M, R, and Jane: all are distinct. But it is still
hard to see how the view sheds light on the Trinity. Are L, M, and R
consubstantial? It is hard to tell. They are events (or event-based
things) involving a common substance; but that doesn’t guarantee
consubstantiality. Suppose I paint a wall red. We then have two
events: the wall’s becoming wet and the wall’s becoming red. But
these two events are not consubstantial, for they are not substances
that share a common nature. Indeed, they’re not substances at all;
and even if they were, what they share isn’t a nature. So too,
apparently, in the case of L, M, and R. Of course, Leftow might have
more to say on the subject; but until more is said, it is, again, hard
to know what to make of the view.27
 The reader will have noticed by now that, though I have o ered
several analogies under the heading of LT, I have not yet said how,
exactly, any of the models solves the problem of the Trinity. The
reason is simple: They don’t—at least not as they have been
interpreted here. The reason they don’t is that, though they o er
‘senses’ in which God is both three and one, they do not explain
how it is that numerically distinct consubstantial beings count as one
God. Leftow’s model aside, the distinct items in so-called Latin
models as they are interpreted here are pretty clearly not
consubstantial. Indeed, they are not substantial at all. And, as we
have seen, Leftow’s model is problematic for other reasons. This
defect is not shared by GT models, which provide straightforward
reasons for rejecting (LPT4), grounded in a clear account of the
consubstantiality of the persons (they are consubstantial by virtue of
being members of a common kind) and varying analogies aimed at
explaining why we cannot conclude (as (LPT4) e ectively does) that
distinct consubstantial divine beings would count as more than one
God. But these models have other problems, as we shall now see.

2.2 Greek Trinitarianism

In the contemporary literature on the doctrine of the Trinity, there

is a family of views that falls under the label ‘social trinitarianism’
(ST).28 Taking the LT-GT classi cation scheme for granted, ST is
normally regarded as co-extensive with GT. What set the
Cappadocian Fathers apart from the LT tradition, according to the
common lore, was precisely their endorsement of ST; and
contemporary versions of ST are basically just contemporary
developments in the GT tradition. Owing in large part to
misunderstandings that lead people to think of LT models as
modalistic, ST models of the Trinity tend to be more popular in the
contemporary literature. Critics, however, charge ST with
polytheism (for reasons I shall explain shortly). This brings to light
the fundamental import of the LT-GT classi cation scheme, and of
recent criticisms of it. The Cappadocian Fathers are universally
acknowledged to have played a vital role in the earliest defenses
and interpretations of the creed of Nicaea. They are among those
who helped to de ne orthodoxy; so it would be surprising, to say the
least, if whatever view they held turned out to be heretical. In short,
saying ‘The Cappadocian Fathers endorsed model M’ is, if true,
nearly decisive as a response to charges that model M is polytheistic
or otherwise heretical. Hence the importance of the contemporary
claim that there is such a thing as GT which (1) is found in the work
of the Cappadocian Fathers and other key eastern theologians, and
(2) di ers from LT by taking plurality rather than unity as its
starting point, as evidenced by their endorsement of ST models.
Historical challenges to the viability of the LT-GT classi cation
scheme take issue with each of these points. They deny that the
Cappadocians endorsed ST; they deny that there is reason to think
that the Cappadocians took di erent ‘starting points’ from key Latin
theologians such as St Augustine; and they deny that there is any
such thing as GT or LT as they are commonly conceived.29 If the
challenges are right, then contemporary criticisms of ST are
suddenly much more forceful; for not only does ST lose the backing
of the Cappadocians that is so vital to warding o those criticisms,
but some other view—whatever view the Cappadocians actually
endorsed—acquires their backing.

In the next section, I will present a view that I take to be a

contemporary development of what was held in common by (at least)
Augustine and the Cappadocian Fathers. But for now, I want to
focus on explaining what ST is and why it is inadequate as a way of
understanding the Trinity.

Contemporary social trinitarians have not been especially clear

about what the central tenets of their view are supposed to be.
Neither have their critics. First and foremost, ST theories are
identi ed by their reliance on analogies that compare the persons of
the Trinity to things that are numerically distinct but share a
common nature—usually rational creatures of some sort, like human
beings. These are called ‘social analogies’ because many of them
(though not all) at least imply, if not explicitly state, that the unity
among the divine persons is some sort of social unity: God is like a
family, or a perfectly uni ed community of rulers, or whatever. This
suggests that ST might fruitfully be characterized as committed to
the following central tenets:

1. Father, Son, and Holy Spirit are not numerically the same
substance. Rather, the persons of the Trinity are consubstantial
only in the sense that they share a common nature; and the
 sharing is to be understood straightforwardly on analogy with
the way in which three human beings share a common nature.

2. Monotheism does not imply that there is exactly one divine

substance. Rather, it implies at most only that all divine
substances—all gods, in the ordinary sense of the term ‘god’—
stand in some particular relation R to one another, a relation
other than being the same divine substance.

3. The persons of the Trinity stand to one another in the relation

R that is required for monotheism to be true.

Di erent versions of ST might then be distinguished in accord with

di erences over what relation R amounts to.

There are many candidates in the literature for being

monotheism-securing relations, but the most popular are the
following:

(a) Being parts of a whole that is itself divine.

(b) Being the only members of the only divine kind.

(c) Being the only members of the community that rules the
cosmos.

(d) Being the only members of a divine family.

(e) Being necessarily mutually interdependent, so that none can

exist without the others.

(f) Enjoying perfect love and harmony of will with one another,
unlike the members of pagan pantheons.
 Most social trinitarians in fact opt for a combination of these,
and most (but not all) of the combinations include at least (a), (b),
and (c). So, for example, Richard Swinburne (1994) focuses on the
fact that YHWH is a composite individual or society whose parts or
members stand in the relations identi ed in (e) and (f).30 But, of
course, he wouldn’t deny that they stand in (b) and (c) as well. J. P.
Moreland and William Lane Craig (2003) focus primarily on (a). On
their view, YHWH is composed of the Persons in a sense analogous
to the way in which the three-headed dog Cerberus, guardian of the
underworld in Greek mythology, might be thought to be composed
of three ‘centers of consciousness’ (ibid. 593). On their view, the
three conscious parts of Cerberus are not dogs; there is only one dog
—Cerberus. But the centers of consciousness are canine, just as any
other part of Cerberus is (derivatively) canine. One dog, then; three
derivatively canine individuals. Likewise in the Trinity: one God;
three derivatively divine individuals. Monotheism is thus secured by
the fact that the Persons are parts of a single fully divine being.
Cornelius Plantinga Jr. (1989: 31), on the other hand, argues that
social Trinitarians may ‘cling to respectability as monotheists’
simply by a rming that the persons are related in the ways
described by (b), (c), and (d). His idea seems to be that monotheism
is true, no matter how many gods there are, so long as all gods
derive their divinity from one source, or share a single divine nature
(as we humans share a single human nature), or are joined together
as a divine family, monarchy, or community. There are other
suggestions in the literature; but they tend to run along very similar
lines.

A standard criticism of social trinitarianism is that these

(re-)characterizations of monotheism are implausible. Those lodging
the criticism typically do so in one of two ways. Some try to argue
that statements in the Nicene and Athanasian Creeds that are
commonly taken to rule out polytheism also speak against social
trinitarianism.31 But from the point of view of those who accept the
standard LT-GT distinction, there are two strikes against this
objection. First, the Athanasian Creed was written in Latin and
incorporates language from Augustine. So it can be dismissed as an
LT treatise—one which we would antecedently expect to re ect an
anti-ST bias. Second, the chief defenders of the creed of Nicaea were
themselves social trinitarians (according to the common lore); so the
claim that ST violates that creed is simply not credible. So, at this
stage anyway, the credal objection is indecisive. Others object to ST
on the grounds that it is not plausible to think that (say) Greek
polytheism would become monotheistic if only we added that Zeus
and the other gods enjoyed perfect love, harmony, and mutual
interdependence with one another; nor would it seem to help if we
were to pare the pantheon down to a single divine family that rules
the cosmos.32 I think that this objection is intuitively decisive
against the suggestion that the relations described in (b), (c), (d),
(e), or (f), singly or in combination, could possibly secure
monotheism. But it has no implications for the more popular
suggestion that Christianity is monotheistic because (a) the persons
are parts of the one and only fully divine being. Moreover, it has the
awkward implication that, if the Cappadocians really were social
trinitarians, then they were polytheists after all. To close the case
against ST, then, two tasks must be accomplished: (1) part-whole
models must be shown to be problematic; (2) one must either
explain how the Cappadocians managed inadvertently to fall into
one of the very errors they were most concerned to avoid, or one
must show that the Cappadocians really were not social
trinitarians.33

Various objections against part-whole trinitarianism are already

present in the literature.34 Rather than summarize those, however, I
want here to present a new one.35 I will take as my target the
version of part-whole trinitarianism developed by J. P. Moreland
and William Lane Craig (2003). I do this because theirs is the most
developed version of part-whole trinitarianism currently available,
but I think that substantially the same objection could be raised
against any version.

As I have already noted, Moreland and Craig compare God to the

mythical dog Cerberus and the persons to Cerberus’s three heads
(or, better, the souls that might be embodied in those heads). So, on
their view, the divine persons—all of them—are parts of God. For
obvious reasons, however, Moreland and Craig want to preserve the
view that God is divine while denying that that God is a fourth
divine thing on a par with the persons. Thus, they distinguish two
kinds of divinity: the full divine nature, which is possessed by God
and implies tri-unity; and a derivative divine nature, possessed by
each person. The distinction is analogous to the (allegedly) two
ways in which something might be said to be feline: something can
be feline by being a cat; or something can be feline in a derivative
way by being part of a cat. Moreover, the making of this sort of
distinction is implicit in every version of part-whole trinitarianism
of which I am aware. Nobody wants to be pressed into a rming a
quaternity in God; thus the composite is always treated, at least
implicitly, as a di erent kind of thing from the parts—a non-person
composed of persons, or a group-mind composed of single minds,
etc. So much for the view itself.

But now we face two problems, both devastating: rst, Moreland

and Craig cannot a rm the opening line of the Nicene creed: ‘We
believe in one God, the Father, almighty.’ For, on their view, God is
a fundamentally di erent thing from the Father.36 Moreover, they
cannot a rm the crucial homoousion clause in the same creed unless
they reject the idea that there is exactly one divine nature—an idea
that the Cappadocians and (so far as I am aware) every other major
interpreter and defender of the creed of Nicaea in the fourth and
fth centuries were in agreement about. Here is why: the only
viable interpretations of the credal claim that the Son is homoousion
with the Father have it that the Son is either numerically the same
substance as the Father or of the same nature as the Father. (Natures
were also referred to as ‘substances’; hence, being consubstantial
with something might just mean having the same nature.) The
former, of course, they reject. The latter they accept; but in
accepting it, they posit, e ectively, two divine natures—one
‘genuine’, possessed only by God; the other derivative, but still
divine, possessed by the two persons. Of course, they could deny
that the derivative nature is a divine nature. But in so doing, they
seem to strip the persons of their divinity, which would con ict with
other parts of the Nicene and Constantinopolitan creeds. If all this is
right, then, part-whole trinitarianism is in serious trouble, at least if
its proponents intend (as Moreland and Craig do) to be o ering a
view that respects Nicene orthodoxy.

But now what of the Cappadocian Fathers? The initial

justi cation for treating the Cappadocian Fathers as social
trinitarians comes primarily from their own apparent employment
of social analogies.37 Under the usual interpretation, the
Cappadocian view was that, ‘just as Peter, Paul, and Barnabas are
each man, so the Father is God, the Son is God, and the Holy Spirit
is God’ (Plantinga 1986: 329–30). In short: the persons of the Trinity
are ‘of the same substance’ only in the sense that they share a
common nature; and the sharing is to be understood
straightforwardly on analogy with the way in which three men
share a common nature.

As I have already indicated, however, this way of interpreting

the Cappadocian Fathers has, of late, come under heavy criticism.
Space will not permit me to present the exegetical case against the
standard interpretation; but in the next section, after laying out the
view that I call Constitution Trinitarianism, I will present reasons
for thinking that the Cappadocians were actually gesturing toward
that view rather than toward ST. If this is right, then one of the
most important reasons for accepting ST is undermined.

3. THE CONSTITUTION MODEL

I turn now to the model that I prefer.38

Let me begin here by identifying a general strategy for solving

the problem of the Trinity that I have so far omitted from our
discussion: the Relative Identity strategy.39 The problem, as I have
characterized it, is generated by the conjunction of the following
premise (from sect. 1 above) with the central tenets of trinitarian
doctrine:

(LPT4) If there are x and y such that x is a God, x is not identical

to y, and y is consubstantial with x, then it is not the case
that there is exactly one God. (Premise)

But (LPT4) is true only if the following principle is also true:

(P1) Necessarily, if x and y are not identical, then x and y are not
numerically the same substance.

If (P1) is false, then (LPT4) simply ignores the possibility that x and
y are distinct but (perhaps by virtue of their consubstantiality) one
and the same God. In other words, (LPT4) presupposes that it is
impossible for an object a and an object b to be numerically the
same F without being absolutely identical. Give up that
presupposition, and the argument that depends on (LPT4) fails.

But how can we reject (P1)? The way to do it is to endorse the

doctrine of relative identity. There is a weak version of this doctrine,
and a strong version. The weak version says:

(RI1) States of a airs of the following sort are possible: x is an F,

y is an F, x is a G, y is a G, x is the same F as y, but x is not
the same G as y.

The strong version is just the weak version plus (RI2):

(RI2) Either absolute (classical) identity does not exist, or

statements of the form ‘x = y’ are to be analyzed in terms
of statements of the form ‘x is the same F as y rather than
the other way around.

RI2 is not needed for solving the problem of the Trinity; but some
philosophers—notably, Peter Geach—endorse it for other reasons,
and it serves as independent motivation for RI1.40

Defenders of the relative identity solution have mostly occupied

themselves with working out the logic of relative identity in an e ort
to show that the doctrine of relative identity itself is coherent, and
to show that the doctrine of the Trinity can be stated in a way that
is provably consistent given the assumption of relative identity.41
(What I have done here falls short of the latter goal. I have shown
how accepting (RI1) enables one to rebut a single argument for the
inconsistency of the doctrine of the Trinity; but I have not shown that
every argument against the consistency of the doctrine as I have
stated it must fail.)

Despite the e orts of its defenders, however, the relative identity

solution has remained rather unpopular. (RI2) in particular is widely
rejected as implausible; and I have argued elsewhere (Rea 2003)
that invoking it in a solution to the problem of the Trinity implies
that the di erence between the persons is theory-dependent, and so
merely conceptual. But without (RI2), (RI1) is (at rst glance,
anyway) unintelligible. The reason is simple: sameness statements
are naturally interpreted as identity statements. So, the claim that ‘x
and y are the same F’ seems logically equivalent to the claim that ‘x
is an F, y is an F, and x = y’. (RI1) is inconsistent with this analysis
of sameness statements. But on its own, it doesn’t supply any
replacement for that analysis. Thus, it renders sameness claims
utterly mysterious. Appealing to (RI1) without a supplemental story
as a way of solving the problem of the Trinity, then, simply replaces
one mystery with another. That is hardly progress.

The constitution view supplies the relevant supplemental story.

The story begins with an example: an artistic building contractor
fashions a marble statue that is to be used as a pillar in the building
he is constructing. So he has made a statue; he also has made a
pillar. It would be strange to say that he has made two material
objects that are simply located in exactly the same spot at the same
time (though many philosophers do in fact say such a thing). What
we are inclined to say is that the statue and the pillar are one and
the same material object, not two. And yet they are distinct. Surface
erosion will destroy the statue without destroying the pillar. Internal
corruption that preserves the surface but undermines the statue’s
capacity to support the weight of a building will destroy the pillar
but (if the statue is removed from its position as a load-bearing
structure) will not destroy the statue. Thus, what we want to say is
that the statue and the pillar are the same material object, even
though they are not identical. If we do say this, we commit ourselves
to (RI1). But we can make (RI1) intelligible by adding that all it
means to say that two things are the same material object at a time
is that those two things share all the same physical matter at that
time.

Let us esh out the story just a bit further. Aristotle maintained
that every material object is a compound of matter and form. The
form might be thought of as a complex organizational property—not
a mere shape, as the term suggests in English, but something much
richer. For Aristotle, the form of a thing was its nature; and forms,
like concrete things (though not in exactly the same sense), count as
substances. Thus, on his view, St Peter would be a compound whose
constituents were some matter and the form, humanity, or human
nature. St Paul would be a compound whose constituents were that
same form, but di erent matter.42 Peter and Paul would thus be ‘of
one substance’ on the Aristotelian view; though (unlike Gregory of
Nyssa) Aristotle would not have spoken of Peter and Paul as being
numerically the same man, nor would he have regarded them as
numerically the same substance. With some minor modi cations
(plus the non-Aristotelian assumption that statues and pillars are
substances, just like men are) Aristotle’s view would permit us to say
that the statue and the pillar are numerically the same substance,
even though they would not be ‘of the same substance’ since they
would not share the same form, or nature. They would be
numerically the same substance, one material object, but distinct
matter-form compounds. They would be the same without being
identical.

In the case of our statue, then, we have two complex properties—

being a statue and being a pillar—both had by the same underlying
subject, some undi erentiated matter.43 This gives us two
compounds, a statue and a pillar. Each is a substance. Thus, the
statue and the pillar are emphatically not mere aspects of a common
substance. They are not properties or relations or anything of the
sort. Furthermore, each is distinct from the other. But they are,
nevertheless, the same substance. Note too that the underlying
matter is not a substance, since it is not a matter-form compound.
But even if it were, it would not be a fourth substance—it would be
the same substance as the statue and the pillar (by virtue of sharing
all its matter in common with them). Nor (obviously) would it be a
third compound. Hence, there is just one substance and two
compounds.
 You might think that if there is just one substance, then we
ought to be able to ask whether it—the one substance—is essentially
a mere statue or essentially a mere pillar (or perhaps essentially a
statue-pillar). But this thought is incorrect. On the present view,
terms such as ‘it’ and ‘the one substance’ are ambiguous: they might
refer to the statue, or they might refer to the pillar. For, again,
statue and pillar are distinct, though not distinct substances.

You might think that if there are three matter-form compounds,

then there are three ‘primary substances’ but only one ‘secondary
substance’ (where a primary substance is a concrete particular, like
a human being, and a secondary substance is a nature, or form). But
this is also incorrect. On the view I am defending, if x and y share
the same matter in common, and x is a primary substance and y is a
primary substance, then x and y are the same primary substance,
despite being di erent matter-form compounds.44

By now the relevance of all this to the Trinity should be clear:

almost everything that I just said about the statue and pillar could
likewise be said about the divine persons. In the case of the divine
persons, we have three properties—being the Father, being the Son,
and being the Spirit; or, perhaps, being Unbegotten, being Begotten, and
Proceeding—all had by something that plays the role of matter. (It
can’t really be matter, since God is immaterial. Suppose, then, that it
is the divine substance, whatever that is, that plays the role of
matter.)45 Each divine person is a substance; thus, they are not mere
aspects of a common substance. Furthermore, each is distinct from
the other. (So modalism is avoided and (T2) is preserved.) But they
are nevertheless the same substance. (Hence (T3).) Thus, there is
just one divine substance, and so the view allows us to say, along
with the creed, ‘We believe in one God, the Father, almighty…and
in one Lord Jesus Christ …begotten, not made, being of one
substance with the Father.’ Since Father, Son, and Holy Spirit count,
on this view, as numerically the same substance despite their
distinctness, (LPT4) is false, and the problem of the Trinity is solved.

So far so good; but there is one loose end that remains to be tied.
What is it that plays the role of matter in the Trinity? And is it a
substance itself? Here I want to o er only a partial view that might
be developed in a variety of di erent ways. What plays the role of
matter in the Trinity is the divine nature; and the divine nature is a
substance. It is not a fourth substance, for reasons already discussed;
nor is it a fourth person (since it is not a compound of ‘matter’ plus
a person-de ning-property). But it is a substance, since (again,
taking cues from Aristotle) natures are substances. What I don’t want
to take a position on here is the question of what, exactly, a nature
is. Is it concrete or abstract? Is it particular or universal? Is it a
property or something else? These questions I will not answer. I
think that they must be answered in a way that allows the divine
persons to be concrete particular non-properties; but I think that
there are various ways of answering these questions that are
compatible with that view.
 This completes my presentation of the constitution view. We are
now in a position to see how the view connects with the views of
fourth- and fth-century defenders of Nicaea. At the heart of the
constitution view is the idea that the divine persons are compounds
whose constituents are a shared divine nature, which plays the role
of Aristotelian matter, and a person-de ning property (like being the
Son, or being Begotten) that plays the role of form. But this is almost
exactly the view that Richard Cross identi es as the fundamental
point of agreement between eastern and western views of the
Trinity. According to Cross, east and west agreed that (a) the divine
nature is a property, and (b) one and the same divine nature is a
constituent of each of the divine persons—i.e. it is the point at
which they overlap.46

So far so good; but mere overlap is not su cient to suggest the

constitution model. For if nature-sharing is a kind of overlap (as it
seems to be for the Cappadocians, at least),47 then two men overlap
on a constituent as well. But, of course, they don’t constitute one
another. In order to attribute something like a constitution model to
Augustine and the Cappadocians, we would have to show that their
view posited not only overlap among the persons but a kind of
overlap that guaranteed the sort of absence of separation that we nd
in objects (like the statue and the pillar) that constitute one another.
Encouragingly, Michel Barnes insists that this idea was as crucial to
both east and west as Cross thinks the idea of overlap was. Thus, he
writes (speaking of both Augustine and the Cappadocians), ‘the most
fundamental conception and articulation in “Nicene” Trinitarian
theology of the 380s of the unity among the Three is the
understanding that any action of any member of the Trinity is an
action of the three inseparably’ (Barnes in Davis, Kendall, and
O’Collins (eds.), 1999: 156, emphasis in original).48 Moreover, as
we’ll see, both Augustine and the Cappadocians endorsed models
that reinforce this reading and the Cappadocians, at least, rejected
models that con ict with it.

Let us return again (for starters) to Augustine’s favored analogy:

self-memory, self-understanding, and self-directed willing. At the
end of Book 10 of On the Trinity, after introducing the favored
analogy, Augustine raises some concerns and then says that the
‘discussion demands a new beginning’. In Book 11, he opens with
two new analogies, the rst of which compares the divine persons
with an external body, vision of the body, and the attention of mind
that joins these two. If Augustine’s goal really were to model the
Trinity on di erent faculties or modes of operation for the human
mind, returning to an analogy like this would be bizarre. For, after
all, an external body has nothing essentially to do with the mind of
one who beholds it. But, of course, this isn’t really Augustine’s goal
at all. The key to understanding this is to attend carefully to the
theory of vision that gets presented along with the analogy. For
Augustine, vision involves a kind of reproduction in the viewer of
the nature or form of the thing viewed.49 So when an external body
is seen, the form of that body is present both in the body and in the
visual faculty of the viewer. Moreover, he seems to think that, in
directing vision toward the external body and retaining the image of
it in the visual faculty, the will takes on the same nature as well—
just as ‘the little body of a chameleon [varies] with ready change,
according to the colors which it sees’.50 Once we see this, however,
the point of the new analogy becomes clear. The case at hand is one
in which three very di erent items are uni ed by virtue of a kind of
overlap: they have (in a way) the very same nature present in them.

Of course, the analogy is imperfect; for, among other things, the

external body’s nature is present in it in a way di erent from the
way in which it is present in the viewer’s visual faculty or will.
Augustine therefore o ers another analogy to correct this defect;
but, in the end, the best analogy is the one that I have called ‘the
favored analogy’. It is the best because, in that analogy, one and the
same substance is present in three distinct things which are
distinguished from one another only by their relations to that one
substance. The things in question are not faculties or modes of
operation. They are, rather, concrete events, each of which is, in
e ect, a complex whose constituents are a substance and a re exive
relation: the mind remembering itself, the mind understanding itself, and
the mind willing itself. These events are not identical; and, as events,
they are at least closer to the category of substance than to the
category of property or aspect. Moreover, since it is the mind itself
that is the common constituent in each complex, and since it is the
mind that is doing the self-remembering, self-understanding, self-
willing, the view strongly suggests the sort of absence of separation
that one nds in the constitution view; and it even suggests shared
agency among the persons.

What of the Cappadocians? The Cappadocians clearly thought of

consubstantiality as the sharing of a nature—on this I think there is
no substantive disagreement. Moreover, it seems clear that they
thought of nature-sharing as the sharing of a common constituent.
Gregory of Nyssa claims (in On Not Three Gods: To Ablabius) that
‘the man in [a group of men] is one’ (Scha & Wace 1900/1999: v.
332). Likewise, in his Fifth Theological Oration: On the Holy Spirit,
Gregory Nazianzen says that ‘the Godhead [or divine nature] is
One’, and he speaks of the Persons as those ‘in Whom the Godhead
dwells’, which suggests that the Godhead—i.e. the divine nature—is
a common constituent of the three (ibid. vii. 322). Moreover, in the
following passage he seems to indicate outright that the relation
between the persons involves overlap on a common nature and
di erentiation by three properties:

the very fact of being Unbegotten or Begotten, or Proceeding has

given the name of Father to the First, of the Son to the Second, and
[to] the Third …of the Holy Ghost, that the distinction of the Three
Persons may be preserved in the one nature and dignity of the
Godhead…. The Three are One in Godhead, and the One three in
properties.

(Fifth Theological Oration, IX, ibid. 320)

 But what of inseparability? I said earlier that not only do the
Cappadocians a rm models that posit the inseparability of the
persons, but they reject models that con ict with it. Thus, though
Gregory Nazianzen acknowledges that Adam, Eve, and Seth are
consubstantial, he goes on to make it very explicit that his point in
saying that they are is only to show that distinct things can be
consubstantial, not to embrace what would now be called a social
analogy:

were not [Adam, Eve, and Seth] consubstantial? Of course they

were. Well then, here it is an acknowledged fact that di erent
persons may have the same substance. I say this, not that I would
attribute creation or fraction or any property of body to the
Godhead . . . but that I may contemplate in these, as on a stage,
things which are objects of thought alone. For it is not possible to
trace out any image exactly to the whole extent of the truth.

(Fifth Theological Oration, XI, ibid. 321)

Elsewhere, he is even clearer about the importance of inseparability,
and the inadequacy of the analogy with three human beings:

To us there is One God, for the Godhead is One …For one [Person]
is not more and another less God; nor is One before and another
after; nor are They divided in will or parted in power; nor can you
nd here any of the qualities of divisible things; but the Godhead is,
to speak concisely, undivided in separate Persons.…[D]o not the
Greeks also believe in one Godhead, as their more advanced
philosophers declare? And with us, Humanity is one, namely the
entire race; but yet they have many gods, not One, just as there are
many men. But in this case the common nature has a unity which is
only conceivable in thought; and the individuals are parted from
one another very far indeed, both by time and by dispositions and
by power. For we are not only compound beings, but contrasted
beings, both with one another and with ourselves; nor do we remain
entirely the same [over time].

(Fifth Theological Oration XIV, XV, ibid. 322)

As I see it, the upshot of these passages is that the analogy with
three humans breaks down at precisely the point where the social
model and the constitution model di er. Three humans are separated
in matter, time, space, and agency; the three Persons are not.

There are other indications in both Gregory of Nyssa and

Gregory Nazianzus that their view of the Trinity was along the lines
of the constitution model.51 I want to close, however, by dismissing
an argument for the claim that they rejected the constitution model.
In Epistle 52, Basil of Caesarea writes:

For they maintained that the homoousion set forth the idea both of
essence and of what is derived from it, so that the essence, when
divided, confers the title of co-essential on the parts into which it is
divided. This explanation has some reason in the case of bronze and
coins made therefrom, but in the case of God the Father and God
the Son there is no question of substance anterior or even
underlying both; the mere thought and utterance of such a thing is
the last extravagance of impiety. (ibid. viii. 155)

Commenting on this passage, William Lane Craig (2005: 80) takes

Basil to be ‘vehemently’ rejecting constitution models outright. But
this seems to me to be exactly the wrong interpretation. Though
constitution theorists would certainly want to say of each coin that
it is constituted by a piece of bronze, Basil is not here considering a
case in which multiple bronze items constitute one another. Indeed,
what Craig does not recognize is that the coins in this example are
consubstantial (if at all) in the way that Father, Son, and Holy Spirit
would be on the social trinitarian view. So, if any contemporary
model is being rejected, it is that one (the one Craig himself favors)
and not the constitution view. Moreover, what is at issue for Basil
here is just the question whether there is an anterior substance
underlying the Persons. But the constitution model is not committed
to an anterior substance. An anterior substance would be one that is
prior to and not the same substance as the persons. But, on the
constitution view, the divine nature is neither.

If the foregoing is right, then, the Cappadocians did not in fact

favor anything like social trinitarianism; rather, their view was
much more in line with constitution trinitarianism. Likewise,
Augustine did not favor the modalistic-sounding view that was
attributed to him in sect. 2.1; rather, he too was defending a view in
line with the constitution view. This is substantial support; and, to
my mind, the constitution view wins out on purely intuitive grounds
as well.

NOTES

I am grateful to Je Brower and Tom Flint for very helpful

comments on an earlier draft of this chapter.

1. Quicumque vult (The Athanasian Creed); trans. Je rey Brower,

quoted from Brower and Rea 2005a: 488.

2. The most important and illuminating contemporary literature

aimed at sorting out these di culties has been written by
theologians working in the eld of Patristic studies. For a start
into this literature, the following sources are especially useful:
Ayres 2004; Barnes 1998; Coakley 1999; Coakley (ed.) 2003;
Stead 1977; Turcescu 2005; and Wolfson 1964.

3. For details on the relevant controversies, see especially Ayres

2004.

4. It is more common nowadays to claim that the central tenets of

the doctrine are these three: (1) there is exactly one God; (2)
the Father is not the Son, the Son is not the Spirit, and the
Father is not the Spirit; and (3) the Father is God, the Son is
God, and the Spirit is God. This way of characterizing the
doctrine more closely follows the Athanasian Creed. But the
language of (3) is less clear than the language of (T3) above
 (due to the fact that the predicate ‘is God’ can be, and has
been, assigned a variety of di erent meanings). Just as
importantly, this formulation obscures the centrality of the
notion of consubstantiality in the doctrine—the very notion that
lay at the center of some of the most important controversies
surrounding the First Nicene Council. My own formulation of
the doctrine is more in accord with formulations found, for
example, in the systematic theologies of Berkhof (1938/1996)
and Hodge (1873).

5. I take this straight from the rst line of the Nicene Creed. The
rst line can be translated in ways di erent from what I have
reproduced here; but every credible way of translating it
strongly suggests that the Father is somehow the same as God.

6. Kelly 1972: 340–1; Scha and Wace 1900/1999: xiv. 189.

7. The common story is that it was settled in 325 and then

challenged by Arian rebels. Lewis Ayres (2004) and Michel
Barnes (1998) tell a di erent story, however—one according
to which the language wasn’t entirely settled in 325, but
rather later in 380/1.

8. The best-known argument along these lines is one that is at

least latent in the thought of a variety of medieval theologians,
explicit in the work of Richard of St Victor, and developed in
detail most recently by Richard Swinburne 1994: ch.8.
9. As the fourth-century defenders of Nicaea were at pains to
show. See, for starters, Athanasius’s ‘Defence of the Nicene
Council’ in Scha and Wace 1900/1999: iv.

10. See e.g. Exod. 20: 3–5; Isa. 42: 8; Matt. 5: 9–13, 7: 21; and
John 2: 16.

11. Compare John 10: 30; Matt. 24: 36; Luke 22: 42; and John 1:
14, 18.

12. e.g. Phil. 2: 10–11.

13. John 14: 9; Acts 5: 3–4 (NIV).

14. Obviously the discussion here is highly compressed. For a

fuller discussion of the biblical case for the doctrine, see (e.g.)
O’Collins 1999: chs. 1–4.

15. According to Michel Barnes, what I am calling the ‘standard’

classi catory scheme takes its cues from the work of the
nineteenth-century theologian, Theodore de Régnon (1892–8).
(Though Ayres notes that de Régnon’s own classi catory
scheme wasn’t so much a division between Greek and Latin
models as between patristic and scholastic models. See Ayres
2004: 302 .) Barnes subjects de Régnon’s scheme and its
relatives to trenchant criticism in Barnes 1995a, b.

16. Cf. e.g. Brown 1985; Cary 1995; Cross 2002; La Cugna 1986,
1991; and Placher 1983: 79.

17. See e.g. Ayres 2004; Barnes 1995a, b, 1998; Cary 1995;
Coakley 1999; and Cross 2002. The idea isn’t that there is no
 di erence whatsoever between eastern and western, or Latin
and Greek, conceptions of the Trinity, but that the di erences
aren’t nearly as sharp as they are commonly construed, that
they aren’t aptly characterized as di erences over ‘starting
points’ or as a fundamental di erence in attitudes toward
‘social’ and ‘psychological’ analogies.

18. On the Trinity, Bk. 9 chs. 3 .; Scha and Wace

1887/1999:127 .

19. On the Trinity, Bk. 10 chs. 10–12 and Bk. 15 ch. 3; Scha and
Wace 1887/1999: iii. 140–3, 200–2. He characterizes the
latter analogy as a ‘more subtle’ treatment of the matter than
the former.

20. Cf. Brown 1985: ch. 7; LaCugna 1991: ch. 3; O’Collins 1999:
ch. 7; Richardson 1955.

21. Moreland and Craig 2003: 585 say that Aquinas ‘pushes the
Augustinian analogy to its apparent limit’. But this makes it
sound as if the ‘subsistent relations view’ isn’t present in
Augustine. I myself don’t mean to suggest that, however. All I
mean to suggest is that it is not explicit in Augustine.

22. A polyadic property is one that applies to several things

together. A property such as being loved is monadic: it applies
to a single subject (even if there are many subjects to which it
applies). Relations such as—loves—, or—thinks that—is —, are
polyadic. The former applies to a pair, the latter to a triad.
 Note, however, that, though it is nowadays common to treat
relations as polyadic properties (as e.g. in van Inwagen 2004),
Je Brower argues that medieval philosophers did not treat
them as such. See Brower 2001 and 2005.

23. I think that the objection can be met, and that the way to do
it is to develop Aquinas’s view along the lines of the view
described in sect. 3. But I won’t attempt to do that here.

24. Morris 1991: ch.9.

25. Admittedly, though, one might think of them as distinct

(subsistent) minds or souls, embodied in or emergent upon the
same material substance. If we think of them this way, then, as
Je Brower and I note elsewhere (Brower and Rea 2005b)
polytheism looms, rather than modalism. Whether the view
actually falls into polytheism, however, depends entirely upon
whether the relevant ‘centers of consciousness’ are in any
sense the same substance. On this question, Merricks is silent.

26. See also Leftow 2007.

27. Thanks to Brian Leftow for helpful correspondence about his

view. Naturally, if I have him wrong here, that is my fault and
not his. My own view is that Leftow’s view can be made
workable by developing it in the direction of the constitution
view, described in sect. 3. But I won’t attempt that here.

28. Proponents of the ST strategy include Timothy Bartel (1993,

1994); David Brown (1985, 1989); Peter Forrest (1998); C.
 Stephen Layman (1988); J. P. Moreland and William Lane
Craig (2003); Cornelius Plantinga, Jr. (1986, 1988, 1989);
Richard Swinburne (1994); Edward Wierenga (2004); and C. J.
F. Williams (1994). Hasker (1970) claims to endorse ST; but
some of what he says seems more in line with the constitution
view (see esp. pp. 27–30). According to Richard Swinburne,
the ‘most in uential modern statement of social
Trinitarianism’ is Moltmann 1981.

29. See n. 17 for references.

30. Cf. Swinburne 1994: 181.

31. See esp. Brower 2004a; Clark 1996; Feser 1997; and Leftow
1999.

32. See esp. Leftow 1999 and Merricks 2006.

33. Or one might just try to sink it with the weight of other
objections. And there are other objections in the literature—
see e.g. Tuggy 2003, 2004. My own view, though, is that even
after these further objections have been piled on, the
cumulative case is not dialectically strong enough apart from
the two further tasks just mentioned.

34. See e.g. Howard-Snyder 2003; Leftow 1999; Rea 2006; and
Tuggy 2003.

35. My objection bears some similarity to objections based on the

Athanasian Creed raised by Je rey Brower (2004a) against
Wierenga’s (2004) social trinitarian view.
36. Dale Tuggy (2003 and esp. 2004) presses roughly this point
against all versions of ST. Part-whole trintiarians could respond
by saying that the Son and Spirit are parts of the Father (who
is also a person). But in that event, their doctrine of two kinds
of divinity (full and derivative) would commit them to
denying that the Son and the Spirit (who would still be
derivatively divine) have the same nature as the Father (who
alone would have the full divine nature. Thus, they would run
into direct con ict with the requirement that the Son be of the
same substance with the Father.

37. Plantinga 1986, 1988, and 1989 provide representative

examples.

38. See Brower and Rea 2005a, b. See also Brower 2004b.

39. Proponents of this strategy include G. E. M. Anscombe and P.

T. Geach (1963: esp. 118–20); James Cain (1989); A. P.
Martinich (1978, 1979); and Peter van Inwagen (1988, 2003).
I count myself as both a critic and a proponent, since I reject
one way of deploying the strategy but endorse another. (See
Rea 2003; Brower and Rea2005 a, b). For other criticism, see
Bartel 1988; Cartwright 1987; Merricks 2006; and Tuggy
2003.

40. See e.g. Geach 1967, 1980: sects. 30, 34, 110.

41. See esp. van Inwagen 1988, 2003.

42. There’s controversy in the literature on Aristotle over whether
distinct human beings really share numerically the same form or
not. Here I assume that they do; but since I feel no special
burden to follow Aristotle on that point, nothing really
depends on the assumption.

43. Or, if you like, a lump of matter. See Brower and Rea 2005a:
504 n. 10 for discussion.

44. Cf. Cary 1992: 381.

45. One might want to say that it is the Father who plays the role
of underlier, that the Son is a compound of the Father and the
property being Begotten, and that the Spirit is a compound of
the Father (or the Son) and the property of Proceeding. Thisview
is suggested by passages in Augustine, and was rst brought to
my attention by Anne Peterson. I take it as a variation on the
present theme, rather than a genuine rival to the view I am
defending.

46. Cross 2002a: 284. See also Cary 1992.

47. On Gregory of Nyssa’s view of universals, see Cross 2002b.

48. I want to avoid views that rule out shared agency; but I don’t
myself want to commit to it. For, of course, it raises serious
questions. The Son prays to the Father. Does the Father do so
as well? Proponents of shared agency will say ‘yes, but not in
the same way.’ (See e.g. Hasker 1970: 29.) Trying to make
 good sense of this response, however, would require another
paper.

49. On the Trinity Bk. 11 ch. 2; in Scha and Wace 1887/1999: iii.
145–6.

50. Ibid. ch. 5; in Scha and Wace 1887/1999: iii. 146–7.

51. See e.g. Gregory Nazianzen’s analogy of three suns in his Fifth
Theological Oration, ch. 14, and the rainbow analogy in Basil of
Caesarea’s Epistle 38 (generally regarded as having been
written by Gregory of Nyssa rather than by Basil).

REFERENCES

ANSCOMBE, G.E.M., and GEACH, P.T.(1963). Three Philosophers.

Ithaca: Cornell University Press.

AYRES, LEWIS (2004). Nicaea and its Legacy: An Approach to Fourth-

Century Trinitarian Theology. New York: Oxford University Press.

BARNES, MICHEL RENÉ (1995a). ‘Augustine in Contemporary

Trinitarian Theology’, Theological Studies 56: 237–50.

——— (1995b). ‘De Regnon Reconsidered’, Augustinian Studies 26:

51–80.

——— (1998). ‘The Fourth Century as Trinitarian Canon’, in Lewis

Ayres and Gareth Jones (eds.), Christian Origins: Theology,
Rhetoric, and Community. London: Routledge, 47–67.
——— (1999). ‘Rereading Augustine’s Theology of the Trinity’, in
Davis, Kendall, and O’Collins (eds.) (1999:145–76).

BARTEL, TIMOTHY (1988). ‘The Plight of the Relative Trinitarian’,

Religious Studies 24:129–55.

——— (1993). ‘Could There Be More Than One Almighty?’ Religious

Studies 29: 465–95.

——— (1994). ‘Could There Be More Than One Lord?’ Faith and
Philosophy 11: 357–78.

BERKHOF, LOUIS (1938/96). Systematic Theology: New Combined

Edition. Grand Rapids: Wm. B. Eerdmanns.

BROWER, JEFFREY (2001). ‘Relations without Polyadic Properties:

Albert the Great on the Nature and Ontological Status of
Relations’, Archiv fur Geschichte der Philosophie 83: 225–57.

——— (2004a). ‘The Problem with Social Trinitarianism: A Reply to

Wierenga’, Faith and Philosophy 21: 295–303.

——— (2004b). ‘Abelard on the Trinity’, in J. E. Brower and K.

Guilfoy (eds.), in The Cambridge Companion to Abelard
(Cambridge: Cambridge University Press), 223–57.

——— (2005). ‘Medieval Theories of Relations’, Stanford

Encyclopedia of Philosophy,
<http://plato.stanford.edu/entries/relations-medieval/#3.3>,
accessed 27 May 2008.
——— and REA, MICHAEL (2005a). ‘Material Constitution and the
Trinity’, Faith and Philosophy 22: 487–505.

——— (2005b). ‘Understanding the Trinity’, Logos 8:145–57.

BROWN, DAVID (1985). The Divine Trinity. La Salle: Open Court.

——— (1989). ‘Trinitarian Personhood and Individuality’, in

Feenstra and Plantinga (1989: 48–78).

CAIN, JAMES (1989). ‘The Doctrine of the Trinity and the Logic of
Relative Identity’, Religious Studies 25: 141–52.

CARTWRIGHT, RICHARD (1987). ‘On the Logical Problem of the Trinity’,

in id., Philosophical Essays. Cambridge, Mass.: MIT, 187–200.

CARY, PHILIP (1992). ‘On Behalf of Classical Trinitarianism: A

Critique of Rahner on the Trinity’, Thomist 56: 365–404.

——— (1995). ‘Historical Perspectives on Trinitarian Doctrine’

Religious and Theological Studies Fellowship Bulletin, Nov/Dec,
<http://www.scribd.com/doc/2385279/Historical-
Perspectives-on-Trinitarian-Doctrine-by-Phillip-Cary>, accessed
27 May 2008.

CLARK, KELLY JAMES (1996). ‘Trinity or Tritheism?’, Religious Studies

32: 463–76.

COAKLEY, SARAH (1999). ‘“Persons” and the “Social” Doctrine of the

Trinity: A Critique of Current Analytic Discussion’, in Davis,
Kendall, and O’Collins (eds.), The Trinity, 123–44.

——— (ed.) (2003). Re-thinking Gregory of Nyssa. Oxford: Blackwell.

CRAIG, WILLIAM Lane (2005). ‘Does the Problem of Material
Constitution Illuminate the Doctrine of the Trinity?’, Faith and
Philosophy22: 77–86.

CROSS, RICHARD (2002a). ‘Two Models of the Trinity?’, Heythrop

Journal 43: 275–94.

——— (2002b). ‘Gregory of Nyssa on Universals’ Vigiliae Christianae

56: 372–410.

DAVIS, STEPHEN T., KENDALL, DANIEL, and O’ COLLINS, GERALD (eds.)

(1999). The Trinity. New York: Oxford University Press.

DE RÉGNON, Theodore (1892–8). Études de théologie positive sur la

Sainte Trinité, 4 vols. Paris: Victor Retaux.

FEENSTRA, RONALD, and PLANTINGA, CORNELIUS, JR. (eds.) (1989).

Trinity, Incarnation, and Atonement. Notre Dame: University of
Notre Dame Press.

FESER, EDWARD (1997). ‘Swinburne ‘s Tritheism International Journal

for Philosophy of Religion 42: 175–84.

FORREST, PETER (1998). ‘Divine Fission: A New Way of Moderating

Social Trinitarianism Religious Studies 34: 281–97.

GEACH, PETER (1967). ‘Identity’, Review of Metaphysics 21: 3–12.

Repr. in Logic Matters. Berkeley: University of California Press,
1972.

——— (1980). Reference and Generality, 3rd edn. Ithaca: Cornell

University Press.
HASKER, WILLIAM (1970). ‘Tri-Unity’, Journal of Religion 50:1–32.

HODGE, CHARLES (1873). Systematic Theology. New York: Scribner,

Armstrong, 1.

HOWARD-SNYDER, DANIEL (2003). ‘Trinity Monotheism’ Philosophia

Christi 5: 375–403.

KELLY, J.N.D. (trans.) (1964). The Athanasian Creed. London: Adam &
Charles Black. -(1972). Early Christian Creeds, 3rd edn. New
York: Longman.

LACUGNA, CATHERINE MOWRY (1986). ‘Philosophers and Theologians

on the Trinity’, Modern Theology 2:169–81.

——— (1991). God For Us: The Trinity and Christian Life. San
Francisco: Harper San Francisco.

LAYMAN, C. STEPHEN (1988). ‘Tritheism and the Trinity’, Faith and

Philosophy 5: 291–8.

LEFTOW, BRIAN (1999). ‘Anti Social Trinitarianism’, in Davis, Kendall,

and O’Collins, The Trinity, 203–49.

——— (2004). ‘A Latin Trinity’, Faith and Philosophy 21: 304–33.

——— (2007). ‘Modes without Modalism’ in P. van Inwagen and D.

Zimmerman (eds.), Persons: Human and Divine. Oxford: Oxford
University Press, 357–75.

LEITH, JOHN H. (ed.) (1982). Creeds of the Churches, 3rd edn.

Louisville, Ky.: John Knox.
MARTINICH, A.P. (1978). ‘Identity and Trinity’, Journal of Religion
58:169–81.

——— (1979). ‘God, Emperor, and Relative Identity’, Franciscan

Studies 39:180–91.

MERRICKS, TRENTON (2006). ‘Split Brains and the Godhead’, in

Thomas Crisp et al. (eds.), Knowledge and Reality: Essays in Honor
of Alvin Plantinga. Dordrecht: Kluwer.

MOLTMANN, JÜRGEN (1981). The Trinity and the Kingdom: The Doctrine
of God, trans. Margaret Kohl. San Francisco: Harper San
Francisco.

MORELAND, J.P., and CRAIG, WILLIAM LANE (2003). Philosophical

Foundations for a Christian Worldview. Downers Grove:
InterVarsity Press.

MORRIS, THOMAS V. (1991). Our Idea of God. Downers Grove:

InterVarsity Press.

O’COLLINS, G ERALD (1999). The Tripersonal God. Mahwah: Paulist

Press.

PLACHER, WILLIAM (1983). A History of Christian Theology.

Philadelphia: Westminster.

PLANTINGA, CORNELIUS, Jr. (1986). ‘Gregory of Nyssa and the Social

Analogy of the Trinity’, Thomist 50: 325–52.

——— (1988). ‘The Threeness/Oneness Problem of the Trinity’,

Calvin Theological Journal 23: 37–53.
——— (1989). ‘Social Trinity and Tritheism’, in Feenstra and
Plantinga, Trinity, Incarnation, and Atonement, 21–47.

REA, MICHAEL (2003). ‘Relative Identity and the Doctrine of the

Trinity’, Philosophia Christi 5: 431–46.

——— (2006). ‘Polytheism and Christian Belief Journal of

Theological Studies 57:133–48.

RICHARDSON, CYRIL (1955). ‘The Enigma of the Trinity’, in Roy

Battenhouse (ed.), Companion to the Study of St. Augustine. New
York: Oxford University Press, 235–56.

SCHAFF, PHILIP, and WACE, HENRY (1887/1999). Nicene and Post-Nicene

Fathers, First Series, 14 volumes. Peabody: Hendrickson.

——— ——— (1900/1999). Nicene and Post-Nicene Fathers, Second

Series, 14 volumes. Peabody: Hendrickson.

SCHWARTZ, E. (ed.) (1914). Acta Conciliorum Oecumenicorum (Berlin:

W. de Gruyter).

STEAD, CHRISTOPHER (1977). Divine Substance. New York: Oxford

University Press.

SWINBURNE, RICHARD (1994). The Christian God. Oxford: Clarendon.

TUGGY, DALE (2003). ‘The Un nished Business of Trinitarian

Theorizing’, Religious Studies 39: 165–83.

——— (2004). ‘Divine Deception, Identity, and Social

Trinitarianism’, Religious Studies 40: 269–87.
TURCESCU, LUCIAN (2005). Gregory of Nyssa and the Concept of Divine
Persons. New York: Oxford University Press.

VAN INWAGEN, Peter (1988). ‘And Yet They Are Not Three Gods But
One God’, in Thomas Morris (ed.), Philosophy and the Christian
Faith. Notre Dame: University of Notre Dame Press, 241–78.

——— (2003). ‘Three Persons in One Being: On Attempts to Show

that the Doctrine of the Trinity is Self-Contradictory’, in Melville
Stewart (ed.), The Holy Trinity. Dordrecht: Kluwer, 83–97.

——— (2004). ‘A Theory of Properties Oxford Studies in Metaphysics

1:107–38.

WIERENGA, EDWARD (2004). ‘Trinity and Polytheism’, Faith and

Philosophy 21: 281–94.

WILLIAMS, C.J.F. (1994). ‘Neither Confounding the Persons nor

Dividing the Substance’, in Alan Padgett (ed.), Reason and the
Christian Religion: Essays in Honor of Richard Swinburne. Oxford:
Clarendon, 227–43.

WOLFSON, H. A. (1964). The Philosophy of the Church Fathers, i. Faith,

Trinity, Incarnation. Cambridge: Harvard University Press.
 CHAPTER 19

ORIGINAL SIN AND ATONEMENT

OLIVER D. CRISP

THE atonement is one of the central and de ning doctrines of

Christian theology. Yet the nature of the atonement—how it is that
Christ’s life and death on the cross actually atone for human sin—
remains a theological conundrum. Over the centuries a number of
di erent theories of the atonement have been proposed that attempt
to make sense of the biblical data and o er an account of the nature
of the atonement. But none has won universal support. This chapter
o ers a new argument for an old theory of the atonement, namely,
penal substitution. This is called for because it seems to me that
traditional arguments for penal substitution (though not the
doctrine itself) face several serious di culties that have led a
number of theologians, past and present, to reject or amend the
doctrine. If my argument is correct, then some of the most serious
problems facing traditional arguments for penal substitution can be
circumvented.

We shall proceed as follows. First I shall set out the theological

context for the argument. This involves giving some account of
alternative theories of the atonement in the tradition, and why
penal substitution might be thought a particularly appealing way of
thinking about the atonement. Then I shall set forth an argument for
a version of penal substitution that involves the application of an
idea found in some Augustinian accounts of the transmission of sin
from Adam to his progeny. At the end of this exposition, I o er
some brief comments on the metaphysics undergirding the theology
of this argument. In a nal section, several theological and
philosophical objections to this reasoning are discussed.

CONTEXT: SOME THEORIES OF ATONEMENT

Systematic theologies usually include a discussion of a number of

di erent accounts of the atonement found in the Christian tradition.
Although di erent taxonomies could be o ered, I shall discuss the
following: the ransom theory; the satisfaction theory; the moral
exemplar theory; and penal substitution. These are not all the
theories of atonement one can nd in the Christian tradition. But
they are arguably the most in uential.1

Some version of the ransom theory is often thought to be the

majority view in the Patristic period of the church. It is certainly not
the view of all the Fathers, but it is re ected in the work of a
number of patristic authors.2 The story informing some historic
versions of this theory runs like this: through sinning against God,
human beings have sold themselves into slavery to the Devil. God,
not wishing to see humanity remain in this benighted situation,
deigns to bring about the means by which human beings can be
emancipated from their vitiated moral condition and indentured
state. This involves the incarnation and work of Christ, who
hoodwinks the Devil into bringing about his death on the cross.
Instead of destroying Christ, this act actually enables Christ to pay
the ransom price required to liberate human beings. Thus, the work
of Christ achieves two interconnected goals: to break the hold evil
(in the person of Satan) has over human beings, and to enable
human beings to be reconciled with God.

The ransom theory (expressed in the sort of story just given) was
dusted down in the mid-twentieth century by the Scandinavian
theologian Gustaf Aulén, whose book Christus Victor claimed that
the ransom model was the ‘classic’ view of the atonement, which
had been overlaid by centuries of theological accretions.3 He
maintains that this classic view is superior to later accounts of the
work of Christ and should be preferred to other ways of thinking
about the atonement. Aulén’s work has been widely discussed,
particularly amongst systematic theologians. But it is not without
problems.

One concern with Aulén’s work is that it is not obvious that the
ransom view is the ‘classic’ account of the atonement, or the best or
most comprehensive way of thinking about the nature of atonement.
That would have to be argued for. Even if the ransom view is the
majority position amongst the Fathers, this does not automatically
give it a privileged status above other theories of the atonement.
The venerable pedigree of a given theory is not, after all, a cast-iron
guarantee of its being the best explanation of the data. A second
problem is that the ransom theory involves a rather naive account of
the relationship between the Fall, human sin, and enslavement to
the Devil, with unpleasant overtones of God being implicated in an
act of deception. That said, the conceptual hard core of the doctrine
is that Christ’s work is remedial, being the price required to
‘purchase’ or ‘buy back’ some number of fallen humanity enslaved
to evil. It is perfectly feasible to set out such a doctrine without the
paraphernalia of the story about deceiving the Devil. Then, Christ’s
work would bring about the reconciliation of some number of fallen
humanity, including their release from the power of evil—the
language about the personi cation of evil having been removed.
This would have the advantage of divesting the theory of the
unwelcome consequence of God’s involvement in deception.
However, if the core of the theory can be stated without the need
for personi ed evil, to whom is the ransom price of Christ’s work
paid? It might be possible to ‘amend’ the theory in the direction of a
satisfaction account of the atonement at this point. In which case
God, whose moral governance has been impugned by human sin (let
us say), requires the ‘ransom’ as the price for human wickedness.
Taken in this way, Christ’s work is not so much a question of paying
the ransom demand of the Devil as paying the price demanded by
the divine moral law for human sin. This moral demand might be
thought a ‘ransom’ in a loose, non-philosophical sense, since it
involves payment of a price set in order to bring about the
reconciliation of some number of fallen human beings. But it is not
a ransom in the sense of a payment made to a diabolical entity that
has enslaved humanity.

But it may be simpler to put to one side those aspects of this

theory that are objectionable, and search for an atonement theory
able to account for what is helpful in this understanding of the work
of Christ. I suggest there are good theological reasons for pursuing
such an alternative. For one thing, New Testament references to the
payment of a ransom (e.g. the Greek word lutron used in Matt. 20:
28 and the parallel in Mark 10: 45) do not require a ransom theory
of atonement to be rendered plausible or consistent with other
things in Scripture. In the context of these passages, the idea of
ransom has to do with saving fallen humans (‘the many’) from a
condition of oppression through an act of self-sacri ce. In fact, there
are no biblical grounds for thinking God enters into some unholy
pact with the Devil over the enslavement and ransom of humanity.
Moreover, satisfaction or penal substitution theories of atonement
can make sense of this biblical idea of ransom, along the lines of
Christ’s work being able to ‘buy back’ some number of fallen
humanity from the moral consequences of sin, without the unhelpful
elements of the ransom theory.4

This brings us to the satisfaction theory, elements of which we

have already segued into the ransom account. Satisfaction is a
theory that is given its rst systematic exposition in St Anselm of
Canterbury’s magisterial Cur Deus Homo (Why The God-Man).
Famously, St Anselm’s version of satisfaction theory is set up remoto
christo—that is, setting aside the incarnation and reasoning simply
on the basis of more general theological claims to the necessity of
the incarnation of a God-Man for atonement to be made. St Anselm
thinks that the nature of God is such that he requires satisfaction for
sin committed against him. God cannot set aside satisfaction
without in some way compromising his own perfect character,
which is impossible. Because human sin is committed against a
being of in nite worth and honour (i.e. God), it generates a
correspondingly in nite demerit that no nite human being can
annul, even if he or she were without sin.5 And any putative
redeemer would have to be without sin because only then would
such a person be in a position to perform an act of supererogation,
that is, an act that is non-obligatory and generates a merit su cient
to atone for human sin (i.e. a merit of in nite worth). Any sinful
human would be incapable of producing such an act since he or she
is already in nitely indebted to God on account of his or her sinful
condition. And any sinless human would still be incapable of
producing a supererogatory act of su cient merit to balance o the
sin of all humanity, since no mere human moral action generates an
in nite merit that can be applied to all humanity. Although he does
not make this explicit, the logic of St Anselm’s position means that
some other creature, e.g. a sinless angel, cannot act on behalf of
human beings in this regard—nor could an angel generate an act of
su cient merit to balance o human sin, since angels cannot
perform acts that have an in nite merit, on account of being nite
creatures.6 In fact, reasons St Anselm, only God is capable of
producing a supererogatory act that would be of su cient merit to
o er satisfaction for human sin. But since one of the stipulations for
satisfaction according to St Anselm is that a human being acts on
behalf of other human beings in atonement,7 only a being that is at-
one-and-the-same-time both fully divine and fully human can
provide satisfaction for (human) sin. Hence, St Anselm arrives at the
conclusion that only the God-Man can perform an act of
supererogation that has the worth requisite to provide satisfaction;
God must become human in order to atone for human sin.

On the Anselmian account, the mechanism for atonement has to

do with the God-Man performing some in nitely meritorious act
that may be used as a merit su cient to balance out the demerit of
human sin. Christ performs such an act in his atonement, o ering
the merit to God on behalf of fallen human beings. His meritorious
act is performed in order to satisfy the moral law to which fallen
humans are, as it were, ‘indebted’. But it is no part of this theory
that the God-Man take upon himself the penal consequences of
human sin, standing in the place of fallen humans as a penal
substitute, as with the penal substitutionary theory of atonement.
The soubriquet ‘the commercial theory of atonement’ by which the
Anselmian view is known in some older textbooks of theology
underlines the transactional character of the Anselmian satisfaction
theory.
 Recently, Richard Swinburne has defended a modi ed version of
satisfaction theory, without St Anselm’s commitment to the notion
that God requires satisfaction for sin, commitment to the moral claim
that sin is in nitely heinous, or the idea that Christ’s atonement
must be strictly equivalent in value to the sin committed against
God.8 He speaks of his theory in terms of the biblical idea of
sacri ce, which he regards as the most helpful way of conceiving of
the work of Christ found in Scripture, namely, Christ o ering
himself up as a sacri ce for human sin. But in Swinburne’s hands
this notion of sacri ce is construed in a way that is consistent with a
version of satisfaction. Christ o ers up some supererogatory act of
sacri ce to God, and fallen human beings are able to somehow
‘access’ that work in order to bring about their own redemption. In
some respects, this echoes the more moderate version of satisfaction
theory advocated by St Thomas Aquinas, according to which Christ’s
satisfaction for sin is a most tting means of atonement, but not
necessarily the only one possible, all things considered. That said, St
Thomas does suggest that the atonement is consequentially
necessary in some sense, that is, necessary as a consequence of God
ordaining to bring about the sort of world he did, with human
beings who would sin and require salvation.9 Swinburne distances
himself from this language about the consequential necessity of
atonement, and several other aspects of the Thomist understanding
of Christ’s work in his own account of Christ’s work.
 The concept of atonement used to do much of the work in
Swinburne’s theory includes four distinct components: repentance,
apology, reparation, and penance. Humans are able to repent and
apologize to God for their sin. But they are not able to o er
adequate reparation or penance—God must arrange this. Swinburne
thinks that ‘the way in which we humans can use Christ’s life and
death as a means of removing sin is by o ering it as our reparation
and penance. To do so, we must join to it our feeble repentance and
halting words of apology.’10 He concedes that God may set aside
satisfaction and simply forgive sin. But God has good reason not to
do this. Human beings need to understand the moral seriousness of
their sins, just as children need to be corrected when they have gone
astray. And it is permissible for God to bring about some
supererogatory act of satisfaction through the work of Christ,
though there is no necessity for God to do so, all things considered.
If we think of Swinburne’s theory as a moderate version of
satisfaction as I am suggesting, then it looks like the hard core of
satisfaction theories has to do with the issue of God’s honour and/or
moral law being satis ed by some act of supererogation which is
performed by the God-Man, and to which we join our own act of
penitence. Whether Christ’s act is a requirement for satisfaction,
what moral value it must have (if it must have a particular moral
value at all) and in what sense it is necessary all things considered,
if it is necessary at all, appear to be matters that not all satisfaction
theorists agree upon.
 To my mind, the satisfaction theory of atonement is less
attractive than penal substitution—though only marginally less
attractive—for reasons that will become apparent presently. In any
case, what I am calling the ‘hard core’ of satisfaction theories can be
assimilated by theories of penal substitution, though the exact
mechanism by which atonement is brought about di ers between
the two theories.11 However, if it transpires that some of the more
controversial aspects of penal substitution are indefensible (e.g. the
claim that culpability for my sin can be transferred to Christ), some
version of the satisfaction theory may be able to succeed where
penal substitution fails.

We come to the moral exemplar theory, the rst clear statement

of which is often thought to be traceable to elements of Abelard’s
theory of atonement, written at least in part in response to St
Anselm’s work. However, caution needs to be exercised in
evaluating this claim, often repeated in textbooks of Christian
theology. Recently, Philip Quinn has challenged the idea that
Abelard’s account of the atonement is reducible to something like
the idea that Christ is merely a moral example whose work we
should emulate, and whose act of love in the atonement ought to
stimulate in us a reciprocal response of love to God.12 Whether or
not Abelard’s medieval theory has elements that sound at times like
an embryonic version of a moral exemplar theory, it is certainly true
that the moral exemplar theory enjoyed a vogue in early modern
theology with the advent of Socinianism, which, in the work of
Faustus Socinus, sought to overthrow the arguments of advocates of
penal substitution and satisfaction theories of atonement.13 Interest
in this theory continued in nineteenth-century liberal theology. A
modern version of this sort of view can be found in the work of the
British philosopher of religion, John Hick.14 As part of his bid to
provide a religiously pluralist account of Christology, he argues that
Christ’s work must be seen primarily in terms of a moral example of
self-sacri cial love, and what, following F. D. E. Schleiermacher, we
might call ‘God-consciousness’. There is no satisfaction brought
about by the work of Christ, on this view, since no such satisfaction
is necessary. Indeed, one persistent objection to this sort of account
of the work of Christ is that it is not really a theory of atonement at
all, since the logic of this view is that nothing is atoned for. On
Hick’s version of the theory, human sin is really a matter which,
though not trivial, may be forgiven without any reparation being
exacted as satisfaction. But even if this view does not necessarily
trivialize sin, it does raise a rather serious problem about the moral
cost involved in the incarnation and death of Christ. One is left
wondering why God could not have provided the requisite moral
example for human beings to emulate without such terrible cost to
himself.

This is not to deny that Christ’s atonement is a moral example

and act of love. The question is whether the work of Christ is
reducible to these elements. In my view it is not. As I have just
indicated, something more needs to be said about the nature of
Christ’s work in order for it to be a work of atonement, although
Christ’s work certainly includes the components of the moral
exemplar theory. Here much depends on the theory of punishment
one adopts. Defenders of the moral exemplar theory reject the idea
that retribution—where there must be a moral ‘ t’ between crime
and punishment—is part of the theory of punishment that informs
the work of Christ. God is not ‘wroth with sin’; he is not bound to
punish; in fact, in some versions of this theory he is bound not to
punish, but to forgive sin. All of which is in stark contrast to the
other theories of atonement canvassed here, which in various ways
include retributive punishment as a constituent of divine
distributive justice.15

The nal theory of atonement we shall consider is penal

substitution. I take it that penal substitution is that family of views
of the atonement according to which Christ’s atoning work on the
cross consists in his acting as a penal substitute, taking upon himself
the penal consequences of my sin and being punished for that sin in
place of the sinner. As I have already intimated, this atonement
theory is able to incorporate the core ideas of the other theories we
have discussed, whilst o ering a di erent understanding of the
nature of the atonement.16 Hence, the advocate of penal substitution
can agree that Christ’s work is remedial, as per the ransom theory. It
also includes the concept of satisfaction (of divine moral law). And,
in common with moral exemplar, satisfaction, and ransom theories,
defenders of penal substitution may speak of Christ’s work as
providing a moral example in life and death of obedience to the will
of God that Christians ought to emulate.

In the theological literature, the relation between Christ’s penal

substitution for the sinner and the sinner’s sin and/or guilt is usually
(though not always) made sense of according to what I shall call a
forensic ction.17 God treats Christ as if he is guilty of the sin of
fallen human beings, and ‘punishes’ him accordingly, bringing about
his death on the cross, which is a suitable act of atonement. God
also treats the sinner as if she was sinless in view of Christ’s work on
the cross, provided the sinner appropriates that work for herself.
(Quite how that is done is beyond the scope of this chapter.) The
important thing to see here is that on this forensic ction account of
penal substitution, the innocent Christ really is treated as if he were
the guilty sinner, though he is not. And God really does punish the
innocent in the place of the guilty. Christ is said to ‘represent’ the
sinner in his act of atonement.

There are, as David Lewis points out, de nite analogues to penal

substitution in human systems of justice, and in many ways we
humans are in two minds about penal substitution, irrespective of
whether or not we believe in its theological application.18 We allow
it under certain conditions for certain sorts of crime where the penal
consequences of the crime committed are pecuniary in nature. But
we do not allow it when the crime is, say, rst-degree murder. In
that case, the penal consequences of the crime committed have to be
met by the one guilty of the crime and we would consider it unjust
if a substitute were to serve the punishment in place of the
perpetrator. Cases of crimes where the penal consequences involved
are pecuniary, or monetary, need not be trivial. They can be very
serious crimes such as large-scale fraud, which carry with them
crippling nancial consequences for the person convicted of that
crime. In such circumstances what is important is that the ne is
paid, not that the perpetrator pays it. Someone generous enough
might pay the full penalty on behalf of the fraudster, if they had the
nancial wherewithal. The law would be satis ed, even though the
perpetrator of the crime has not actually paid the penal consequence
set out by the law—a substitute has done so instead. But
importantly, we nd no examples of legislation allowing
substitution when the crime is a serious felony, such as murder. In
such cases, the one guilty must meet the penal consequences of that
crime, and we would consider it a terrible miscarriage of justice
were a substitute punished in place of the perpetrator.

Hence, there are circumstances in which penal substitution in

human jurisprudence is deemed appropriate, while there are others
—perhaps more precise analogues to our situation as condemned
sinners before God—where, it seems, it is not. What the defender of
the theological application of penal substitution needs is some
reason for thinking that it is just for God to punish Christ in place of
the sinner. We need to know why the theological application of the
doctrine is morally permissible, as with the case of pecuniary penal
substitution, outlined by Lewis. For, whatever else the defender of
penal substitution says, unlike pecuniary penal substitution, it
certainly looks unjust that Christ, an innocent individual, should be
punished in place of me, a sinful individual. This concern goes all
the way back to the Socinians of the immediate post-Reformation
period, and can still be heard today.19

REALIST PENAL SUBSTITUTION: THE ARGUMENT

Nevertheless, I think an argument for penal substitution can be had,

which gets around these di culties. It is to that argument we now
turn. A central plank of the reasoning I will set forth involves
appropriating a theological concept called Augustinian realism,
deriving from St Augustine of Hippo’s work on original sin.20
Advocates of this view claim that original sin includes both the state
of moral corruption with which all humanity post-Adam are born,
which inevitably leads humans into acts of sin, and the culpability
aspect of guilt that accrues to Adam’s rst sin.21 Both the moral
corruption engendered by Adam’s primal sin and the culpability
aspect of his guilt is transferred, on this way of thinking, from Adam
to his progeny. This transfer of Adam’s sin and guilt to me is just
provided God somehow organizes things such that Adam and his
posterity are one metaphysical entity so that what the rst part of
that entity does has moral implications for later parts of the same
entity.22 Hence, Adam’s sin and guilt on this view is (somehow)
really mine. For this reason, the view is called Augustinian realism.
 A number of theologians who have defended robust accounts of
original sin including the concept of original guilt have thought that
it makes much more sense to conceive of the transmission of
original sin as a question of God imputing Adam’s sin and guilt to
me, treating me as if I was originally sinful and guilty of Adam’s sin,
though I am not, strictly speaking (which is, of course, another
instance of a theological forensic ction). But defenders of
Augustinian realism usually object that treating someone as if they
are guilty of a crime is not the same as their being guilty of that
crime. Moreover, being held culpable for the crime of another,
especially another that is removed in time and space from me, and
whose course of action I did not and could not have approved or
otherwise endorsed, is monumentally unjust. Thus, if God imputes
the sin and guilt of Adam to me, he acts unjustly. But God cannot
act unjustly. Scripture states that God does transmit the sin and guilt
of Adam to his progeny (Rom. 5:12–19). So there must be some
other account of this transmission that does not have the morally
unpalatable consequence of God imputing Adam’s sin and guilt to
me. And Augustinian realism is such an explanation.23 St Augustine
does not put it quite like that, but there are certainly later
Augustinian realists who do think in these terms.24

The central Augustinian realist notion that Adam and his

progeny are somehow metaphysically united so that original sin
may be justly transmitted from Adam to his posterity is often
con ated with a particular theory about the manner of such
transmission. But Augustinian realism is not a thesis that requires a
particular account of the exact mechanism by which this
transmission occurs. It is simply the theological notion that such an
arrangement obtains between Adam and his progeny. How such a
metaphysical arrangement obtains is not the same issue as that it
obtains (if it does obtain). Some Augustinian realists claim that the
mechanism by which this arrangement obtains involves
traducianism. Individual souls are produced by ssion or
parturition, being generated by the soul of one or both parents as
part of the process of natural generation. The parent souls are
produced in the same fashion from their parents, and so on, going
all the way back through the generations to the rst human pair. In
which case, my soul is literally, though mediately, a chip o the old
(Adamic) block; my soul was part of the soul of Adam when he
sinned prior to its being ‘individualized’, so to speak, through a
process of ssion or parturition down through the generations from
Adam to me.25 According to the traducian, it is just for God to
punishme for Adam’s sin since my soul was part of Adam’s soul
when it sinned. But an Augustinian realist need not be a traducian,
although, as a matter of historical fact, most have been. That is, an
Augustinian realist need not be committed to a particular theory of
how God brings about the distinction between Adam and his
progeny whilst still making the idea of original unity plausible, in
order for sin to be justly transmitted from one human being to the
other.
 Let us now attempt to apply this theological ‘realist’ thesis about
the union of Adam and his progeny in the transmission of original
sin to the atonement in order to help make sense of the union
between Christ and those human beings he saves through his work.
In so doing I shall not be concerned with expounding Augustine or
his followers, but merely with trying to make sense of a particular
construal of Augustinian realism, and its application to the work of
Christ. This construal avoids the traducian version of Augustinian
realism and opts instead for a di erent, fourdimensionalist account
of the ‘realist’ transmission of sin.

We begin with the idea that Adam and his progeny are
(somehow) one metaphysical entity such that God may justly pass
on the moral consequences of Adam’s sin to his heirs because they
are all members of one persisting entity, or object, that we might
call Fallen Humanity. To be clear, this object is composed only of
Adam and his progeny all of whom are ‘parts’26 or phases in the life
of the whole entity. On the sort of view of Fallen Humanity I have
in mind, if the rst ‘part’ of the object sins, later ‘parts’ of the same
object (including all of humanity after Adam) have a vitiated moral
nature as a result of the action of Adam. It is rather like an acorn
that is infected with some disease that then a ects all the later
stages in the life of the sapling and tree into which it grows.27

Of course, it does not follow that just because Adam and his
progeny are ‘parts’ of one persisting object, Adam’s progeny share
the same morally vitiated condition as Adam. Nor does it follow that
just because an acorn is infected with some disease all the later
stages in the life of the tree into which it grows are a ected by this
disease. Some diseases might only a ect the rst phase of the life of
the tree, rather than all phases of its life. But the sort of reasoning I
have in mind is o ered as a possible explanation of how original sin
is transmitted from one generation to the next from some rst
human parents. It is true that later phases of the life of a persisting
entity may not be a ected by what happens in an earlier phase of
the same life. But the Augustinian realist story is concerned to make
sense of the claim that just such an arrangement does obtain in the
case of Adam and his progeny. All that needs to be granted here is
that it is possible for a given entity to have phases of its life that are
a ected by earlier phases of its life, and that all phases of the life of
a given persisting entity may be a ected by what happens to one of
the rst ‘parts’ or phases of that entity. This seems plausible just as
it seems plausible to think that an infected acorn may grow into a
tree that is diseased in every subsequent phase of its existence. And
this ts with the Augustinian realist story of original sin I am
drawing upon.28

Still, even if there are diseases that a ect the life of the tree in
all its phases, the disease a icting the life of the mature oak is not
necessarily the same as the disease a ecting the life of the sapling,
or the acorn. But suppose it is the same disease introduced to the
acorn that a ects all the later phases of the life of the oak into
which it grows. Is the corruption a ecting the oak the same as the
corruption a ecting the acorn? Much depends here on the ontology
one adopts. If one thinks persisting objects are wholly present at
each moment of their existence, then one might think that this is the
same disease. But if one thinks that the acorn and oak are di erent
temporal parts of one perduring four-dimensional whole object, then
one has a reason for thinking the disease a ecting the oak is not
numerically the same as that a ecting the acorn. For presumably
any disease persisting through time has temporal parts just as the
tree it infects has. And the temporal parts of a given thing are said
to be numerically distinct. So the fact that on this way of thinking
the oak tree does not have numerically the same disease as the
sapling does not pose any particular problem that defenders of
temporal parts are not already familiar with.

Temporal parts theorists sometimes speak of ‘tokens’ possessed

by individual temporal parts of an object, and the property
possessed by a whole object. On this way of thinking, the guilt-
token of the temporal part of Trevor existing now is numerically
distinct from the guilt-token of Trevor existing a moment ago. Yet
both are tokens of the guilt of Trevor, the four-dimensional entity
who has these temporal parts. The Augustinian realist can claim that
the original sin (and guilt) Trevor has transmitted to him from the
rst human is like the example of the guilttoken later temporal parts
of Trevor share with earlier temporal parts of Trevor. God arranges
things such that Adam and his progeny are a four-dimensional entity
to whom such moral-tokens can be attributed, and the morally
vitiated condition Adam post-Fall incurs is one that is transmitted to
all later phases of the life of this entity, Fallen Humanity.

But this may not be the only metaphysical story consistent with
Augustinian realism. Perhaps one can say both that objects that
persist through time are normally wholly present at each moment of
their existence, and that in the case of the transmission of original
sin the human species propagates sin through natural generation.
Construed along the lines of traducianism, this means that I have
Adam’s sin and guilt because my soul is either a fraction of the soul
Adam possessed (if souls are generated through ssion) or the
product of the soul of at least one of my parents, going all the way
back to Adam (if souls are parturient). Though ‘individualized’ or
otherwise brought about through natural generation, I retain the
property of original sin that has been passed on to me, as would be
the case with inherited physical diseases. The important di erence
between inherited disease and original sin, on this way of thinking,
is that my soul was present when Adam sinned as part of the
‘unindividualized’ soul that Adam possessed. Or, perhaps, the ‘seed’
of my soul was present in the soul of Adam, if souls are parturient
rather than ssiparous (compare Heb. 7: 10). So I am guilty of
Adam’s sin on account of possessing a fraction of Adam’s soul or
‘seed’ thereof that is now ‘individualized’ through soul- ssion or
parturition. But it seems consistent with this theological account of
the manner by which sin is transmitted to say that persisting objects
are wholly present at each moment of their existence. If this is right,
then an Augustinian realist account of the transmission of original
sin can be underpinned by more than one metaphysical account of
persisting objects, depending on what the Augustinian realist thinks
about the mechanism by which sin is transmitted.

This reasoning about an Augustinian realist account of the

doctrine of original sin can be applied, the relevant changes having
been made, to Christ and those for whose sins he came to atone—
call them, with a nod to the tradition, the elect. Consider the
possibility that Christ and the elect together compose one
metaphysical entity that persists through time, just as, on the
Augustinian realist way of thinking, Adam and his progeny do. This
object we shall dub Redeemed Humanity. Christ is in some sense the
rst member of this entity, and the elect are subsequent members.

But here the analogy between Adam and Christ begins to break
down. For, according to the Augustinian realist account of the
transmission of sin, Adam as the rst human can a ect the moral
condition of all subsequent humans in a way that some later human
being could not (because some earlier humans would already have
perished, or be beyond the in uence of this later person, or could
not be in possession of the whole ‘unindividualized’ soul that Adam
had, or whatever). However, Christ’s in uence cannot just follow
the arrow of time, moving from the moment his atonement is
complete to include all those who are elect that exist after that time.
Were this the case, then we would have no reason to think that
anyone who lived temporally prior to Christ’s act of atonement was
amongst the elect. But, of course, Scripture says otherwise, with
Hebrews 11 recounting a great cloud of witnesses for the faith in the
Old Testament, and Paul in 1 Corinthians 10 speaking of the
presence of Christ with the children of Israel wandering in the
desert, that drank from the rock (at Meribah) that was Christ. So, a
realist account of the atonement must also be able to include within
the ambit of salvation those of the elect who lived prior to Christ.

One way of making sense of this29 would involve co-opting Karl

Barth’s doctrine of election, or something very like it, as found in
Church Dogmatics ii pt. 2.30 According to Barth, Christ is the Elect
One and all those for whom Christ’s work atones are somehow
derivatively elect ‘in’ Christ (which, for present purposes, we shall
take to be functionally equivalent to Redeemed Humanity). This
solves the problem of those who are derivatively elect and die
before Christ by making Christ the ‘ rst’ amongst the elect, or,
perhaps better, the agent through whom election is derived. He
would not be temporally rst amongst all the derivatively elect as
Adam, on the Augustinian realist account, must be temporally the
rst human being for his sin to a ect all those living after him.
Instead, he would be somehow ‘logically’ or ‘metaphysically’ prior
to the derivatively elect, whose derivative election depends, in some
proleptic sense, on Christ’s work as the Elect One. A ‘realist’ version
of a Barth-like doctrine of election would also go some way towards
explaining the relation between Christ and the derivatively elect,
that is, Redeemed Humanity. They are all somehow members of one
metaphysical entity with Christ being metaphysically ‘prior’, as it
were, to those elect-in-Christ. Each of the derivatively elect is a part
of a larger whole, including Christ. But—and here the Augustinian
realism comes to the fore—the punishment due for the sin of the
derivatively elect is transferred to Christ, or to that ‘part’ of Christ
on the cross such that he really does take upon himself the penal
consequences of sin for humanity. Although he is not the one who
has sinned, or the part of the mass of Redeemed Humanity that has
sinned, because he is a member of this larger entity, he may pay the
consequences of the sin of other members of the same entity, by
which I mean the derivatively elect, like you and me.

Conversely, the atoning work of Christ means that the

derivatively elect are really united with Christ as ‘parts’ of one
whole entity (Redeemed Humanity). Notice that there is no
‘imputation’ involved in this process, no forensic ction, whereby
God treats Christ as if he were the sinner for the purpose of bringing
about atonement, as with many traditional accounts of penal
substitution. The transference of the punishment for sin from fallen
humanity to Christ, and the union of those selfsame humans with
Christ are two aspects of Christ’s atoning work that involve a real
union between Christ and the human beings concerned. And,
plausibly, this Barth-like doctrine of election could be conjoined
with a construal of the nature of the atonement as an act of penal
substitution.31 If what has been said so far sketches out how a
Barth-like doctrine could tell an internally consistent story about the
way in which the punishment due for human sin is transferred to
Christ and how sinful human beings may be united to Christ in
virtue of his act of atonement such that the nature of Christ’s
atonement is realist, then what remains to be explained is the
manner in which his work brings about atonement.

The explanation goes like this. At the cross Christ has transferred
to him the penal consequences of the sin of the derivatively elect—
that is, the derivatively elect members of Redeemed Humanity. At
that moment, he takes the penal consequences for which they are
guilty, as one ‘part’ of the larger metaphysical whole. As a result he
su ers for that sin. Having made atonement by expiating human sin
in his own person, the members of Redeemed Humanity for whose
sin he has atoned are reconciled to God.

A variation on this Barth-like version of realist penal substitution

that involves a more traditional Reformed account of election could
be applied just as well (nor do I suppose these are the only two
ways of construing this doctrine32). As with the Barth-like version of
the doctrine, the key problem is accounting for those amongst
Redeemed Humanity who die prior to Christ’s atoning act. A
distinction must be maintained between the natural union of Adam
and his progeny, in which Adam had to be the rst human being in
order that his sin might be transferred to the rest of humanity, and
the forensic union of Christ and the elect. The forensic nature of the
atonement does not require that Christ is chronologically the rst
human of a new, elect humanity. All it requires is that somehow his
atonement is the catalyst by which human sin is atoned for. In this
respect, as before, he is metaphysically or logically ‘prior’ to the
other parts of Redeemed Humanity, although many of those who
make up Redeemed Humanity live chronologically prior to Christ.

The manner of atonement on this more traditional Reformed

account is as follows. Christ on the cross has transferred to him the
penal consequences for the sin of the rest of Redeemed Humanity,
and he atones for their sin on the cross. As a consequence of this act,
the members of Redeemed Humanity other than Christ, including all
those who lived and died prior to the atonement, are reconciled to
God.33 The important di erence between Christ and the elect and
Adam and his progeny on this view is that because the relation
between Christ on the cross and the elect is a forensic one, Christ is
not a ‘part’ of Redeemed Humanity that is guilty of sin. He is
innocent. But he is able to take upon himself the penal consequences
of the sin of the elect because he is a part of the same object.34

So, we have at least two ways in which an Augustinian realist

account of the atonement could be set forth, namely, the Barthian
and the traditional Reformed versions of the doctrine. Note that
although both the Barthian and more traditional, Reformed realist
versions of penal substitution require that the penal consequences of
the sin of the elect is transferred to Christ, they avoid the moral
problems besetting other, standard arguments for penal substitution
that require only that Christ acts on behalf of sinful humans as their
representative. This is usually called representationalism, and it is this
argument that requires a forensic ction in order to make good on
the act of atonement. It is this that has generated so much of the
di culties facing apologists for penal substitution. This problem is
not so much overcome as circumvented by the realist accounts of
penal substitution o ered here, because Christ’s act of atonement is
not merely representationalist; he really takes upon himself the
penal consequences for human sin. This does not mean that the
other members of Redeemed Humanity are no longer guilty of
sinning. As many classical theologians have noted, there is a
di erence between being the one guilty of committing a sin, and being
the one whose guilt requires punishment for that sin. My sin can be
atoned for. Yet I remain the one who committed that sin. It is just
that the penal consequences of it have been dealt with so that I am
no longer liable to punishment for that sin. (The same is true
according to representationalism, of course, mutatis mutandis.) But
according to the realist doctrine, these penal consequences of sin are
transferable to Christ because he and I are members of one
metaphysical whole, Redeemed Humanity.

Earlier, in outlining Augustinian realism, I pointed out that the

central theological thesis of this theory is consistent with more than
one metaphysical story concerning the mechanism by which God
transmits original sin from Adam to his progeny. But it is di cult to
see how the same could be true of the several realist accounts of
atonement just o ered. For it looks like the advocate of a version of
‘realist’ penal substitution is committed to some metaphysical story
that includes a four-dimensionalist ontology, since, as I have set out
the argument above, both Fallen Humanity and Redeemed
Humanity are, in some sense, four-dimensional entities. But there
are di erent ways in which one could extrapolate a fourdimensional
ontology.35 All I have said thus far concerning the ‘realist’
component of realist penal substitution commits its advocates to this
sort of ontology. But there is still plenty of scope to argue over how
one makes sense of that ontology, and in particular, how it is that
the ‘parts’ or phases in the life of Fallen and Redeemed Humanity
are related to one another, and to the four-dimensional entity as a
whole.

OBJECTIONS

Now this, I will admit, is strong mead that may be di cult to

swallow. But then, the problems of the transmission of Adam’s sin
(and guilt) and of the transmission of the penal consequences of my
sin to Christ are theological di culties at the heart of the Christian
faith that many of the greatest minds in Christendom have struggled
to overcome. What I am suggesting is novel in some respects,
although not without precedent (Tobias Crisp, John Owen, and even
Eusebius of Caeserea use realistsounding language of the atonement
at times).36 But even if there have been a small number of
theologians in the past willing to consider something approaching a
realist doctrine of the atonement, or who have made much of the
language of ‘mystical union’ between Christ and the believer via the
atonement, no one, to my knowledge, has mounted this sort of
argument for the conclusion that penal substitution is just.

One problem with this is that many Christian thinkers are

somewhat suspicious of theological novelty because any articulation
of a particular doctrine must conform to Scripture and the tradition.
I do not think the doctrine just outlined is contrary to Scripture. In
fact, it seems to me to make good sense of Isa. 53: 6, Rom. 5: 12–19,
and 2 Cor. 5: 21 in particular, where union with Christ and the
notion of Christ ‘becoming sin for us’ are stressed. But it is true to
say that this is not a traditional argument and that may be a
problem. However, if this realist doctrine makes better sense of the
theological version of penal substitution by o ering a di erent
metaphysical underpinning for that doctrine that avoids certain
problems that beset traditional accounts of the doctrine, then I think
it is worthy of serious consideration. After all, our understanding of
the truth of the Gospel is sometimes mistaken. And, though I have
not argued for this here, it seems to me that penal substitution is an
important component of a true account of what Scripture teaches
concerning the atonement.

In addition to such theological concerns, there are objections of a

more philosophical nature that could be raised. One of the most
pressing concerns how an apparently gerrymandered object such as
Redeemed Humanity can possess moral properties, such as ‘being
sinful’ or ‘having temporal parts that are sinful’. It makes sense to
think that you or I can be thought of like this because we are
recognizably objects that persist through time (whatever that
means). But Redeemed Humanity is not like that. It is an object
cobbled together from ‘real’ persisting or perduring objects, that is,
human persons, which seems to lack the integrity and personhood
requisite to have such moral properties.

One line of response might be to query whether Fallen and

Redeemed Humanity really are the subjects of moral properties, like
persons. ‘Being sinful’ looks like a moral property, and one that does
not apply to Fallen or Redeemed Humanity as a four-dimensional
entity. But is, say, ‘having temporal parts that are sinful’ the sort of
property that requires the entity of which it is predicated to be a
moral subject? Not obviously. I suggest that Fallen Humanity does
not have any moral properties that would require it to be a moral
subject, like a person. Though Fallen Humanity is a real object,
according to the Barth-like and traditional Reformed versions of
realist penal substitution, it is not a person, nor a moral subject.
Some of its members have moral properties. But it does not have
moral properties that are predicated of it over and above the moral
properties of some of its members. And the same is true, mutatis
mutandis, for Redeemed Humanity.

Still, if this means that some members of Fallen Humanity and

Redeemed Humanity have moral properties, this might be
problematic. Earlier, in explaining how an advocate of Augustinian
realism might appropriate a temporal parts construal of a four-
dimensionalist ontology, I said perdurantists maintain that moral
properties apply to certain sorts of 4-D wholes, like persons, the
temporal parts of which have tokens of moral properties. Perhaps
this component of a temporal parts doctrine can be applied here as
well. Assume Fallen Humanity and Redeemed Humanity are objects
that have temporal parts. Some of the temporal parts of these
objects are themselves parts of other entities, distinct human agents
that are themselves 4-D perduring objects that are moral subjects,
whose temporal parts may only have tokens of moral properties.
Now, Fallen and Redeemed Humanity have human agents as
temporal parts—or, perhaps better, as phases (the term ‘stages’
having already been co-opted by another sort of four-dimensionalist,
about which more in a moment). But from this it does not
necessarily follow that Fallen or Redeemed Humanity as an object
has properties that only a moral subject may.

There is another way of understanding a four-dimensionalist

ontology that is also of help to the theologian in making sense of the
transference of sin from Adam to his progeny, and of righteousness
from Christ to his elect. This is stage theory.37 According to the
stage theorist, what we have been calling the temporal parts of 4-D
wholes—but what they call ‘stages’ of 4-D entities—do have moral
properties and it does make sense to say ‘this stage of a 4-D agent is
guilty of sinning at tn’. Stage theorists need not posit such entities as
Fallen and Redeemed Humanity either. They can tell a story
according to which God treats certain stages as one persisting (i.e.
exduring) entity, or ‘makes’ this the metaphysical truth of the
matter. Or, perhaps there are certain immanent causal relations
between stages of Adam up to his moment of original sin and all his
progeny, on the one hand, and of Christ on the cross and all the
elect, on the other, which corresponds to what I have been calling
Fallen and Redeemed Humanity, respectively. But the argument I
have o ered thus far, which depends on the sort of organic
connection between Adam and his progeny and Christ and his elect
spoken of by a number of theologians, is not quite the same as this.
Perhaps a stage-theoretic account of original sin and union with
Christ is preferable to the perdurantist story. For present purposes
three issues need to be made clear: that the 4-D ontology assumed in
the foregoing argument for realist penal substitution can be
construed in at least two ways, namely, a doctrine of temporal-parts
or stage theory; that the temporalparts construal of Fallen and
Redeemed Humanity need not end up requiring that these two 4-D
objects are moral agents of some kind; and that one could take a
stage-theoretic account rather than a temporal-parts account of
these things, making use of the same ontology understood rather
di erently. Either way, the problem of attributing moral properties
to ‘gerrymandered’ 4-D wholes may be avoided.

I have already hinted that the 4-D ontology (however this is

construed) might be thought to depend on a divine conventionalism,
which seems consistent with Scriptures such as Rom. 9, Mal. 1, and
other biblical passages that suggest there is no partiality with God,
even that God’s choice does not depend on creaturely actions (e.g. 2
Chron. 19: 7; Job 34: 19; Luke 20: 21; Gal. 2: 6; Col. 3: 25, etc.). But
then it looks as though divine convention is doing all the work,
which is yet another sort of philosophical objection to a realist penal
substitution argument conjoined with some doctrine of divine
conventionalism with respect to God bringing about certain 4-D
objects. In fact, there are two parts to this objection, which might be
put like this: Couldn’t God bring about something like realist penal
substitution, or an Augustinian realist account of the transfer of
original sin from Adam to his progeny, by divine convention and
without a doctrine of temporal parts or stage theory? And couldn’t
something much more like the commonsense view I mentioned
earlier have purchase here if God simply ordains that a realist penal
substitution, and/or a realist account of the transfer of original sin,
obtains?

We have already answered the rst of these questions in the

a rmative: there are several ways of construing the 4-D ontology
that underpins the versions of realist penal substitution we have set
forth. The argument will need to be tweaked, depending on which
construal of this ontology one opts for. But I am much more dubious
about using a di erent ontology to make sense of realist penal
substitution, such as that presupposed by the commonsense
position, according to which persisting objects are wholly present at
each moment of their existence. The obvious way forward for a
commonsense way of thinking about these things is to say that God
imputes Adam’s sin to me, and my sin to Christ. But, it should be
clear that such a way of thinking about these matters is mired in
real moral di culties, having to do with the very idea of imputing
sin from one person to another, which implies a sort of moral
and/or legal ction that appears unjust. If it is said that divine
conventionalism means God may treat any two persons or their
parts as related in certain sorts of ways—even certain sorts of moral
ways—as he sees t, this seems to me divine conventionalism run
amok. It is no part of the version of penal substitution outlined thus
far that there are no morally relevant similarities between the
temporal parts that make up Fallen or Redeemed Humanity. Nor is
it any part of this doctrine that God may simply gerrymander
temporal parts willy-nilly. For I suppose that God acts in accordance
with his divine character, doing that which seems to him most
tting.

There is much more to be said about the atonement. What I have

set out here is one theological story that connects the atonement
with the doctrine of original sin in a way that involves a somewhat
novel use of Augustinian realism, but remains faithful, I think, to
those passages in Scripture wherein the union between Adam and
his progeny on the one hand, and Christ and his elect on the other,
are treated. I have tried to indicate some of its virtues, not the least
of which is that on this view we can say with the Apostle that, ‘as by
one man’s disobedience many were made sinners, so also by one
man’s obedience many will be made righteous’ (Rom. 5: 19).

NOTES
 My thanks to Gavin D’Costa, Tom Flint, Paul Helm, Daniel
Hill, Joseph Jedwab, Brian Leftow, Mike Rea, Bill Schweitzer
and Richard Swinburne for their comments on or concerning
the argument of previous drafts of this chapter. Thanks too to
the members of the Systematic Theology Seminar in the
Faculty of Divinity, Cambridge University, The Joseph Butler
Society, Oriel College, Oxford, and the Theology Seminar at St,
Mary’s College, University of St Andrews, where earlier
versions of the latter part of this chapter were read in 2007.

1. I have dealt with two other views of the atonement elsewhere,

namely, the governmental and the idea of Christ’s vicarious
humanity (stemming from the work of the nineteenth-century
Scottish theologian John McLeod Campbell). Interested
readers should consult Crisp 2008 and 2007b respectively.
There has been renewed interest in discussion of the
atonement as essentially a sacri ce amongst contemporary
systematic theologians. There have also been a number of
recent revisionist accounts of the atonement, which I will not
enter into here. For discussion, see Fiddes 2007.

2. St Gregory of Nyssa’s Great Catechism epitomizes the ransom

view amongst the Patristic authors, whereas parts of St
Athanasius’ On the Incarnation, especially chs. 2 and 4, sound
much more like a primitive satisfaction account of the
atonement. For a standard historical discussion of the
development of atonement doctrine, see Franks 1962.
3. Aulén 1931.

4. Other atonement theories may also o er a way of making

sense of the biblical material on ransom, e.g. the governmental
theory of atonement. I am not suggesting satisfaction or penal
substitution theories are the only plausible alternatives.

5. This does have the apparently counterintuitive consequence

that all sin entails an in nite demerit, which appears to make
all sin equal for the purposes of culpability. There are possible
responses to this problem, but it would take us too far from
the main argument to discuss them. See e.g. Crisp 2003 and
the literature cited there.

6. Like St Augustine of Hippo, St Anselm does think that one

motivation for the redemption of fallen human beings is that
their number make up the number of angels that fell with the
Devil. See Cur Deus Homo 1.16–18 in St Anselm 1998.

7. ‘But the obligation rests with man, and no one else, to make
the payment referred to. Otherwise mankind is not making
recompense.’ Cur Deus Homo 2. 6. Cf. 2. 7 in St Anselm 1998:
320–1.

8. Swinburne 1989.

9. Aquinas 1948, III q. 46.a.1–4.

10. Swinburne 1989: 161.

11. One recent attempt to include aspects of Swinburne’s account

in a moderate doctrine of penal substitution can be found in
 Porter 2004.

12. Quinn 1993. See also Bynum 2004.

13. Franks 1962: 366 says that for Faustus Socinus, ‘the death of
Christ operates to redeem us in so far as it is an example of
obedience, leads us to trust God, and gives us hope of
deliverance from punishment’. The major work on the
atonement by Faustus Socinus is De Jesu Christo Servatore
(1594). A recent historical-theological reassessment of Socinus’
importance can be found in Gomes 1993.

14. Hick 1993: 112–33. Another in uential modern theological

statement of the moral exemplar view sans commitment to
religious pluralism can be found in Rashdall 1919.

15. Swinburne’s theory means God could set aside divine

retribution, though he does not. But saying God may not exact
some reparation or satisfaction from humanity for human sin
is not the same as saying God cannot exact some reparation or
satisfaction from humanity because the very idea that God
would act in this way is somehow immoral, or beneath the
divine dignity. Those moral exemplarists that make this latter
claim seem to be committed to a very di erent understanding
of divine justice, according to which Christ’s work is about
reconciliation, but not atonement.

16. This is not a novel idea. In the modern literature it is

discussed in Packer 1999.
17. Murray 1959 o ers a classic treatment of this. The
nineteenth-century Presbyterian theologian Charles Hodge has
an idiosyncratic version of the forensic ctional view that is
noteworthy. See Hodge 1940.

18. See Lewis 1997.

19. Thus Vincent Brümmer 2005: 75 ‘a theory of penal

substitution strikes many of us today as highly immoral since
it claims that God punishes the innocent for the transgressions
of the guilty’.

20. St Augustine seems to think that all humanity was somehow

‘seminally present’ in Adam. See Augustine 1972, 13. 14. For a
theological recapitulation of this view, see Berkhof 1939: 241–
2.

21. A number of Augustinians distinguish between that aspect of

guilt which is non-transferable, i.e. ‘that Trevor was the one
guilty of stabbing Dean’ and the aspect of guilt that is
transferable, having to do with liability to punishment. As we
shall see in a moment, the punishment for a crime might be
transferred to a penal substitute, though the person who
committed the crime is still the one guilty of perpetrating the
crime.

22. In fact, as we shall see, one construal of Augustinian realism

is stronger than this. On this stronger view, the transfer of
Adam’s sin and guilt to me is just provided God somehow
 organizes things such that Adam and his posterity are one
metaphysical entity at the moment of primal sin, and only
thereafter di erentiated into distinct human natures.

23. The idea that God imputes Adamic sin and/or guilt to Adam’s
progeny is found in much, though not all, Reformed theology.
‘Realism’ is a somewhat plastic term, having a variety of
di erent applications in the current philosophical literature.
But, as I indicate here, Augustinian realism uses the term in a
rather di erent way from current analytical discussions about
realism vs. anti-realism.

24. A robust defence of the imputation view can be found in

Murray 1959. An example of the Augustinian realist view can
be found in Shedd 2003: 434–82, 557–64, 630–1.

25. I take it that traducians deny a doctrine of the simplicity of

souls, according to which souls are, as Chisholm puts it,
‘simple substances’, that is, without more basic parts from
which they are composed. See Chisholm 1991.

26. I place ‘parts’ in parentheses here and in what follows

because at this stage of the argument I am not concerned to
assimilate the particular theological story concerning a realist
account of penal substitution with the doctrine of temporal
parts, which I take to be a particular metaphysical account of
perduring objects. A doctrine of temporal parts is one possible
means of underpinning this theological story. But it is not the
only means—a matter to which we shall return.
27. Strictly speaking, Fallen Humanity has parts that are not
‘fallen’, since Adam prior to his primal sin was unfallen, and
Christ does not possess a ‘fallen’ human nature (i.e. a human
nature that has original sin), though he is a ‘Son of Adam’. So
it seems that possession of original sin is a contingent property
of almost all the ‘parts’ or phases of the life of Fallen
Humanity. Alternatively, Fallen Humanity comprises the phase
of Adam’s life from his primal sin onwards and all other
humans born after Adam, except Christ, whose human nature
is miraculously preserved from being tainted with original sin.

28. Advocates of Augustinian realism often point to Heb. 7: 9–10

as evidence that Levi was (somehow) in the loins of Abraham
when Abraham paid a tithe to Melchizedek. But it seems to be
consistent with other passages of Scripture too, including Rom.
5:12–19 and 1 Cor. 15: 22.

29. Another way of making sense of this would involve using a

doctrine of backwards causation. Then, Christ’s work could
backwardsly cause the election of those who lived before him.
But I do not really understand backwards causation, and it is a
controversial idea anyway. So I shall not pursue this possibility
here.

30. Barth 1957–69.

31. Given that this doctrine is only inspired by Barth’s account

rather than being an exposition of it, it does not matter
whether it tracks all that Barth himself believed about election
 and atonement. What matters is that these notions can be
tted together into a coherent whole.

32. It may be possible to construct a version of realist penal

substitution that factors in a doctrine of traducianism. Then
Christ has a fraction of Adam’s soul that is miraculously
cleansed at the moment it is individualized, so to speak. But I
shall not outline such a view here, since I am dubious about
traducianism.

33. And lest we forget, the problem of Christ being temporally

subsequent to some of the elect is a problem common to all
accounts of Christ’s work that suppose Christ’s death is an
atonement for sin. He lives before some of his elect, and after
others, but somehow his death has consequences for all the
elect existing at many di erent moments in time, before and
after Christ.

34. This could be taken in a more radically ‘realist’ direction

where Christ is really united with sinners, and really becomes
guilty for their sin. Then justice is truly being carried out at
the cross, exacting punishment for one guilty of the sin of
other temporal parts of the same Redeemed Humanity, not
wrongfully a icting the innocent. But this radically ‘realist’
doctrine does also have the unwelcome consequence that
Christ is truly guilty of sin. And that seems theologically
untenable. So I shall not pursue this option here.
 35.I suppose one could o er a ‘ ctionalist’ account of Fallen and
Redeemed Humanity, according to which these entities are not
real but ctional, or entities that are in some sense merely
conventional. But it is di cult to see how this would be an
improvement on those forensic ction accounts of the
atonement according to which Christ represents me, and is
treated as if he is the one culpable for my sin though he is not,
strictly speaking.

36. See Crisp 2007a: 102 n. 7 for references.

37. For an excellent treatment of this, see Rea 2007.

REFFEREMCE

ST ANSELM OF CANTERBURY (1998). St Anselm of Canterbury, The Major

Works, ed. Brian Davies and Gillian Evans. Oxford: Oxford
University Press.

AQUINAS, ST THOMAS (1948) [1911]. Summa Theologica, trans. Fathers

of the English Dominican Province. IIa IIae qq. 149–89; IIIa qq.
1–73. New York: Benziger Brothers, iv.

ST AUGUSTINE OF HIPPO (1972). City of God, trans. Henry Bettenson,

Introduction by John O’ Meara. Harmondsworth: Penguin
Books.

AULEN, GUSTAV (1931). Christus Victor. London: SPCK.

BARTH, KARL (1957). Church Dogmatics, ed. G. W. Bromiley and T. F.
Torrance. Edinburgh: T&T Clark, ii pt. 2.

BERKHOF, LOUIS (1939). Systematic Theology. Edinburgh: Banner of

Truth.

BRUMMER, VINCENT (2005). Atonement, Christology and The Trinity,

Making Sense of Christian Doctrine. Aldershot: Ashgate.

BYNUM, CAROLINE WALKER (2004). ‘The Power in the Blood: Sacri ce,
Satisfaction, and Substitution in Late Medieval Soteriology’, in
Stephen T. Davis, Daniel Kendall, and Gerald O’Collins (eds.),
The Redemption. Oxford: Oxford University Press, 177–86.

CHISHOLM, RODERICK M. (1991). ‘On the Simplicity of the Soul’, in

James Tomberlin (ed.), Philosophical Perspectives v. Philosophy of
Religion (1991). Atascadero: Ridgeview, 167–81.

CRISP, OLIVER D. (2007a). An American Augustinian: Sin and Salvation

in the Dogmatic Theology of William G. T. Shedd. Milton Keynes:
Paternoster; Eugene, Ore.: Wipf & Stock.

———(2003). ‘Divine Retribution: A Defence’, in Sophia 42: 35–52.

———(2007b). ‘Non-penal Substitution’, in International Journal of

Systematic Theology 9: 415–33.

———(2008). ‘Penal Non-Substitution’, in Journal of Theological

Studies NS 59:140–68.

FIDDES, PAUL F. (2007). ‘Salvation’, in John Webster, Kathryn Tanner,

and Iain Torrance (eds.), The Oxford Handbook of Systematic
 Theology. Oxford: Oxford University Press, 176–96.

FRANKS, R.S. (1962) [1918]. The Work of Christ. Edinburgh: Thomas

Nelson & Sons.

GOMES, ALAN W. (1993). ‘De Jesu Christo Servatore: Faustus Socinus

on the Satisfaction of Christ’, in Westminster Theological Journal
55: 209–31.

HICK, JOHN (1993). The Metaphor of God Incarnate, Christology in a

Pluralistic Age. Louisville: Westminster John Knox.

HODGE, CHARLES (1940) [1871]. Systematic Theology, ii. Anthropology.

Grand Rapids: Eerdmans.

LEWIS, DAVID (1997). ‘Do We Believe in Penal Substitution?’, in

Philosophical Papers 26: 203–9.

MURRAY, JOHN (1959). The Imputation of Adam’s Sin. Grand Rapids:

Eerdmans.

PACKER, JAMES I. (1999). ‘The Logic of Penal Substitution’, in The J.I.

Packer Collection, ed. Alister McGrath. Downers Grove, Ill.: IVP.

PORTER, STEVEN (2004). ‘Swinburnian Atonement and the Doctrine of

Penal Substitution’, in Faith and Philosophy 21: 238–9.

QUINN, PHILIP L. (1993). ‘Abelard on Atonement: “Nothing

Unintelligible, Arbitrary, Illogical, or Immoral about it” in
Eleonore Stump (ed.), Reasoned Faith. Ithaca, NY: Cornell
University Press, 281–300.
RASHDALL, HASTINGS (1919). The Idea of Atonement in Christian
Theology. London: Macmil-lan.

REA, MICHAEL C. (2007). ‘The Metaphysics of Original Sin’, in Peter

van Inwagen and Dean Zimmerman (eds.), Persons: Human and
Divine. Oxford: Oxford University Press, 319–56.

SHEDD, WILLIAM G. T. (2003). Dogmatic Theology, Third Edition, ed.

Alan W. Gomes. Phillipsburg, NJ: Presbyterian & Reformed.

SWINBURNE, RICHARD (1989). Responsibility and Atonement. Oxford:

Oxford University Press.
 CHAPTER 20

THE INCARNATION

RICHARD CROSS

THE Christian doctrine of the Incarnation maintains that the second

person of the Trinity became a human being, retaining all attributes
necessary for being divine and gaining all attributes necessary for
being human. As usually understood, the doctrine involves the claim
that the second person of the Trinity is the subject of the attributes
of Jesus Christ, the rst-century Jew whose deeds are reported in
various ways in the New Testament. For most theologians, this
requires the further claim that the second person of the Trinity is
identical with Jesus Christ. It certainly seems to have been taken as
such by the Council of Chalcedon (451 CE), generally understood to
be the standard for orthodox teaching on the matter. The Council
teaches that

one and the same Christ, Son, Lord, only-begotten, [is]

acknowledged in two natures which undergo no confusion, no
change, no division, no separation;…the property of both natures is
preserved and comes together into a single person and single
subsistent being (hypostasis); he is not parted or divided into two
persons, but is one and the same only-begotten Son, God, Word,
Lord Jesus Christ. (N. P. Tanner 1990: i. 86*)

The view of Chalcedon, then, is that one and the same person—the
second person of the Trinity (the ‘Word’, identical with the ‘Lord
Jesus Christ’, according to the nal list of Christological titles)—
exists ‘in two natures’, divine and human.1 One of the views that the
Council opposes is that known as Nestorianism, according to which
there are somehow two distinct subjects in Christ—one divine, and
one human. The point of rejecting this view is to ensure that the
divine person—the second person of the Trinity—is the subject not
only of divine but also of human attributes or properties. Now, the
Council does not de ne its terms, and some recent theological
commentators maintain that it gives no more than particular
syntactic rules governing the use of the terms ‘person’ and ‘nature’.
But it seems to me—as it has seemed to most recent philosophical
commentators—that the linguistic rules are secondary to some
apparently substantive metaphysical claims: namely, that there is
some divine subject of properties—the second person of the Trinity
—that gains human properties (a human nature). Clearly, the
metaphysics leaves open just what a nature is—for example, a
universal or a particular—and as we shall see, this issue crops up in
a number of recent discussions too.

As thus put, the doctrine raises a series of issues of general

interest to philosophers: for example, questions of the relation
between individuals and their individual-essential and kind-essential
natures, and of identity across time. But the fundamental
philosophical problem speci c to the doctrine is this: how is it that
one and the same thing could be both divine (and thus, on the face
of it, necessary, and necessarily omniscient, omnipotent, eternal,
immutable, impassible, and impeccable) and human (and thus, on
the face of it, have the complements of all these properties). Most of
the philosophical e ort devoted to the speci cs of the doctrine has
centered on this issue, and I shall concentrate on it below. In
addition to this array of topics, there are a number of other
potential problems with the doctrine. Some of these were nicely
focused in a book that in e ect kick-started the modern
Christological debate. Thirty years ago a group of British
philosophers and theologians published The Myth of God Incarnate
(Hick (ed.) 1977) in which they outlined a series of objections to the
orthodox doctrine of the Incarnation: in particular, the coherence
objection just mentioned, and in addition worries about the
particularity of the Incarnation (given the vastness of the universe),
and concerns about the lack of—or impossibility of—evidence for
the Incarnation. But I shall focus here on the coherence objection, as
indicated.2

I have mentioned the Council of Chalcedon as a formative event

in the history of the doctrine. Recent work in the history of
Christian doctrine has tended to emphasize the open-endedness of
Chalcedon, and its (historical) failure to reach any kind of resolution
of the various Christological controversies of the early- to mid- fth
century. Philosophers of religion often (though not always) pass
over the equally important developments at the Third Council of
Constantinople (680-1 CE). Chalcedon teaches that Christ has all the
properties necessary for being divine, and all the properties
necessary for being human. Constantinople III clari es that these
properties must include both divine causal powers (energy and will)
and human causal powers (energy and will). The point might sound
obvious, but anyone who wants to associate causal power—or mind,
for that matter—with personal identity might nd it di cult to
assert the existence of two irreducibly distinct varieties of such
power in one person, or two irreducibly distinct minds in one
person. As we shall see, the kinds of position defended at
Constantinople III are central to at least one very important
development in modern philosophical Christology—the insight that
Christ must in some sense have two minds.

Before considering in detail the approach to the coherence

problem adopted by philosophers of religion in the analytic
tradition, it is worth considering a suggestion made by certain
philosophically inclined theologians who, while they would not
count themselves as part of this philosophical tradition, nevertheless
have some sympathy for it and adopt approaches that are on
occasion analogous to it. I have in mind the suggestion made by
thinkers such as Herbert McCabe, Kathryn Tanner, and David
Burrell, that the whole project of thinking of the issue in terms of a
problem about contradictory properties is misconceived. The two
natures, divine and human, are not competing for existence in the
same ‘logical world’ (McCabe 1987: 57); giving an account of the
metaphysics of the Incarnation is not a case of practising a ‘zero-
sum game’ (Burrell 2004: 469). Thus Kathryn Tanner (2001:11):
‘Because divinity is not a kind, God is not bound by apparent
contrasts between divine and creaturely qualities; God is thereby
free to enter into intimate community with us, without loss to the
divine nature.’ Why is this? Tanner (ibid. 15) explains by means of
an example: ‘Divine omniscience is not something like human
knowledge, just better, minus the ignorance or the limitations of
nitude.’ Presumably the point is that the divine attributes are such
that they are neither contradictories nor contraries of the relevant
human attributes. The problem with this is fundamentally one of
plausibility: whatever divine omniscience amounts to, for example,
it is hard to see how it can be compatible with the kinds of
ignorance that Jesus exhibited on earth. Neither does a mere appeal
to God’s (allegedly) not belonging to a kind su ciently secure the
conclusion that his attributes are such that they fail to be the
contraries of corresponding creaturely attributes—indeed, if they
fail to be the contraries of corresponding creaturely attributes, we
might reasonably wonder whether there is such a thing as divine
transcendence, or, alternatively, whether our doctrine of God is not
so mysterious as to amount to nothing much at all.

Still, there is a substantive point here. We do not know—and

certainly do not know in advance of the doctrine of the Incarnation
—what the complete list of divine and human essential attributes
consists in. So we do not know in advance whether or not the list
contains contradictory pairs. So the agnostic strategy just outlined
could under lie at least two possible approaches to the doctrine that
have proved popular in recent debate, one of which uses the
doctrine to restrict the list of kind-essential human attributes, the
other of which uses the doctrine to restrict the list of kind-essential
divine attributes—both discussed in sect. 4 below. Of course, it
would be just as possible to combine these two strategies in some
shape or form, and I shall argue below that this may well turn out to
be the best way forward. In what follows, then, I shall try to outline
very brie y the most important recent attempts by philosophers of
religion to deal with the fundamental Christological problem that I
have identi ed here, and to come to some kind of assessment of the
value of each attempt.

1. REDUPLICATION AND ADYERBIAL MODIFIERS

The Council of Chalcedon suggests one way of dealing with putative

contradictions. It claims of Christ that he was ‘begotten before the
ages from the Father as regards his divinity (kata tēn
theotēta/secundum divinitatem), and in the last days the same for us
and for our salvation from Mary the virgin God-bearer (theotokou) as
regards his humanity (kata tēn anthropotēta/secundum humanitatem)’
(N. P. Tanner 1990: i. 86*). This suggests that we can defuse the
problem by dividing potentially contradictory properties into two
groups, associating one with the divine nature, and the other with
the human nature. There are various ways of signalling gramatically
the kind of adverbial modi cation suggested by Chalcedon. I shall
talk of quaquali ers: thus ‘S qua N is F.’ Thomas Senor has recently
summarized the three possible construals of such adverbial
modi ers known in the tradition, and I shall add a couple more as
well in this section and the next.

The rst is the one that the medieval theologians knew as the
‘reduplicative’ analysis (see Cross 2002: 193-5). On the
reduplicative analysis, ‘S qua N is F’ should be understood as ‘“In
virtue of being N, S is F”’ (Senor 2002: 229). This understanding, as
the medievals noticed, does nothing to help deal with putative
contradictions:

Sentences of the form ‘in virtue of N, S is F’ entail sentences of the

form ‘S is F.’ So ‘In virtue of being divine, Christ is omnipotent’
entails ‘Christ is omnipotent’; and ‘In virtue of being human, Christ
is not omnipotent’ entails ‘Christ is not omnipotent.’…Hence,
understanding reduplicative sentences like this will not help defend
orthodoxy because in the end we will still be left with simpliciter
properties and the contradictions they promote.
(ibid.)

Thomas Morris clari es:

Consider any conjunctive reduplicative proposition of the form ‘x as

A is N and x as B is not N.’ If the subjects of both conjuncts are the
same and the substituends of N are univocal across the conjunction,
then as long as (1) the reduplication predicates being A of x and
predicates being B of x, and (2) being N is entailed by being A, and
not being N is entailed by being B, then the reduplicative form of
predication accomplishes nothing except for muddying the waters,
since in the end the contradiction stands of x being characterized as
both N and not N. (Morris 1986: 48-9)

In fact, this is not the only form of reduplication that is possible.

Douglas K. Blount has objected to Senor’s attempt to dismiss
reduplication by noting that the in virtue of’ analysis is not
appropriate for some cases that seem very close to those that Senor
considers. For example,

When one claims that

Jesus Christ qua man read in the synagogue before Jesus Christ
qua man carried his cross,

One typically does not mean to be claiming that

Jesus Christ in virtue of being human read in the synagogue before
Jesus Christ in virtue of being human carried his cross. (Blount
2002: 239, original emphasis)
This is clearly true, at least if we understand the ‘in virtue of’
locution to state a su cient condition, and shows that not all cases
can be analyzed in just the same way. Still, even in these cases the
reduplication provides some sort of explanation for the truth of the
predication—as in the more standard reduplicative case considered
 rst. In fact, the reduplication here seems to signal something like
an INUS condition: an insu cient but necessary member of a set of
unnecessary but jointly su cient conditions (hence ‘INUS’). Christ’s
being human is an INUS condition for his reading in the synagogue,
and for his carrying the cross. But he still read in the synagogue,
and carried his cross. So this secondary sort of reduplicative
sentence does not provide a way to prevent the ascription of the
predicates to their subjects simpliciter; thus it does not prevent one
and the same person being characterized in contradictory ways.

Marilyn McCord Adams has recently argued in a similar way,

proposing what is in e ect a further variant on the reduplicative
strategy. Because Christ’s human nature is non-essential to him, we
can claim that Christ has simpliciter all the attributes of deity, but
that he has human attributes only in a quali ed way. So Christ is
simpliciter omniscient, and limited in knowledge qua human; this is
su cient to circumvent the putative contradiction. According to
Adams (2006: 135-6), the relevant attributes are the same as in their
usual contexts; but the way in which the human attributes are
possessed is di erent in the case of Christ from that of all other
human beings. According to Adams, ‘qua’ here means ‘by virtue of’;
but ‘by virtue of’ is ambiguous, and in these cases ‘N is the nature by
virtue of which x is F because N entails the real possibility of F’
(ibid. 136-7). Still, this strategy alone does not seem su cient to
block the alleged contradiction, since the obvious Christological
problem is not just that Christ is possibly ignorant or mistaken, but
that he is actually ignorant or mistaken; qua’ must amount to more
than just a modal quali cation of the kind proposed by Adams. In
any case, essential omniscience is incompatible with even possible
ignorance.

On Senor’s second way of understanding qua-propositions, we

should take the qua-clause

to be imbedded in the grammatical subject of the sentence. We can

make this clear by appropriately placing hyphens. ‘Qua N, S is F’
becomes ‘S-qua-N is F.’ The properties of omnipotence and non-
omnipotence [for example] are predicated simpliciter, but
contradiction is avoided because the properties are predicated of
di erent subjects, i.e., Christ-qua-God and Christ-qua-human. (Senor
2002: 229-30)

Senor comments that the cogency of this grammatical analysis

seems to presuppose the Nestorian claim that there are two real
subjects in Christ (ibid. 230). This may be a little quick. For the
medieval theologians, the kind of analysis that Senor is describing
was known as the ‘speci cative’ analysis, and they usually took it as
a mark of synecdoche—ascribing to a whole a property of a part of
that whole (see Cross 2002: 195-203). As far as I can see the
cogency of this analysis does require some theological
commitments: namely that we adopt a mereological account of the
person of Christ, or something that draws on a close analogy to such
accounts. I return to this in sect. 3 below, where I shall try to judge
whether or not such mereological Christologies are eo ipso
Nestorian.

On a third way of understanding qua-sentences proposed by

Senor, also known to the medievals (see ibid. 203-4), we should
understand

‘S qua N is F’ as ‘“S is F-qua-N.” That is, the qua-clause gets packed

into the predication in such a way as to form as single, non-
compositional, simple property. The F-entailment is blocked here
because the properties being ascribed bear no non-trivial logical
relation to each other’ (Senor 2002, 230).

Of course, coherence here comes at a price: speci cally,

considerable agnosticism about what we mean when we make
Christological (and other) claims. For take the locutions
‘incorporeal-qua-God’ and ‘corporeal-qua-man’. Is incorporeality-
qua-God a determinate of incorporeality, and corporeality-qua-man
a determinate of corporeality? If so, then we simply have another
variant of reduplication properly so called, and the entailment to a
contradiction will go through. If not, then it is not clear what the
attributes are supposed to be. We could claim that our agnosticism
extends only as far as the divine attributes, and adopt a strategy
similar to that proposed by Kathryn Tanner in the discussion cited at
the beginning of this chapter. We could claim, for example, that
Christ is corporeal-simpliciter, and incorporeal-qua-God, where
‘incorporeal-qua-God’ picks out the unknown attribute that is God’s
incorporeality. But we need to understand what we are committing
ourselves to if we adopt this approach: if being incorporeal-qua-God
is compatible with being corporeal, we really have no idea what
‘incorporeal-qua-God’ means. And the more attributes we add to the
list, the less we will be able to talk meaningfully about God at all.
And I do not think that many theologians would want to make such
a strong claim, even at the price of preserving some kind of formal
coherence to Christological formulations.

2. RELATIVE IDENTITY SOLUTIONS

One way of utilizing these qua-quali cations is by appealing to the

notion of relative identity in this context. Adherents of relative
identity claim that identity is ‘sortal relative’, such that

‘a and b could be the same F but not the same G (where “F” and “G”
stand in for sortal terms, or general count nouns with associated
criteria for individuation and reidenti cation)’ (Morris 1986: 28).

Peter van Inwagen, for example, has argued that the doctrine of the
Incarnation can be shown to be ‘free from formal contradiction’ (van
Inwagen 1994: 225) if we adopt such an account, such that two
distinct beings could be the same person (ibid. 217). Equally, the
position allows for one being to be (e.g.) human and the other non-
human without contradiction, and one being to be (e.g.) passible
and the other impassible without contradiction, while yet allowing
both beings to be one and the same person (ibid. 221-3). This is a
version of the speci cative analysis outlined in the previous section,
since it treats the two beings in Christ as subjects distinct from each
other, though not from the divine person with which each severally
is identical. But unlike that analysis, it does not rely on a
mereological account of the Incarnation, since the two beings
postulated to be the same person in van Inwagen’s account are not
parts of that person. Rather, each being is severally the same person
as the second person of the Trinity and as Jesus of Nazareth (ibid.
217-18).3

There are various ways in which this kind of approach can have
some appeal. For example, it could do so if we were to adopt some
more wide-ranging anti-realism: if we were to adopt, for example,
the view that there are no theory-independent facts about the
natures of things, or the distinctions of one thing from another.
Many, though not all, theologians nd such anti-realism at the very
least counterintuitive, and in any case some anti-realists would be
more than happy to deny relative identity: relative to each theory
about how to cut up the world, identity remains absolute. In fact,
van Inwagen’s own working out of the doctrine along these lines
does not involve making any explicitly anti-realist claims. What it
does involve is accepting a logic of relative identity that lacks a rule
permitting inferences that conform to the principle of the
indiscernibility of identicals. The logic thus licences the use of
premises that can violate the principle of the indiscernibility of
identicals: something that most philosophers would be reluctant to
accept. So van Inwagen’s approach is ultimately of very limited
appeal. And as he himself notes, even if we accept his account,
much work would be left to be done: I [do] not claim to have
penetrated the mystery of the Incarnation, but at most to have
shown that that doctrine can be stated without formal contradiction’
(ibid. 202). The interest of the logic turns out to be restricted to its
application—and the philosophical price would be felt by many to
be too high.

3. PARTS AND PROPERTIES

We often talk of things having the same features as each other—of

two or more things sharing certain features. And we usually talk not
merely of things having such-and-such features, but of being such-
and-such. In such contexts, we are likely to claim that we are
committed to there being properties of things, whatever those
properties be. In more technical jargon, we claim that these
properties are instantiated or exempli ed. For some philosophers,
properties are universals, shared by the various particulars that
exemplify them. On this view, to explain the truth of a conjunction
such as ‘Socrates is white and Plato is white’ we need to posit some
entity, whiteness, in some sense the same (usually numerically the
same) in both Socrates and Plato. Other philosophers hold that the
properties are as particular as the things of which they are
properties—that they are analogous to parts, but parts of a very
abstract kind. These ‘thin’ parts nowadays usually go under the
designation of ‘tropes’. Still other philosophers think that properties
are merely linguistic items, things that can be predicated of, or said
of, particulars. Whatever our view of properties, we should in this
context make a distinction between properties and substances.
Substances are concrete objects, with causal powers; properties are
abstract, and lack such powers. They do not interact with other
things in the world in the way that substances do (though some of
them could of course be the causal powers of substances).

Some thinkers claim that all properties are parts of substances.

But whatever we think about this, it is more important for our
purposes here that some parts of substances are not properties but
rather (themselves) substances.4 For there are some Christologies
that explicitly rest on the claim that the two natures could count as
parts rather than properties—and thus as (in some sense) substances
(rather than properties: concrete, not abstract, items). Setting aside
for a moment the distinction between substances and parts, the
relevance of the universal property/particular property distinction
to the doctrine of the Incarnation is well known. Richard
Swinburne, for instance, claims that we should think of a nature—or
at least the human nature—as some kind of universal, on pain of
Christological incoherence. An individual nature

‘would be the essential core of the individual—what made him who

he is—and could not be possessed temporarily or accidentally by
anyone else. Christ therefore could not have an individual human
nature. His human nature must be universal, in no way peculiar to
Christ—it is just a set of properties which he acquires’ (Swinburne
1994: 212-13; see too Williams 1967).

Now, one obvious way of thinking of the Incarnation is to think of

the two natures not as properties but as (concrete) parts of the
divine person. Making such a move involves denying Swinburne’s
claim that an individual nature is in every case the ‘essential core of
the individual’. This strategy was adopted by, for example, many of
the medieval theologians (for some of the ways in which they
blocked Swinburne’s inference, see Cross 2002: 246-309). Indeed,
talk of parts in the Christological context is initially quite appealing
as a way of dealing with putative contradictions, since the di erent
properties can be attributed to the di erent parts or substances that
(on a mereological analysis) the two natures are. As I mentioned
above, the usual way of understanding the speci cative sense of
‘qua’ is mereologically, and in line with this we can assert that the
two natures are parts of the whole person, and ascribe the relevant
properties to the person by piggy-backing’: the properties piggy-
back from the parts to the whole.

Eleonore Stump has developed this view with considerable

sophistication, starting from some common-sense assumptions: ‘On
the [speci cative] strategy, the attributes that are incompatible with
each other are also segregated from each other in the incarnate
Christ in virtue of inhering in di erent natures of his’ (Stump 2003:
412). On this view, not only are the natures in some sense real
constituents of a person, but also they are themselves the subjects of
properties—and thus in some sense substances. It is in virtue of a
property’s inherence in a nature that it can inhere in a person. As
Stump notes,

there is a distinction between a property a whole has in its own

right and a property it has in virtue of having a constituent that has
that property in its own right; it is possible for a whole to ‘borrow’ a
property from one of its constituents. This distinction between ways
in which a whole can have a property gives us one helpful manner
in which to analyze locutions of the form x qua A is N. In such a
locution, the property of being N is predicated of x, but it is
predicated of x just in virtue of the fact that x has a constituent C
which has the property of being N in its own right. (ibid.)

In the case of the Incarnation, then, we can think of the two

natures, divine and human, as somehow or other parts of the
incarnate second person of the Trinity. This mereological account is
not Nestorian. On the one hand, it does not preclude some kind of
duplication of di erent properties of the same basic kind (two wills,
or two minds, for example—something I return to below). On the
other, it also allows that all properties are properties of the person,
and thus avoids positing a complete duality of ultimate subjects in
Christ. Clearly, not every kind of constituent and every kind of
property can allow piggy-backing of properties in the right kind of
way. (Consider quantitative predicates.) But this is perhaps not
surprising: di erent properties have di erent natures, which might
interact di erently with the part-whole relation.

There seems to be a problem with Stump’s proposal nevertheless.

If we deny that an individual nature is the ‘essential core’ of a
person, it is perhaps easy enough to see how the human nature
could be a concrete particular part of a divine person. But it is
harder to see how the divine nature could be such a concrete part.
The divine nature or divinity, in standard Christian orthodoxy, is
supposed to be shared somehow by the divine persons, such that in
virtue of their possession of divinity, each person is said to be
divine. It will thus be more like a property than like a part, at least
in the kinds of sense outlined above. For it is hard to see how any
such shared object could be a concrete part: we usually think of
such shared objects, those in virtue of which something is said to be
such-and-such, as paradigm cases of abstract objects. Suppose that
the divine nature is a concrete substance, and suppose we claim that
(for example) this substance is impassible. On the Christological
solution I am now considering, the divine nature is an impassible
substance, while a divine person (at least the second person of the
Trinity) turns out to be a passible substance, since he borrows
passibility from his human substance. But in this case, what is it in
general for a divine person to have a divine nature? Is it perhaps for
the person to ‘borrow’ divinity, and (for example) impassibility,
from the nature? This is certainly suggested by the Christological
application. But this makes it look as though there are four concrete
objects in God (the three divine persons and the concrete divine
substance), something that would take considerable e ort to defend
from charges of quaternitarianism.

One way of developing the insight that the divine essence is a

concrete object is by maintaining that the relation between the
divine essence and the divine persons is some kind of inclusion
relation: the divine essence is more than any one person. For
example, suppose that each person is a sphere of consciousness of
one divine substance. But the substance is not, on this view, a part
of the person (as it would need to be on Stump’s understanding); if
anything, the mereological relationship would go the other way
round.

It is important to keep in mind here that the whole Christ is, on

the Chalcedonian de nition as I am reading it, identical with the
second person of the Trinity. Before the Incarnation, this person
comprises the divine essence along with whatever property
distinguishes the person from the other two divine persons—
standardly, liation: the relation of being generated/begotten. After
the Incarnation, this person comprises the divine essence, personal
property, and human nature. This identity claim is important from
an orthodox point of view, but it seems to be denied in a signi cant
recent mereological account of the Incarnation. According to Brian
Leftow, Christ is a whole whose parts are the second person of the
Trinity, a human body, and a human soul. This avoids the problem
that I identi ed above in Stump’s proposal—namely, the di culty
of seeing how the divine nature could be a concrete part of anything.
Here, the divine person is a concrete part of Christ. Leftow identi es
Christ as a whole whose parts are those outlined because he believes
that by so doing he can make full use of qua-quali ers: we can
ascribe divine and human attributes to Christ because Christ is a
whole made up of the divine person and a human body and soul
(Leftow 2002: 288-9).

Leftow is clearly right to think that his view allows full use of
qua-quali ers, but his way of construing the part-whole relationship
seems nevertheless to engender certain Christological di culties. In
particular, on his proposal it is hard to see how the second person of
the Trinity could be the subject of human attributes. We can see this
by considering a close relation to Leftow’s view found in the early
church, namely the view of Nestorius’s (explicitly condemned at the
Council of Ephesus in 431 CE) that we should hold not that Mary
was the mother of God (theotokos), but rather that she was the
mother of Christ (Christotokos). The claim that Mary was mother of
Christ is true, according to Nestorius, in virtue of the fact that Mary
was the mother of the man Jesus who, along with the second person
of the Trinity, constituted Christ, such that the second person of the
Trinity is a part of Christ, not identical with him. The problem here,
identi ed by Nestorius’s main orthodox opponent, Cyril of
Alexandria, is that the second person of the Trinity is not himself
the subject of human attributes. (Hence Mary is not his mother.) I
conclude that Leftow’s impressive strategy avoids the problems of
contradiction only at too high a price—the danger of a Nestorian
denial that the Son of God (as opposed to the whole of which he is a
part) is human, or has human attributes, at all.

Underlying Leftow’s account is a belief that it is not possible to

be a human person without ‘“owning” a human body, soul, mind
and will’ (ibid. 278). Body and soul are clearly concrete particulars,
parts of a person. Thus, as Leftow puts it, ‘abstract-nature
incarnation takes place only if concrete-nature incarnation does’
(ibid. 279). There seems to me to be reason to doubt this. For we
might hold that (whether we think of natures as particulars or
universals) abstract-nature incarnation merely amounts to this: that
the second person of the Trinity becomes a body, or becomes a soul,
or becomes something that is identical with the conjunction of body
and soul (this last being, as we shall see, Swinburne’s view—a view
that seems to defend abstract-nature incarnation without concrete
nature incarnation). The case would be analogous to that of any
human person, given relevant claims in the philosophy of mind. The
divine person begins to exemplify human nature, such that the only
concrete object is the person.

In an important paper, Trenton Merricks has suggested

something very like this. He presents two options. Suppose a dualist
account of human nature is true, such that a human person
somehow includes both body and soul. Merricks assumes an account
of dualism that identi es a human person as a soul related to one
body in such-and-such a way, and not related to any other body in
just that way. So rather than say that the Son of God, in the
Incarnation, gains a (concrete) human soul, we should rather say
that the Son of God becomes a human soul (Merricks 2007: 293). Or,
if we make physicalist assumptions about human nature, denying
that a human person includes a soul, then we should claim not that
the Son of God gains a body, but that he becomes a body:

Physicalism has a straightforward account of embodiment. You have

a body if and only if you are identical with that body. I assume that,
in the Incarnation, God the Son is related to the body of Jesus just
as you and I are related to our respective bodies. So, given
physicalism, God the Son, in the Incarnation, is identical with the
body of Jesus. (ibid. 294)

It would be possible to argue similarly on the assumption that a

human person is a composite of body and soul: the divine person
becomes a human body-soul composite. (I made a similar suggestion
a few years ago in Cross 2003, and, as we shall see, something
analogous is defended by Swinburne.) I take it that, depending on
the preferred theory of properties, all of these could count as cases
of abstract-nature assumption without concrete-nature assumption.
On the physicalist view, for example, the sum of physical parts that
is Christ constitutes a human person (the second person of the
Trinity), not a concrete human nature. As Merricks (ibid. 294) puts
it, ‘to say that the Son became a physical object is just to say that he
came to have physical properties’. On theological grounds, Merricks
prefers the physicalist option, since he believes that, on dualist
assumptions, it is impossible to give an account (other than an
entirely ad hoc and unsatisfactory one) of the way in which the Son
of God could come to have only the body of Jesus (ibid. 282-94). On
the physicalist view, abstract-nature assumption without concrete-
nature assumption entails that there are no parts had in common by
one and the same person pre-and post-incarnation. Putting it
bluntly, a person who previously lacked any proper physical parts
comes to be constituted of a collection of such parts. Something
analogous obtains in the case of the dualist options too. If a human
person is a soul, then the divine person becomes a soul; if a person
is a body-soul composite, the divine person becomes that composite
(on this see Cross 2003, and the account of Swinburne, that I discuss
below). Before the incarnation, the person had no proper parts; after
it, he comes to be constituted of body and soul as proper parts. It is
not clear to me that this is incoherent. It is not obvious that
something at rst part-free could not come to gain parts over time—
be those parts entirely physical, or physical and non-physical.

Clearly, this abstract-nature approach will not in itself better

allow for putatively contradictory properties to belong to one and
the same thing than any of its rivals; neither is it intended to do so
by its originators. I suggested above that this is the fundamental
problem for a defence of the coherence of the Incarnation, and thus
such views do not address the central problem raised by the
doctrine of the Incarnation. (This does not entail that the view could
not allow for, say, more than one mind or will in Christ: as
suggested by some of the church Fathers, Christ may have a series of
distinct human wills, and he may too have a highly unusual and
unique will associated with his divine status.) Thus Merricks (2007:
299): ‘Physicalism is silent on how to reconcile Christ’s divinity with
his apparently not knowing the hour of his return (Matt. 24: 36)’.
And I doubt that any view that maintains that the natures are akin
to properties (rather than concrete parts) will allow for such
putatively contradictory properties, be these properties universals or
particulars (tropes). For there simply are not di erent parts—in
e ect di erent substances—to which we could ascribe the relevant
inconsistent properties.

4. RESTRICTION SOLUTIONS

A range of treatments of the problem concedes that the only way

forward is ‘to trim’ the list of attributes necessary for being human
and/or divine. More popular are those that trim the list of human
attributes. As Swinburne notes:

Our criteria for use of the word ‘human’ are just not precise enough
for there to be a right answer as to whether the possession of
powers far beyond those of normal humans [e.g. a power to move
mountains on distant continents just by willing, or to know what is
going on in distant galaxies without using a telescope or listening to
what others tell him] would rule out someone from being a human.
(Swinburne 1994: 28)

The most extensive example of the use of this strategy is by Thomas

Morris. Morris makes a signi cant distinction between being
‘merely’ human, and being ‘fully’ human:

The kind-nature exempli ed distinctively by all human beings is

that of humanity. To be a human being is to exemplify human
nature. An individual is fully human just in case he fully exempli es
human nature. To be merely human is not to exemplify a kind-
nature, a natural kind, distinct from that of humanity; it is rather to
exemplify humanity without also exemplifying any ontologically
higher kind, such as divinity. (Morris 1986: 66)

The point here is that something that is fully but not merely human
lacks certain imperfect properties commonly (but not necessarily)
associated with human beings: for example, contingency,
peccability, fallibility, limitations on power. These properties are
had by persons that are merely human, but not by other human
persons. It seems to me that this is the correct way forward, and in
principle there seems to be no reason why we should assume that a
being that fails to exemplify these common limitations fails to be
human.

Still, it seems that we need to modify too our classical

understanding of the divine nature, and perform some analogous
‘trimming’ here as well. For example, many theists hold God to be
immutable and impassible. Clearly, anything that is human is
mutable and (on the face of it) passible. These are modal attributes,
and thus had (if at all) throughout the duration of the existence of
whatever possesses them. I infer, therefore, that, in the absence of
the success of the kinds of solution canvassed in the rst three
sections of this chapter, no divine person is immutable or impassible
(other than in the weak sense of being indestructible and morally
reliable, and being such that nothing happens to him that he could
not avoid). The same goes for timelessness. Anything that is human
is temporal; nothing can be rst timeless and then temporal; no
divine person is essentially non-human; so no divine person is
timeless. And if it be doubted that anything that is human is
temporal, it at least cannot be doubted that Christ was, since he
changed through time, sometimes eating, sometimes drinking,
sometimes laughing, and sometimes weeping. As Morris (ibid. 70)
notes, there are still further problems too. For example, there are
some scriptural data that seem to suggest that the Son of God was
ignorant (Matt. 24: 36) and perhaps even fallible (Matt. 16: 28), and
that he was subject to temptation (which latter might require either
the mistaken belief that he could sin, or indeed that he could sin,
and so that he was either fallible or peccable: see Heb. 4: 15).

Some theologians infer that the Son of God gave up his

omniscience, and perhaps infallibility, when he became incarnate.
And it is not infrequent for these same theologians to hold (again on
the basis of scriptural data) that the incarnate Son of God failed to
be omnipotent too. These sorts of Christology are known as
‘kenotic’, on the basis of the early Christian hymn in Phil. 2: 5-11,
according to which Christ, in becoming incarnate, ‘emptied’
(ekenōsen) himself (Phil. 2: 7). Curiously, the clearest and best
exponent of this view is a thinker who emphatically rejects it,
namely Thomas Morris. Morris (ibid. 94-5) considers three ways of
developing the position. First, we could simply assume that the
relevant divine attributes are non-essential. God is merely
contingently omniscient and omnipotent, such that any divine
person could fail to exemplify these properties.5 (Perhaps they could
not all do so at once, or at least do so and continue to sustain the
universe in existence, since these properties may be required for
sustaining it.) If we do not want to abandon the claim that God has
his attributes necessarily, then we would claim instead that each
relevant divine attribute has a ‘kenotic clause,’ such that ‘a being
would be said to be, for example, omniscient just in case that being
naturally (in the sense of “ordinarily” or “typically”) knows all true
propositions and fails to know any true proposition only when freely
and temporarily restricting its knowledge…[where] the kenotic
clause is the clause following the italicized connective’ (ibid. 97-8).
This of course allows God to be omniscient even while incarnate,
but, as Morris (ibid. 99) points out, this ‘seems to amount to no
more than a verbal maneuver which departs dramatically from the
traditional concept of omniscience’. A third option retains the
traditional de nition of each divine attribute, but claims that what
God has (necessarily) is not the attribute but the attribute quali ed
by ‘kenotic limitation possibilities’ (ibid.). For example, God has ‘the
property of being omniscient-unless-freely-and-temporarily-
choosing-to-be-otherwise’ (ibid.). This strategy has proved popular
amongst kenoticists,6 though Morris (ibid. 100-1) rejects it on the
grounds that it still involves a considerable departure from
traditional views of the divine nature.

Morris rejects the rst strategy on the grounds that it is

inconsistent with even a weak view of divine immutability (one
endorsed by Morris) according to which ‘any individual who has a
constitutive attribute of deity can never begin to have it, and can
never cease to have it’ (ibid. 97). Of course, it can be objected to
this that Morris has, in o ering this weak account of immutability,
already signi cantly departed from the God of classical theism,
since the account is perfectly consistent with the view that
being/becoming incarnate involves change. This is in accordance
with the position outlined above, according to which we learn from
the doctrine of the Incarnation that no divine person is timeless,
immutable, or impassible. This move does not entail the kind of
kenoticism that Morris rejects, of course, although it does push
away from classical theism in the way suggested above. Swinburne
(1994: 230-1) o ers a further consideration against the kenotic
strategy, namely, that ‘the hypothesis of the existence of a being
who has the divine properties essentially…is much simpler than the
hypothesis of the existence of a kenotic God’. Our evaluation of this
will depend on our judgment as to the success of those models of
the Incarnation that attempt to make it compatible with God’s
possession of all traditional divine attributes essentially. It seems to
me that omnipotence can be dealt with easily: Christ is necessarily
omnipotent, though there is only a limited amount that he can bring
about by means of standard human powers. In principle, I can see
no objection to Christ’s being necessarily omniscient as well, since it
is by no means apparent to me that being human requires being
limited in knowledge. But on the face of it this is precluded by
reasonably secure scriptural data about Christ’s life, according to
which he was ignorant (and possibly even mistaken), and subject to
temptation. Considerations such as these lead Morris, Swinburne,
and others to introduce a further element to their theories, and I
turn to this in the next section. As we shall see, there are various
replies to this worry in both Morris and Swinburne.

5. TWO-MINDS SOLUTIONS

Two-minds solutions build on the kind of material dealt with in the

previous section—particularly, the insight that certain divine
properties are compatible with failing to be merely human—but
make particular modi cations to deal with Jesus’s apparently
limited knowledge. The basic suggestion is that the divine person
includes two ‘minds’, or two ‘systems of belief’, or two
‘consciousnesses’. The position has been defended at length by
Thomas Morris. On the position, we can identify in Christ
‘something like two distinct ranges of consciousness. There is rst
what we can call the eternal mind of the Son with its distinctively
divine consciousness…. And in addition there is a distinctly earthly
consciousness that…drew its visual imagery from what the eyes of
Jesus saw, and its concepts from the languages that he learned’
(Morris 1986: 103-4). Morris uses the example of those cases of
multiple personality, discussed in psychoanalytic research, in which
one personality has full cognitive access to the experiences of the
others, to argue that there could be an ‘asymmetric accessing
relationship’ between the divine and human minds in Christ (ibid.
106), such that the divine mind has access to all the thoughts and
experiences contained in the human mind, though not conversely.
Clearly, as Morris acknowledges, the divine mind needs more than
mere access to the human mind, since each divine person has such
an asymmetric accessing relationship with every created mind. So
Morris (ibid. 162) argues that there is in addition a ‘unity of
cognitive and causal powers productive of the contents of each
mind’. Such a view gives unity of person, presumably because it
gives unity of mental subject: persons are mental subjects; minds on
this view are merely collections of mental items such that the
mental items have some kind of unity relation to each other alone,7
such that the relevant unity relationship does not require unity of
mental subject, but merely of the mental items within one
consciousness.
 On this view, having two minds does not entail being two
persons: ‘Ordinarily, minds and persons are individuated in a one-
one correspondence…. So among mere humans, the individuation of
two minds at any one time will su ce for the identi cation of two
persons. But this leaves open the possibility that outside that
context, there is no such one-one correlation’ (ibid. 157). Given
some of the insights of psychoanalysis in relation to divided minds,
there does not seem to be an overwhelming problem with this
distinction between minds and persons (I consider some of these
claims in a little more detail below). But these insights do not help
deal with the problem of the divine person’s including incompatible
beliefs, since they do not account for a person’s consciously holding
incompatible beliefs. So Morris (ibid. 159) maintains that not all
beliefs included in the human mind are contained in the divine
mind: ‘Most theologians who take seriously the real humanity of
Jesus…will want to allow at least the possibility that the full belief-
set of the earthly mind of Jesus…did not even constitute a proper
subset of the belief-set ingredients in the omniscient mind.’ But
there is nevertheless a Christological problem here: the
Christological point is supposed to be that the divine person has all
these beliefs as his own, so maintaining that not all the beliefs
contained in the human mind are included in the divine mind does
not prevent those beliefs from belonging to the person. As Tim Bayne
(2001:137) succinctly puts it, ‘Consciousnesses don’t believe things,
people do.’8 Conversely, if it is held that the relevant beliefs do not
belong to the divine person, it will be hard to resist the conclusion
that there are two mental subjects in Christ: the divine person, and a
subject for the false human beliefs too.

Suppose that we accept this last conclusion, namely that there

are two mental subjects in Christ—something that Bayne’s analysis
of Morris’s position seems to require. In this case, it looks as though
we need to make a further distinction in order to avoid
Nestorianism: namely, a distinction between mental subjects and
persons. How might we make this distinction? The anti-Nestorian
desideratum is that human attributes and experiences be ascribed to
the divine person: that they are his beliefs and his experiences. But
it is not clear that this requires that all human attributes and
experiences be attributed to the divine person. Perhaps most are;
but any that raise otherwise insoluble contradiction-problems could
be mental experiences of some other mental subject without being
experiences of the divine person, much as Morris implicitly suggests.

The Incarnation would on this view consist of two overlapping

mental subjects, such that some of the attributes of one would be
attributes of the other—something I attempted to defend a few years
ago (in Cross 2002: 318-23).9 We might in this context think of
some of the experiences of one mind—the human— owing to the
other—the divine:10 i.e. being such that they are the experiences of
the divine mental subject himself. And we might think that unity of
person requires maximal ow: that is to say, the experiences of the
human mental subject ow to the divine person unless there are
logical reasons why they cannot (e.g. because the beliefs of the
human mental subject are incompatible with beliefs of the divine
mental subject). So maximality of unavoidable ow (within the
logical constraint just mentioned), and not mere voluntary overlap
of minds, would be required in order to avoid the danger of two
distinct persons. On this view, any two mental subjects, united in as
strong a sense as is metaphysically possible, count as one person—
or, better, only the recipient subject counts as a person; the other is
a mental subject that fails to be a person, because most of its
experiences automatically maximally ow to some other entity: they
are not uniquely its experiences.11 This distinction between mental
subjects and persons need not present insurmountable problems:
after all, some such distinction seems to be implicitly at work in the
passages in Morris, examined above, where the problem of
contradictory beliefs is considered. It ts neatly too with cases of
divided minds where it seems plausible to maintain that there is one
person but two mental subjects (for example, in cases in which
neither mind appears to have any access to the experiences of the
other). This account of the Incarnation perhaps presents problems
for that explicitly o ered by Morris, since he identi es persons as
subjects of mental properties. But this would merely constitute a
local objection to the accounts o ered by Morris (in terms of
internal coherence), rather than a global objection to two-minds
Christologies as such. All that would be required of Morris would be
a distinction between persons and mental subjects.
 In any case, there is another well-known defence of a two-minds
Christology that avoids making any of these moves, because the
account defends the two-minds Christology while avoiding making a
distinction, whether implicit or explicit, between persons and
mental subjects. The view is defended by Swinburne. Before I
consider the way in which his account avoids any sort of distinction
between persons and mental subjects, we need (in order to
understand his position) to grasp certain metaphysical commitments
that he espouses. According to Swinburne (1994: 25-6), a human
being is a humanly embodied soul and a humanly animated soul.
Soul and body are two substances, and each is an instantiation of
distinct bundles of universal properties. The bodily properties are
instantiated in matter (ibid. 11 42-3, 46), and the soul’s properties
are instantiated in something that is individuated by a ‘thisness’ or
haecceity’ (ibid. 39-42). Personal identity derives from the identity
of the soul, in the sense that the soul is the ‘thisness’ of the human
person. In the case of the Incarnation, the Son of God takes on a
human body, and in gaining a relationship to a body (of the kind
usually had by a soul to its body), the Son of God begins to
instantiate human mental properties in addition to his divine mental
properties. The Son is the ‘thisness’ that determines the identity of
the embodied person. On this view, the Word becomes a composite
of body and soul, such that the soul, which previously constituted
the Son alone, becomes a part of the relevant human composite. The
idea, in short, is that the Word gains a contingent part—a body—
that he previously lacked.12
 Given this, what does Swinburne have to say about the Son’s
essential omniscience? Swinburne accepts a version of the two-
minds theory, and thus maintains that a solution can be found in the
shape of a person having two systems of belief, or somehow
including two consciousnesses. The example—taken from Freudian
psychotherapy—is of a person who, while clearly acting on the basis
of certain (true) beliefs, refuses to allow these beliefs into her
consciousness, and indeed consciously (and falsely) believes the
opposite of the relevant beliefs. In this case, the subject’s two belief-
systems are to some extent separate, in the sense that she does not
admit some of her beliefs to consciousness (ibid. 201). So the person
does not consciously simultaneously believe that p and not-p. I take
it that this is un-controversial: it does not imply that the same
person consciously has incompatible beliefs. Only one of the beliefs
is consciously entertained. For this reason Swinburne prefers to
categorize the other ‘belief’ not properly as a belief but as an
inclination to believe. The object of such an inclination is a
proposition that ‘does not form part of a general view of the world,
but merely guides the subject’s actions in certain circumstances’—as
in the case of a person who acts on certain inclinations to believe
without consciously having the contents of such inclinations as
beliefs of hers (ibid. 202).13 The key point, however, turns out to be
not that the inclinations are not consciously entertained. Rather,
they fail to count as beliefs because they are not expressive of the
subject’s overall world-view. If asked to give an account of her
beliefs about the world, the subject would not include in that
account any of those propositions that count merely as inclinations
to belief.

Swinburne spells out what this implies for the case of Christ. In
becoming incarnate, the divine person takes on a human belief-
system, one that comes to include a set of propositions that would
under normal circumstances form the beliefs of a human person. In
the case of Christ, this set of propositions forms merely inclinations
to belief, for some of the propositions (the false ones) do not
constitute the divine person’s world-view.14 Christ’s human actions
(including ‘thoughts consciously entertained connected with the
brain’) are performed ‘in the light of his human belief system’
(ibid.). Christ, then, is one person with two consciousnesses. But
while remaining essentially omniscient, and consciously so, the
divine person consciously keeps this knowledge from forming part
of the human belief-system. So Christ’s human actions are
performed on the basis of a belief-system that, rst, does not include
the belief that the belief-system is included in the divine belief-
system in this way; and, secondly, includes some inclinations to
false beliefs. But since inclinations to belief are not beliefs, there is
no danger of the divine person’s having false beliefs: as Swinburne
(ibid.) puts it, ‘his divine knowledge-system will inevitably include
the knowledge that his human system contains the beliefs [i.e.,
inclinations to belief] that it does; and it will include those among
the latter which are true’. And since the inclinations to belief in the
human belief-system are not beliefs, there is no worry about two
mental subjects: beliefs that p, on the one hand, and inclinations to
believe that not- p, on the other, can belong to one and the same
mental subject, namely, in this case, the divine person.

The way in which Swinburne sets out the metaphysics of his

theory has nevertheless given commentators some theological
unease. For as he himself acknowledges, at rst glance his view
might seem to be Apollinarian, since it denies to Christ a human
soul (see ibid. 196-7)—a view of Apollinarius’s that was condemned
at the rst Council of Constantinople (381 CE). Swinburne claims
that if Christ had a human soul in this sense, he would be two
individuals or persons—something that he rightly holds cannot have
been intended by early church councils when they a rmed that
Christ had a ‘rational soul’. Rather, according to Swinburne (ibid.
252), what is intended is that Christ has a human ‘way of thinking
and acting’; and it is certainly the case that this is what Apollinarius
wished to deny. And surely Swinburne is right about this. The Son
of God is not identical with the rational soul thus understood: the
Son is the composite of body and soul, and the identity of the person
derives from the Son’s thisness’. Since human souls generally
individuate human persons, there is no soul (in Swinburne’s sense of
‘soul’) other than the Son in the Incarnate Christ. But this does not
mean that there is no sense in which the Incarnate Christ has a
human soul in the way required by the councils. Swinburne allows
that there is a di erent but legitimate sense of ‘soul’—one that is
accepted by Aristotle—according to which a soul is ‘a way of
thinking and acting’—namely, a human way of thinking and acting.
And Swinburne’s view of course requires that the Word acquired
such a soul; he holds too that the early church councils require that
the Word acquired such a soul, and that the councils would deny
that the Word acquired a soul in Swinburne’s sense of ‘soul’. This
too is correct. Swinburne’s express claim is that the Son began to
instantiate human mental states, and this is something that
Apollinarius was not prepared to countenance. So on this reading,
the charge of Apollinarianism is unjust. In any case, it is important
to realize that Swinburne’s two-minds solution is in principle
independent of the metaphysics he outlines: we could wholly deny
his view about human identity and still nd his two-minds solution
fruitful and appealing.

Overall, Swinburne’s distinction between beliefs and inclinations

to belief provides, it seems to me, a more elegant way of dealing
with the Nestorian worry than the one I proposed above, and that I
found implicitly by Morris too, according to which we need to make
a distinction between persons and mental subjects. But if
Swinburne’s distinction between beliefs and inclinations to believe
is felt to be unsatisfactory, I believe that distinguishing mental
subjects from persons is still probably su cient to deal with the
Nestorian problem.

Modern discussion in philosophical theology on the doctrine of the

Incarnation still represents to some extent work in progress. But it
has, I believe, shown where best to look for solutions, and what the
likely shape of such solutions will be: namely, rst, an abandonment
of a strong form of classical theism, and secondly the clari cation of
the relation between persons and minds, and perhaps between both
of these and mental subjects. On the rst of these issues, we need to
abandon divine impassibility, immutability, and timelessness. The
two-minds views that I consider above all seem to entail this kind of
modi cation to classical theism (at least with regard to
impassibility, immutability, and timelessness). And it is surely right
to worry about the kinds of concern that philosophers raise.
Coherence is, after all, a necessary condition for truth, and
theological ‘wisdom’, or appeals to the necessarily ‘mysterious’
nature of the doctrine, should not be an excuse for failing to worry
about possible fundamental di culties with this central locus of
Christian theology. I have tried to suggest here that the most
hopeful strategy for dealing with possible contradictions lies in a
combination of something like Morris’s merely human/fully human
distinction with a modi cation of classical theism, such that no
divine person is timeless, or immutable, or impassible. Excluding
contradictory pairs of attributes does not, of course, entail
abandoning Chalcedon by adopting a ‘one-nature’ Christology. After
all, the Christologies I am advocating do not involve denying that
Christ has all properties necessary and jointly su cient for being
human (i.e. a human nature), and all properties necessary and
jointly su cient for being divine (i.e. a divine nature). The point is
merely that these lists of properties cannot contain contradictory
pairs.
 On the second point—the clari cation of the relation between
persons, mental subjects, and minds, the proposals made above are
certainly not inconsistent with the claim made at Constantinople III
that Christ has two wills and two energies—two modes of acting, or
two kinds of causal power—divine and human. Talk of divine will
and divine power here is roughly a way of saying that Christ can do
things (e.g. miracles) that only God can do; and talk of human will
and human power is roughly a way of saying that Christ can do
things (e.g. talk, touch things) that human beings can do. And talk
of two wills does not entail any serious con ict—even I, for
example, can have short-term goals di erent from and incompatible
with my long-term goals. What I cannot do, of course, is
simultaneously act in a way that achieves two incompatible goals—
and neither can an incarnate divine person. Nestorianism is avoided
both by the maximal- ow proposal, and by Swinburne’s distinction
between beliefs and inclinations to belief. The capacity of the
maximal- ow proposal to avoid Nestorianism perhaps remains
controversial, since what counts as Nestorianism is not a settled
matter, either today or in the light of the objections raised in early-
church orthodoxy (e.g. by Cyril of Alexandria). Overall, the
achievements thus far of this theological ‘research project’ (so to
speak—governed as it is by shared aims, materials, methods, and
assumptions) seem to be orthodox; but it is not yet clear that further
work is not required to show this.

NOTES
 Thanks to the editors of the volume for comments, and
especially to Brian Leftow and Joseph Jedwab. Jedwab’s
insightful feedback greatly improved the whole of this chapter.
Leftow’s comments greatly helped me deal with issues in sect.
3. We await eagerly Jedwab’s important Oxford doctoral
dissertation on the whole issue. It should of course be added
that neither Leftow nor Jedwab would endorse the conclusions
I come to here.

1. This reading of Chalcedon is not universal amongst scholars.

Sarah Coakley (2006: 246 n. 1), for example, notes that ‘In the
Chalcedonian de nition itself it is not clear that the hypostasis
is to be straightforwardly identi ed with the pre-existent
Logos [i.e. the Word], since it (the hypostasis) is mentioned
only in the context of its being the point of “concurrence” of
the humanity and the divinity.… The explicit identi cation of
the hypostasis with the Logos is made only later, at the second
Council of Constantinople [in 553].’ But in Chalcedon, Christ
is the ‘point of “concurrence” and in the Council the Son of
God (referred to as the Word) is identi ed with Christ, as just
noted. Thus, the Word is the subject of whatever properties
Christ is the subject of—and this amounts to the claim that the
hypostasis is the Word.

2. The worry about the particularity of the Incarnation receives

an extensive treatment in Morris 1986: 163-86; the question of
the evidential basis for the Incarnation ibid. 187-204, and
 especially in Swinburne 1994: 216-38 and Swinburne 2003. I
ignore these further debates here, because it seems to me that
the most important philosophical work has focused on the
coherence issue, and because this question (unlike the others)
seems to be a speci cally philosophical problem, one that
cannot be tackled without certain intellectual specializations
of a kind typically had by philosophers and philosophically
minded theologians.

3. Brian Hebblethwaite endorses this relative identity approach

too, though he rather confusingly chides Richard Swinburne
for accepting what is in e ect a reduplicative analysis without
accepting relative identity (Hebblethwaite 2005: 65-6). I
discuss Swinburne’s position below, and suggest why
Hebblethwaite’s criticism is misguided.

4. I thank Joseph Jedwab for this way of spelling out the relevant
distinction between parts and properties. For a di erent
attempt that strikes me now as less clear and elegant, see Cross
forthcoming.

5. For other endorsements of kenotic views by modern

philosophers of religion, see e.g. Davis 1983: 118-31, 2006;
Evans 1996: 132-6, 2002, 2006a; Feenstra 1989, 2006; Forrest
2000; and Thompson and Plantinga 2006.

6. See e.g. its endorsement in Feenstra 2006: 152.

7. I thank Joseph Jedwab for this way of putting the matter.

 8. A two-minds view related to Morris’s can be found in Brown
1985. But Brown’s account seems to involve some unresolved
issues. For example, he associates minds with natures (ibid.
265), but also talks of e.g. the divine nature experiencing
things ‘through the medium of the human nature’ (ibid. 266)
—suggesting that minds for Brown are mental subjects
(subjects of mental experiences), and thus, on the face of it,
persons. But Brown’s contribution nevertheless represents an
important rst step on the way to the more fully developed
two-minds solutions found in Morris and Swinburne.

9. Mymotivation in this case was to provide some kind of defence

of medieval Christologies, which generally concern themselves
with metaphysical rather than psychological accounts of what
it is to be a person. Thus, for the medievals, there is no sense
in which persons should be identi ed with mental subjects.
This medieval view raises questions exactly analogous to the
ones I am now considering in relation to Morris’s view. And
this, at the least, demonstrates that there are opinions close to
Morris’s view, and generally considered to be orthodox, that
require something like the distinction that I am now
proposing. The modern debate brings with itself, quite rightly,
additional insights that require further discussion.

10. I borrow talk of ‘ ow’ from Brown 1985: 264-5.

11. It may be that this account of the Incarnation, and of maximal

mental ow between two subjects, will require some further
 nessing of the doctrine of the Trinity, since it may be that
there is such maximal ow between the divine persons’ mental
experiences too, and the criterion is not intended to show that
there is just one divine person. But that is a topic for a
di erent paper.

12. It is hard to keep together all Swinburne’s claims here. On the

one hand, as we saw above, he insists that natures are
universals. On the other hand, he holds that the Word becomes
a human soul that has an (individual) body. The body, as an
individual, cannot be a universal, or part of a universal.
Furthermore, Christ’s human soul is supposed to be, like any
human soul, an individual essence, a thisness: the individuator
of a human person. Like any human soul, it is also supposed to
be a mental subject. But an individual essence is a (non-
universal) property: that is to say, an abstract object, whereas
a mental subject is supposed to be a concrete object. (Again, I
thank Jedwab for this point.) But this constitutes a worry
about his general philosophical anthropology rather than a
speci c objection to his Christological outworking of this
anthropology.

13. It is perhaps for this reason that Hebblethwaite believes that

Swinburne is committed to a use of qua-quali ers that requires
him to be committed to relative identity. But this is not right:
Swinburne claims that we can partition beliefs into di erent
consciousnesses, and that we can do so without endorsing
 either relative identity or ascribing con icting conscious
beliefs to one and the same person.

14. I am not sure whether Swinburne holds that all Christ’s human
beliefs are in fact inclinations to belief, or whether only the
false ones are. Swinburne introduces the distinction to deal
merely with the problem of contradictory beliefs, but ends up
arguing without quali cation that ‘the “beliefs” belonging to
the human perspective would be mere inclinations to belief’
(Swinburne 1994: 202-3). The true propositions contained in
the human belief-system would seem to be expressive of the
divine person’s world-view, and thus count as fully- edged
beliefs.

REFERENCES

ADAMS, MARILYN MCCORD (2006). Christ and Horrors: The Coherence of

Christology, Current Issues in Theology. Cambridge: Cambridge
University Press.

BAYNE, TIM (2001). ‘The Inclusion Model of the Incarnation:

Problems and Prospects’, Religious Studies 37:125-41.

BLOUNT, DOUGLAS K. (2002). ‘On the Incarnation of a Timeless God’,

in Ganssle and Woodru (eds.), 236-48.

BROWN, DAVID (1985). The Divine Trinity. London: Duckworth.

BURRELL, DAVID (2004). Review of Cross 2002, Speculum 79: 467-9.

COAKLEY, SARAH (2006). ‘Does Kenosis Rest on a Mistake?’, in Evans
(ed.), 2006b: 246-64.

CROSS, RICHARD (2002). The Metaphysics of the Incarnation: Thomas

Aquinas to Duns Scotus. Oxford: Oxford University Press.

——— (2003). ‘Incarnation, Omnipotence, and Action at a

Distance’, Neue Zeitschrift für systematische Theologie und
Religionsphilosophie 45: 293-312.

——— (forthcoming). ‘Parts and Properties in Christology’, in

Martin Stone (ed.), Reason, Faith and History: Essays for Paul
Helm. Aldershot: Ashgate.

DAVIS, STEPHEN T. (1983). Logic and the Nature of God. London:

Macmillan.

——— (2006). ‘Is Kenosis Orthodox?’, in Evans (ed.), 2006b: 112-

38.

——— KENDALL, DANIEL, and O’COLLINS, GERALD (eds.) (2002). The

Incarnation: An Interdisciplinary Symposium on the Incarnation of
the Son of God. Oxford: Oxford University Press.

EVANS, C. STEPHEN (1996). The Historical Christ and the Jesus of Faith:
The Incarnational Narrative as History. Oxford, Clarendon Press.

——— (2002). ‘The Self-Emptying of Love: Some Thoughts on

Kenotic Christology,’ in Davis, Kendall, and O Collins (eds.),
246-72.
——— (2006a). ‘Kenotic Christology and the Nature of God,’ in
Evans (ed.), 2006b, 190-217.

——— (ed.) (2006b). Exploring Kenotic Christology: The Self-Emptying

of God. Oxford: Oxford University Press.

FEENSTRA, RONALD J. (1989). ‘Reconsidering Kenotic Christology’, in

Ronald J. Feenstra and Cornelius Plantinga, Jr. (eds.), Trinity,
Incarnation and Atonement: Philosophical and Theological Essays,
Library of Religious Philosophy. Notre Dame: University of
Notre Dame Press, 128-52.

——— (2006). ‘A Kenotic Christological Model for Understanding

the Divine Attributes,’ in Evans (ed.), 2006b: 139-64.

FORREST, PETER (2000). ‘The Incarnation: A Philosophical Case for

Kenosis’, Religious Studies 36: 127-40.

GANSSLE, GREGORY E., and WOODRUFF, DAVID M. (eds.) (2002). God and
Time: Essays on the Divine Nature. New York: Oxford University
Press.

HEBBLETHWAITE, BRIAN (2005). Philosophical Theology and Christian

Doctrine, Exploring the Philosophy of Religion. Malden:
Blackwell.

HICK, JOHN (ed.) (1977). The Myth of God Incarnate. London: SCM.

LEFTOW, BRIAN (2002). ‘A Timeless God Incarnate’, in Davis, Kendall,

and O’Collins (eds.), 273-99.

MCCABE, HERBERT (1987). God Matters. London: Geo rey Chapman.

MERRICKS, TRENTON (2007). ‘The Word Made Flesh: Dualism,
Physicalism, and the Incarnation’, in Peter van Inwagen and
Dean Zimmerman (eds.), Persons Human and Divine. Oxford:
Oxford University Press, 281-300.

MORRIS, THOMAS V. (1986). The Logic of God Incarnate. Ithaca: Cornell

University Press.

SENOR, THOMAS (2002). ‘Incarnation, Timelessness, and Leibniz’s Law

Problems’, in Ganssle and Woodru (eds.), 220-35.

STUMP, ELEONORE (2003). Aquinas, The Arguments of the

Philosophers. London: Routledge.

SWINBURNE, RICHARD (1994). The Christian God. Oxford: Clarendon

Press.

——— (2003). The Resurrection of God Incarnate. Oxford: Oxford

University Press.

TANNER, KATHRYN (2001). Jesus, Humanity and the Trinity: A Brief

Systematic Theology, Scottish Journal of Theology Current Issues
in Theology. Edinburgh: T. & T. Clark.

TANNER, NORMAN P. (ed.) (1990). Decrees of the Ecumenical Councils, 2

vols. London: Sheed & Ward; Georgetown, Washington DC:
Georgetown University Press.

THOMPSON, THOMAS R., and PLANTINGA, CORNELIUS, Jr. (2006). ‘Trinity

and Kenosis’ in Evans (ed.), 2006b: 165-89.
VAN INWAGEN, Peter (1994). ‘Not by Confusion of Substance but by
Unity of Person’, in Alan G. Padgett (ed.), Reason and the
Christian Religion: Essays in Honour of Richard Swinburne. Oxford:
Clarendon.

WILLIAMS, C.J.F. (1967). ‘A Programme for Christology’, Religious

Studies 3: 513-24.
 CHAPTER 21

THE RESURRECTION OF THE BODY

TRENTON MERRICKS

RABBINIC Judaism, Christianity, and Islam agree that there is life

after death. Moreover, all three religions agree that we shall not
spend eternity as mere spirits or as disembodied souls. Instead, we
shall have hands and feet and size and shape. For we shall have
bodies. And not just any bodies. Each of us will have the very same
body that he or she had in this life, although that body will be
‘glori ed’. Each of us can have the same body because, at some
point in the future, all those bodies that have died will rise again to
new life. That is, dead bodies will be resurrected. Indeed, we
ourselves shall be resurrected. This is the doctrine of the
resurrection of the body.

This chapter will focus on two questions about the doctrine of

the resurrection, questions that will occur to most philosophers and
theologians interested in identity in general, and in personal identity
in particular. The rst question is: How? How could a body that at
the end of this life was (e.g.) frail and feeble be the very same body
as a resurrection body, a body that will not be frail or feeble, but
will instead be glori ed? Moreover, how could a body that has
passed out of existence—perhaps as a result of decay or cremation—
come back into existence on the Day of Resurrection?

The second question is: Why? Why would anyone want a

resurrection of the body? And even if the resurrection delivers
something that we want—maybe one’s current body has some
sentimental value and so having it back would be nice—we might
still wonder why any religion would give the doctrine a central
place, as Judaism, Christianity, and Islam all do.

CHANGE AND TEMPORAL GAPS

I have a body. And it is the very same body that I had earlier today.
It is even the same body I had as a child. That is why, for example,
my foot now bears a scar from an injury I su ered as a toddler.
Again, the body I have now is numerically identical with—is one
and the same object as—the body I had as a child. To deny this
would imply, I think quite implausibly, that I literally lost one body
and then acquired another at some point or other between my
childhood and now.

Of course, my body is now quite di erent from how it was when

I was a child. It has changed in size and in shape and in many other
ways. That is why someone who had last seen my body when I was
a child would not recognize it today. Again, my body has persisted
through enormous amounts of change. And so has every other
human body, at least every other body that has been around for any
appreciable amount of time. So too have many other objects.

We are familiar with a single object’s being one way, even

though it was or will be a very di erent way. So the doctrine of the
resurrection’s implication that one and the same body is one way at
death (e.g. sickly) but will be a very di erent way at resurrection
(e.g. glori ed and healthy) should not strike us as strange. So let us
assume that there is nothing worrisome about this implication.1

Dead bodies often pass out of existence. For example, some dead
bodies decay completely; others are cremated; some are even eaten
by wild animals. But, given the doctrine of the resurrection, even
those bodies that have gone out of existence will one day rise—and
so will one day exist—again. Thus the doctrine of the resurrection
implies a ‘temporal gap’ in the career of many bodies.2

To understand better what a temporal gap is supposed to be,

consider this story: you build a time machine that can send you—
and your body—to ‘the future’. You push the start button. You
disappear. You then reappear at some later date. That is, this
machine sends you to some future time, allowing you to ‘skip’ all
the times between now and then. Thus this machine causes a
temporal gap in your life, and also in the career of your body. Or
consider a watch that is disassembled, perhaps for cleaning. Assume
that, as a result, it ceases to exist. Assume further that when its parts
are reassembled, the watch comes back into existence. If all these
assumptions are right, the watch ‘jumps’ a temporal gap via
disassembly and reassembly.

The doctrine of the resurrection implies a temporal gap in the

career of many bodies. So objections to a temporal gap in a body’s
career are thereby objections to the doctrine of the resurrection. We
shall consider two such objections.

Suppose, rst, that, necessarily, whenever identity holds between

an object at one time and an object at another time, something must
account for or ground that identity. And suppose, second, that
nothing could account for, or ground, identity’s holding across a
temporal gap between a body that has (e.g.) been cremated and a
resurrection body. These two suppositions jointly rule out the
coming back into existence on Resurrection Day of a body that has
been cremated. So these two suppositions amount to an objection to
the doctrine of the resurrection.

A compelling account of what grounds the identity of a

resurrection body with that of a cremated body would undermine
this objection. And a number of accounts have been o ered. Over
the centuries, the most common account among Christian
philosophers and theologians, an account that was also
countenanced by Islamic and Jewish thinkers, has been this: on the
Last Day, God will gather up the very small bits that composed a
body at death and will ‘reassemble’ them, which will thereby bring
the body that died back into existence (see Bynum 1995; Smith and
Haddad 2002).3
 One potential problem with this account is that even a body’s
smallest parts might themselves sometimes go out of existence. For
example, perhaps some of those parts get converted (as in nuclear
explosions) from matter to energy. If this were to happen, then
those parts would not be around for reassembly on the Last Day,
and so that body—given this account—could not be resurrected.4
But every body is supposed to be resurrected.

Another objection to resurrection as reassembly trades on the

many ways in which the small parts of one body can end up as parts
of another body. The most sensational versions of this objection
involve cannibalism. A cannibal eats me, incorporates parts of my
body into his, and then dies. So some of the small bits that
composed my body at my death also composed the cannibal’s body
at his death. As a result, when Resurrection Day arrives, God cannot
(totally) reassemble both the cannibal’s body and mine. So—given
resurrection as reassembly—God cannot resurrect both my body and
the cannibal’s body. But, again, every body is supposed to be
resurrected.

In the second century, Athenagoras replied to this objection by

insisting that human esh is not digestible. As a result, he would
have maintained, the very small bits that compose my body at death
never could become parts of a cannibal’s body. So on Resurrection
Day my body shall be the only one with a claim to those bits, even if
a cannibal ate me and then died (Bynum 1995: 33). Unfortunately
for Athenagoras’s bold reply, human esh is (so I understand)
digestible.

A nal, and I think deeper, problem with resurrection as

reassembly was pointed out at least as far back as Origen (ibid. 64–
6). It is that our bodies are constantly changing with respect to the
very small bits that compose them. (This is one way in which our
bodies are quite di erent from inanimate objects such as watches.)
Therefore, a body’s identity from one moment to the next is not a
matter of having exactly the same very small parts. So it seems to be
a mistake to claim that its identity from death to resurrection is a
matter of having exactly the same very small parts.

Look at it this way. Because my body is constantly changing its

very small parts, it might be that none of the atoms that was a part
of my body when I was 5 years old is now a part of my body. Let us
suppose that none is. Now suppose that today God gathers together
all those atoms that composed my body when I was 5 years old and
reassembles them. This would produce a body, a body that looks
just like mine did when I was 5 years old. But that body would not
—could not—be my body. For that body is over there, where God
did the reassembling, but my body is right here (cf. van Inwagen
1978:120).

We know for sure that reassembling the atoms that composed my

body when I was 5 years old, if done today, would not produce my
body. This suggests that reassembling those same atoms, if done on
the Day of Resurrection, would not produce my body. And this
suggests, in turn, that reassembling the atoms that will compose my
body when I die, if done on the Day of Resurrection, would not
produce my body. At any rate, I conclude that a Resurrection Day
reassembly (of last parts, or indeed of parts from any time during a
body’s life) would not ground or account for the identity of a
resurrection body with a body that existed in this life.

Here is an alternative to resurrection as reassembly, an

alternative rst endorsed by early rabbis. Each body has an
indestructible bone at the base of its spinal cord, and even if a body
goes out of existence at some point after death, that body will come
back into existence when a resurrection body is constructed around
that bone (Bynum 1995: 54; cf. Smith and Haddad 2002:131).
Unfortunately for this proposal, there is no indestructible bone at
the base of the spinal cord, no bone that survives cremation, decay,
and every other threat to a dead body’s existence.5

Some might suggest that my current body will be identical with

whatever resurrection body has the same (substantial) soul as is had
by my current body. But a soul is not part of a body. And I doubt
that the identity of one physical object (such as a body) might be
entirely a matter of the identity of a second object (such as a soul)
when that second object is not itself a part of the rst object. In this
regard, taking a soul to be the guarantor of bodily identity is less
plausible than taking the bone from the base of the spinal cord to be
that guarantor. For at least that bone is a part of the relevant body.
 Moreover, explaining bodily identity across a temporal gap in
terms of the same soul seems to presuppose that having the same
soul is su cient for being the same body. This straightaway entails
that it is impossible for a soul to switch bodies. But, surely, if there
really are souls, it is possible for a soul to switch bodies. (Some even
argue for the existence of the soul by way of the alleged possibility
of a person’s—and so, presumably, a soul’s—switching bodies (see
e.g. Swinburne 1986:151).) And, nally, as will become clear in the
next section, the doctrine of the resurrection itself provides a reason
to deny that bodily identity across a temporal gap is secured by
having the same soul.

We have brie y considered three accounts of what might ground

the identity of the resurrection body with the body had in this life. I
do not think that any of these three accounts—or any other extant
account—is plausible.6 Nor do I have a novel account of my own to
o er. But, as I shall now point out, none of this implies that the
above objection to the doctrine of the resurrection is successful.

Recall that that objection assumes, rst, necessarily, something

must account for or ground every instance of identity over time;
and, second, nothing could account for or ground identity across (at
least some) resurrection-induced temporal gaps. As I have argued
elsewhere, possibly, some instances of identity over time have no
ground. Moreover, it would be no surprise if the identity of a body
had in this life with a resurrection body were just such an instance
(see Merricks 2001b). If so, then the rst assumption of the above
objection to the resurrection is false, and so that objection fails.

Moreover, suppose we concede, just for the sake of argument,

that something or other must account for, or ground, every instance
of identity over time, including every instance that would result
from a coming resurrection. Then believers in the resurrection can
block the above objection by denying that objection’s second
assumption. That is, they can simply conclude that there will be
something in virtue of which each resurrection body will be
identical with a body had in this life, something that will ground or
account for that identity. Crucially, they can conclude this even
given their inability to discover that ground, an inability evidenced
by the failure of proposed accounts such as reassembly. After all, no
one should presume to know exactly how God pulls o any miracle,
including the resurrection of the body.

A second objection to the temporal gaps implied by the doctrine

of the resurrection, and so a second objection to the doctrine itself,
claims that a certain condition is necessary for bodily identity over
time. Moreover, this objection adds, when it comes to the purported
identity of a body that has (e.g.) been cremated with a body that
will exist on Resurrection Day, this condition cannot be satis ed.7

There are various species of this second objection, each di ering

from the others with respect to what is alleged to be necessary for
bodily identity over time. One familiar allegation is that
spatiotemporal continuity is thus necessary. This allegation is
equivalent to the claim that, rst, spatial continuity is thus
necessary and, second, temporal continuity is thus necessary.8 The
claim that temporal continuity is thus necessary just is the claim
that one and the same body cannot exist at two times without
existing at all the times in between. And that claim just is the claim
that temporal gaps in a body’s career are impossible.

So the thesis that spatiotemporal continuity is necessary for

bodily identity over time says exactly that ‘spatial continuity’ is
necessary for bodily identity over time and, moreover, that temporal
gaps in a body’s career are impossible. Given that this is what it
says, the thesis that spatiotemporal continuity is necessary for
bodily identity over time presupposes that temporal gaps in a body’s
career are impossible. So that thesis is a question-begging reason to
conclude that such temporal gaps are impossible. So it is not a good
reason for that conclusion.

Some have suggested that, necessarily, the way a body is at one

time must appropriately cause the way a body is at a later time, if
the body at the later time is to be identical with the body at the
earlier time.9 This suggestion threatens temporal gaps in a body’s
career, and so threatens the doctrine of the resurrection, only if the
relevant sort of causation cannot occur across a temporal gap.

Suppose, once more, that the time machine sends you to the
future. You arrive in the future with a familiar tattoo on your leg.
That tattoo’s being on your leg was caused not only by a youthful
lapse of judgment, but also—and more importantly—by your having
that very tattoo on your leg before entering the time machine. This
implies that causation can occur across a temporal gap, since the
way your leg was before time travel causes it to be a certain way
after time travel. In fact, this seems to be just the sort of causation
that is allegedly necessary for bodily identity over time.

Now some might object that the time machine story just told is
absolutely impossible. But I do not see why we should agree with
them.10 More generally, and more to the point, I see no compelling
reason to conclude that there is any condition that is both necessary
for bodily identity over time and also cannot possibly be satis ed
across a temporal gap. On the other hand, the considerations raised
above do not show that there is no such condition. As far as those
considerations go, we should be agnostic about the existence of such
a condition.

Similarly, I think that, although we might have hunches one way

or the other, philosophical reasons of the sort surveyed above
deliver nothing conclusive about the possibility of the temporal gaps
implied by resurrection. As far as standard philosophical re ection
on these matters goes, resurrection of the cremated or decayed
human body might be possible, but then again it might not be
possible.

Of course, those who believe in the resurrection in the rst place

do not believe in it because of standard philosophical re ection.
Rather, we believe that God has revealed that there will be a
resurrection of the body, a resurrection that—given cremation,
decay, etc.—implies bodily identity across a temporal gap. And to
the extent that revelation justi es belief in the resurrection, I think
it also justi es belief in bodily identity across a temporal gap. So it
likewise justi es the conclusion that there are no necessary
conditions for bodily identity that cannot possibly be satis ed across
a temporal gap.

BODIES AND PERSONS

Even in polite company, one may admit to believing in life after

death, at least if one’s beliefs are appropriately spiritual, involving
leaving one’s body behind, heading toward the light, and so on.
Such beliefs are controversial, of course, but most people will
maintain eye contact with you, and perhaps even murmur
sympathetically, while you express your hope that, for example,
your spirit will live on past the grave. Alas, the resurrection is
another matter altogether. At least, my own experience suggests that
averted eyes and an awkward silence are the typical results of
expressing one’s hope that, at some time after one’s death, every
dead body, including one’s own, will come back to life.

The idea of a coming resurrection of every body seems strange

not just to the non-religious, but also to many of the religious,
including many who believe that Jesus Christ rose from the dead.
Perhaps discomfort with the idea of many bodies rising from the
dead is why—I merely speculate—my childhood Sunday school
teachers never asked me to memorize this passage from the New
Testament, which concerns events around the time of Christ’s death
and resurrection: ‘The tombs also were opened and many bodies of
the saints who had fallen asleep were raised. After [Christ’s]
resurrection they came out of the tombs and entered the holy city
and appeared to many’ (Matt. 27: 52–3, NRSV). At any rate, it is
safe to assume that the signs at sporting events that read ‘John 3:
16’ will not soon be replaced with signs saying ‘Matthew 27: 52–3’.
Lots of bodies rising from the dead is just too strange.

Nevertheless, Christianity, Islam, and traditional rabbinic

Judaism all teach that lots of bodies—in fact, every human body
that has ever died—will rise from the dead on the Last Day.
Moreover, this teaching is central to all three religions. For example
—and now I narrow my focus to Christianity—the Apostle’s Creed
closes with an a rmation of ‘the resurrection of the body, and the
life everlasting’.11 And here are the nal words of the Nicene Creed:
‘I look for the resurrection of the dead, and the life of the world to
come. Amen.’12

These creeds are short documents, meant to summarize the most

important points of Christian theology. And resurrection of the body
gets explicit mention in both, which I take to be evidence of the
centrality of the doctrine of the resurrection to Christianity. This
centrality should bring to mind some of the questions that opened
this chapter: Why should anyone want there to be a resurrection of
the body? And why would any religion have that doctrine at its
heart?
 I shall defend an answer to these questions. My defense begins
with a three-step argument, each step of which is plausible, though
controversial. The rst step says that each of us—each human
person—has physical properties. For example, you literally have a
certain weight. You likewise have a certain shape. You have a
certain location in space. And so on.13

The second step says that you are neither heavier than nor
lighter than your body; that is, your weight is the same as your
body’s weight. Moreover, you are not one shape, and your body
another; rather you and your body have the same shape. Nor are
you o in one corner of the room, while your body is to be found in
another; instead, you are located just where your body is located.

The third step says that there is only one human-shaped object
exactly and entirely located where you are exactly and entirely
located, and, more generally, only one object with all of the physical
properties had by you and had by your body. Once we have taken
these three steps, we must conclude that you are identical with your
body.

So if the above argument is sound, you are identical with your

body. (And even if it is not sound, it is still illuminating, clarifying
what your being identical with your body amounts to.) At any rate,
let us suppose, just for the sake of argument, that you really are
identical with your body. More generally, let us suppose, for the
sake of argument, that each of us is identical with his or her body.
 As was noted in the preceding section, most dead bodies
(eventually) cease to exist. This implies, given our supposition that
each of us is identical with his or her body, that most of us shall
cease to exist after death. And this implies that, for most of us, our
only hope for existence after death is the hope that our bodies (i.e.
we ourselves) will come back into existence. Moreover, this implies
that, for all of us, our only hope for life after death is the hope that
our bodies (i.e. we ourselves) will live again, that is, that our bodies
will be resurrected. The doctrine of the resurrection of the body—
despite its appearing quite strange at rst glance—is as motivated as
belief in life after death itself.

Look at it this way. Suppose you know that you are about to die.
And you hope that death is not the end. Add that you know that you
are one and the same thing as your body. Then your hope for life
after death should have a very clear focus. You should hope that
your body (i.e. you yourself) will one day live again. Your body’s
living again will not happen on its own, of course. It will take a
miracle, especially if your body passes out of existence by way of
(e.g.) cremation. But that miracle—and, more generally, God’s
raising every dead body—will not be merely some sort of spooky
sideshow. Instead, it will be your only shot at life after death.

Or suppose that a close friend has just died. Suppose that you
know for certain that your friend was identical with his or her body.
Then you have only one hope for seeing your friend again: the
resurrection. Your only hope is the hope that, someday, God will
raise the dead. Thus we see once more that our being identical with
our bodies makes the motivation for, and importance of, the
doctrine of the resurrection perfectly clear.

If we are bodies, then when our bodies are resurrected, we

ourselves are resurrected. This result ts with the way the creeds
couple belief in resurrection with belief in everlasting life, and also
with the way various passages of Scripture describe the resurrection.
Consider:

At that time Michael, the great prince, the protector of your people,
shall arise. There shall be a time of anguish, such as has never
occurred since nations rst came into existence. But at that time
your people shall be delivered, everyone whose name is found
written in the book. Many of those who sleep in the dust of the
earth shall awake, some to everlasting life, and some to shame and
everlasting contempt. (Dan. 12:1–2, NRSV)

Do not be astonished at this; for the hour is coming when all who
are in their graves will hear [the Son of Man’s] voice and will come
out—those who have done good, to the resurrection of life, and
those who have done evil, to the resurrection of condemnation.

(John 5: 28–9, NRSV)

For the Lord himself, with a cry of command, with the archangel’s
call and with the sound of God’s trumpet, will descend from heaven,
and the dead in Christ will rise…

(1 Thess. 4: 16, NRSV)

If we take the above passages at face value—or take any of a
number of others at face value (e.g. Matt. 14: 12–14; Acts 24; 1 Cor.
15)—it is not just dead bodies that will be raised to life, but dead
people. Our being identical with our bodies makes perfect sense of
the idea that the resurrection of our bodies will be the resurrection
of us.14

What if we were not identical with our bodies? Then it would be

hard, if not impossible, to make sense of the idea that dead people
will be resurrected. Moreover, the importance of the doctrine that,
on the Day of Resurrection, one gets a body identical with the body
one had in this life would be di cult to explain. Indeed, I cannot
think of any plausible explanation at all, much less one that rivals
the very straightforward and absolutely compelling explanation that
ows directly from the claim that each of us is identical with his or
her body.

I think that all of this gives those of us who believe in the

resurrection of the body—and who are committed to its importance
—a good reason to conclude that we are identical with our
respective bodies. Thus we have a new reason to conclude that each
of us is identical with his or her body, a reason in addition to my
three-step argument above that began with the claim that we have
physical properties.15
 This is but one reason to conclude that each of us is identical
with his or her body. I believe that there are further reasons to
endorse this conclusion. And there are, of course, alleged reasons to
reject this conclusion, including some speci cally theological
reasons. Below I shall respond to a few reasons one might o er for
rejecting this conclusion. But I do not pretend to respond to every
such reason, just as I do not pretend to present every reason one
might have for a rming that each of us is identical with his or her
body.16

Let us begin with what I suspect is the most common reason that
many Christians, and others, deny that a human person is one and
the same thing as his or her body. They deny this because they want
to make sense of life after death, life after the destruction of one’s
body. And they think that this can be done only if one is not the
body that will be destroyed, but instead something else, such as a
soul. But as should now be perfectly clear, this ‘reason’ is no good.
On the contrary, I argued above that we are identical with our
bodies precisely because this identity makes the best sense of
speci cally Christian claims surrounding life after death, even life
after the destruction of one’s body. (Our being identical with our
bodies is, I confess, entirely inconsistent with the pictures of life
after death found in, for example, pagan Greek philosophy and the
movie Ghost.)

Let us turn to another objection. On the view I am here

suggesting, we cease to exist at (some point after) death and then
come back into existence on the Day of Resurrection. Put otherwise,
my view implies that we jump ahead in time from our death to the
Day of Resurrection, skipping all the times in between. And some
might object that Christians are committed not just to life after
death, but to life after death and before resurrection.

Some might thus object because of certain scriptures. For

example, the book of Revelation speaks of souls, under the altar,
prior to the Day of Resurrection:

When he opened the fth seal, I saw under the altar the souls of
those who had been slaughtered for the word of God and for the
testimony they had given; they cried out with a loud voice,
‘Sovereign Lord, holy and true, how long will it be before you judge
and avenge our blood on the inhabitants of the earth?’ They were
each given a white robe and told to rest a little longer… (Rev. 6:
9–11, NRSV)

This passage, taken completely literally, suggests that martyrs exist

as souls between death and resurrection (and also that souls can
wear robes). But I think we should not take this passage completely
literally. And the same goes for other passages that seem to suggest
that we shall exist after our death but before our resurrection.17

I say this partly because I take di erent passages literally, such

as those that say that dead people are raised to life. And I say this
partly because I am convinced by the above explanation of the
motivation for, and centrality of, the doctrine of the resurrection. In
the background here is my opinion that the emphasis on
resurrection as our hope for eternal life is more central to Christian
Scripture and creed than is the idea of an ‘intermediate existence’
between death and resurrection.

Some might respond that intermediate existence is required for a

practice central to the devotional lives of many Christians: seeking
the intercession of the saints. Most of the saints have died, but have
not yet been resurrected. (The exceptions that prove the rule are,
assuming her assumption, the Blessed Virgin Mary, along with any
Saint that never died in the rst place, such as unfallen angles.) So
—if human beings do not exist between death and resurrection—
most of the saints do not now exist. This seems to threaten the
practice of asking the saints for help.

But consider the following. I ask Saint Frideswide to pray for me,
to ask God to grant a certain request. God, being omniscient, knows
that I have asked her this. So suppose that God, after the
resurrection, will communicate my request to her. She will then ask
God to have granted my petition. God even now knows that she will
do this. And so he now grants my petition, on account of
Frideswide’s future intercession. And so it goes, in general, with
how the saints intercede for us.

This seems to accommodate the practice, and e cacy, of asking

the saints for help. More generally, I think that it o ers one way of
securing the ‘communion of saints’. But none of this requires that
the saints exist right now. So I conclude that the various practices
involving the saints do not require that they exist between death
and resurrection.

Of course, the ‘mechanism’ for saintly intercession I have just

outlined is not what petitioners are likely to have in mind. But I am
not sure what mechanism, if any, they do have in mind. After all,
those who seek the help of the saints do so not only in many
languages, but also often completely silently. How are the saints to
know what is being asked of them? I think that the best answer is
that God, in his omniscience, knows what is requested of each saint,
and somehow communicates that to him or her.

So I suspect that, whatever we say about the existence of human

beings between death and resurrection, any petitions that reach the
saints do so by ‘going through God’ in some way or other. But once
we concede this, I see nothing objectionable about the mechanism I
have suggested. And that mechanism is consistent with each saint—
like each of us—jumping ahead in time from his or her death to the
Day of Resurrection.

NOTES

Thanks to Mike Bergmann, Tom Flint, Mark Murphy, Mike

Murray, Mike Rea, and Nick Wolterstor .

1. Merricks 1994 discusses, and defends, identity through change.

Some might object that, while identity can be preserved
 through ordinary change, one’s glori ed resurrection body will
be so di erent from one’s current body that that resurrection
body cannot be identical with the body one has now. But I
reply that we do not know enough about what a resurrection
body will be like to conclude this, especially when we recall
that some ordinary change is quite stunning, as when a single
body goes from being the body of an infant to that of a full-
grown adult.

2. In what follows, I shall say that the doctrine of the resurrection

implies a temporal gap in the career of many bodies. This is
shorthand for the claim that that doctrine and the fact that
many dead bodies go out of existence jointly imply such gaps.
My own view is that a human body ceases to exist
immediately after dying (see Merricks 2001a: 53). But nothing
I say below turns on this view. The arguments below require
only that at least some human bodies cease to exist at some
point or other after dying.

3. We shall consider objections to resurrection as reassembly. But,

at least until quite recently, virtually no self-styled believer in
the resurrection would have raised the following objection:
‘What is the point of reassembly? Why even try to account for
the identity of a resurrection body with a body had in this life?
After all, resurrection bodies need not be numerically identical
with bodies had in this life.’ Virtually no one would have thus
objected because debates among believers in the resurrection
 have been over how (not whether) a body that has ceased to
exist will secure identity with a resurrection body. Again, over
almost all the past two thousand years, those debates have
uniformly presupposed that the very body that dies (and
perishes) will rise again. This is why I say that, according to
the doctrine of the resurrection, the very body that dies will
rise again. Christians have another reason to insist that the
very body that dies will rise again: they believe that Christ’s
body at cruci xion, that is, the body that was cruci ed on the
cross, is one and the same as the body that walked out of the
tomb; and they believe that our resurrection will be patterned
after Christ’s.

4. Objection: Those parts could come back into existence on the

Day of Resurrection and so be available for reassembly. Reply:
To do that, the parts themselves would have to jump a
temporal gap. But they could not jump a temporal gap by way
of reassembly, since the idea here is that when they convert to
energy (or otherwise cease to exist), so too do all their parts (if
they have parts at all). So this objection requires that there are
ways other than reassembly to jump a temporal gap. Thus this
objection in defense of resurrection as reassembly undermines,
at least to some extent, the motivation for resurrection as
reassembly.

5. Bones perish when a body is cremated. But some smaller parts

of a body do not. For example, an electron that is now part of
 my body would not perish if my body were now cremated.
Thus one might suggest that there is some very small part—a
certain electron, say, as opposed to a very hard bone—of each
body such that that body would come back into existence, if a
resurrection body were constructed around that electron.

This suggestion is a icted by analogues of the problems

that a ict resurrection as reassembly. First, that electron itself
might go out of existence, precluding resurrection of the
relevant body. Second, that electron might become part of
another body, as a result of (e.g.) cannibalism; this
demonstrates, among other things, that having that electron as
a part is not su cient for being the body that originally had it.
Third, human bodies are constantly changing their very small
parts, including the electrons that compose them; so it seems
mistaken (if not positively bizarre) to say that bodily identity
across a temporal gap could be entirely a matter of having a
single special electron as a part.

6. Hud Hudson 2001: 190 o ers a memory-based account of a

physical person’s jumping the temporal gap between that
person’s death and resurrection (cf. John Locke 1975: 542;
Essay, 4. 3. 6). But this is not—as Hudson himself agrees—
even a purported account of a body’s jumping a temporal gap.
Peter van Inwagen 1978, unlike Hudson, does believe that the
very body that has died will be resurrected. And van Inwagen
o ers an account of how this could be. But his account rejects
 my assumption that dead bodies typically cease to exist.
Rather, according to his account, each dead body is squirreled
away somewhere by the Lord to await resurrection, while a
replica decays (or is cremated, etc.) in its stead.

7. This second objection turns on (alleged) necessary conditions

for identity over time. The rst objection—the objection in
terms of what grounds or accounts for transgap bodily identity
—turned on (alleged) su cient conditions for identity over
time. Unlike the rst objection, this second objection cannot
be blocked by denying that identity across temporal gaps must
be grounded. Nor can it be blocked, again unlike the rst
objection, by admitting ignorance about how God will
resurrect long-gone bodies.

8. More precise statements of spatiotemporal continuity are

o ered by, among others, George Mavrodes 1977: 37 and Eli
Hirsch 1982: 15–21.

9. Van Inwagen 1978 not only endorses this causal requirement,

but also turns it into a new objection to resurrection as
reassembly, an objection in addition to those considered
above. The key to van Inwagen’s objection is that, he argues,
the reassembly of a dead body’s last parts is not su cient for
the body that has died to cause, in the relevant way, the
features of the body that results from reassembly.

10. Dean Zimmerman defends the possibility of the relevant sort of

causation occurring across a temporal gap. Zimmerman 1999:
 204 summarizes his defense thus: ‘Of course the supposition
that causal processes can be spatiotemporally gappy in this
way is contentious. But it should be much less so than it once
was, for the following reasons: there is no a priori reason to
think it is impossible, and some a posteriori reason to think it
happens; the theories of causation which imply that it is
impossible have been exploded; and the most promising
theories still in the water can accommodate it.’

11. The doctrine of the resurrection is also central to Islam and

Judaism. Smith and Haddad 2002: 63 tell us: ‘The promise, the
guarantee, of the day at which all bodies will be resurrected
and all persons called to account for their deeds and the
measure of their faith is the dominant message of the
Qur’an…’ The Basic Principles of Moses Maimonides are
widely taken to articulate the central beliefs of rabbinic
Judaism. The thirteenth and nal principle a rms the
resurrection of the dead.

12. The Athanasian Creed tells us that, at Christ’s coming, ‘all men
shall rise again with their bodies’.

13. Our having physical properties does not imply that our only
properties are physical. For example, our having physical
properties is consistent with our having mental properties,
even if those mental properties are themselves in no way
physical. Thus our having physical properties is consistent
with ‘property dualism’ about the mental.
14. I am not saying that the human authors of Scripture, or those
who formulated the creeds, believed or meant to teach that we
are identical with our bodies. I am saying that they believed
and meant to teach that our bodies will be resurrected, that
this is intimately related to our hope for life after death, and
that dead people will rise again. Because I believe what they
taught, I conclude—for reasons given in this chapter—that we
are identical with our bodies.

Compare my approach here to Scripture and the creeds

with the following. A document written in 1350 describes
those dying all around the author; the dying experience
nausea, fever, and other symptoms of bubonic plague. Because
we believe what the author had to say, we might conclude—
for reasons that a pathologist might give—that those people
had an infection caused by the bacterium Yersinia prestis. Of
course, the author of that document did not believe or mean to
say anything about that bacterium; the author had never even
heard of bacteria.

15. I have just argued from the doctrine of the resurrection to the
claim that we are identical with our bodies. That claim implies
that we lack (substantial) souls. This is why I said, in the
preceding section, that the doctrine of the resurrection
provides a reason to deny that bodily identity across a
temporal gap is secured by having the same soul.
 Those who deny that we are identical with our bodies might
say that they have no idea why the doctrine of the resurrection
is important, even though it is important. Perhaps this is a
reasonable thing for them to say. And they might add that
saying this is analogous to saying that they have no idea what
grounds bodily identity across a temporal gap, even though
(some might maintain) something must ground it. But I do not
think that the cases are appropriately analogous. As I argued
in the previous section, there is nothing remotely like a
genuinely live option that, if true, would deliver a full and
satisfying account of what grounds the identity of a
resurrection body with a body that was (e.g.) cremated. So an
appeal to ignorance here is unavoidable for believers in the
resurrection, at least for those who think there must be some
ground for bodily identity across a temporal gap. On the other
hand, there is (what I take to be) a genuinely live option that,
if true, would deliver a full and satisfying account of the
importance of the doctrine of the resurrection: namely, the
identity of a person with his or her body.

16. For example, the Incarnation provides considerations that bear

on the identity of a person with his or her body, but I shall not
discuss them in this paper. Plantinga 1999 and Leftow 2002
both object that the Incarnation cannot be squared with the
claim that human persons are identical with their bodies. But,
on the contrary, I argue (Merricks 2007) that it is easier to
 reconcile the Incarnation with the claim that human persons
are identical with their bodies than with any other thesis
about the relation of a person to his or her body.

17. My dying results in my literal non-existence. Nevertheless, to

die is to jump ahead in time to the Day of Resurrection. Thus I
could think to my self, as I am about to die, that so far as
things seem to me—and only because of the resurrection of the
body—this day I shall be with the Lord in paradise (cf. Luke
23: 43).

REFERENCES

BYNUM, CAROLINE WALKER (1995). The Resurrection of the Body in

Western Christianity 200–1336. New York: Columbia University
Press.

HIRSCH, ELI (1982). The Concept of Identity. Oxford: Oxford University

Press.

HUDSON, HUD (2001). A Materialist Metaphysics of the Human Person.

Ithaca: Cornell University Press.

LEFTOW, BRIAN (2002). ‘A Timeless God Incarnate’, in Stephen T.

Davis, Daniel Kendall, and Gerald O’Collins (eds.), The
Incarnation. Oxford: Oxford University Press.

LOCKE, JOHN (1975). An Essay Concerning Human Understanding, ed.

Peter H. Nidditch. Oxford: Clarendon.
MAVRODES, GEORGE I. (1977). ‘The Life Everlasting and the Bodily
Criterion of Identity’, Noûs 11: 27–39.

MERRICKS, TRENTON (1994). ‘Endurance and Indiscernibility’, Journal

of Philosophy 91: 165–84.

——— (2001a). Objects and Persons. Oxford: Clarendon.

——— (2001b). ‘How to Live Forever without Saving Your Soul:

Physicalism and Immortality’, in Kevin Corcoran (ed.), Soul,
Body, and Survival: Essays on the Metaphysics of Human Persons.
Ithaca: Cornell University Press.

MERRICKS, TRENTON (2007). ‘The Word Made Flesh: Dualism,

Physicalism, and the Incarnation’, in Peter van Inwagen and
Dean Zimmerman (eds.), Persons: Human and Divine. Oxford:
Oxford University Press.

PLANTINGA, ALVIN (1999). ‘On Heresy, Mind, and Truth’, Faith and
Philosophy 16: 182–93.

SMITH, JANE IDLEMAN, and HADDAD, YVONNE YAZBECK (2002). The Islamic
Understanding ofDeath and Resurrection. Oxford: Oxford
University Press.

SWINBURNE, RICHARD (1986). The Evolution of the Soul. Oxford:

Clarendon.

VAN INWAGEN, PETER (1978). ‘The Possibility of Resurrection’,

International Journal for Philosophy of Religion 9: 114–21.
ZIMMERMAN, DEAN (1999). ‘The Compatibility of Materialism and
Survival: The “Falling Elevator” Model’, Faith and Philosophy 16:
194–212.
 CHAPTER 22

HEAVEN AND HELL

JERRY L. WALLS

IN the introduction to their anthology of readings about heaven,

Carol and Philip Zaleski observe that the classical view of
philosophy represented by the likes of Plato and his followers was
that our highest calling as human beings is the eternal
contemplation of truth, beauty, and goodness. So understood, the
Zaleskis observe that ‘philosophy itself … is nothing less than the
quest for heaven’.1

While versions of the doctrines of heaven and hell appeared in a

number of ancient cultures long before Christ was born,2 the
Christian account of heaven raised the signi cance of the quest for
truth, beauty and goodness to new heights. The ultimate destiny of
every person is either eternal joy of unimaginable glory and delight
or eternal misery of unspeakable horror. The distinctively Christian
account of God with its attendant doctrines of Trinity, incarnation,
and atonement, gave particular shape to the hope for heaven and
the horror of hell. Heaven is the climax and perfection of an
intimate relationship with a personal God whose very nature is love.
This love was revealed de nitively in the incarnation, atonement,
and resurrection of Jesus, and the salvation thereby provided will
reach its highest end in the blissful experience of seeing the Trinity.3
The choice either to receive this salvation or to reject it is a matter
of momentous importance. It is precisely the prospect of losing a
good so extraordinary that makes hell so terrible. The truth was
never so beautiful, and the stakes never so high in the quest to nd
the truth and follow after the good.

With the prospects for happiness and misery so magni ed, the
meaning of our lives and the signi cance of our choices are both
elevated to dramatic proportions. Indeed, both heaven and hell have
stirred the imagination of western culture for centuries, inspiring
great literature as well as visual art. Moreover, heaven and hell have
played an undeniable role in the moral foundations of western
culture, providing direction as well as hard motivation in the form
of ultimate sanctions.

Despite this impressive pedigree, heaven and hell have faded

from the consciousness of many people in recent times. Citing what
he sees as ‘irreversible changes’ in our ways of thinking and
perceiving the world, Harvard theologian Gordon Kaufman opines
that ‘I don’t think there can be any future for heaven and hell.’4 For
an example of the loss of heavenly belief in recent times that may be
of particular interest to philosophers, consider the autobiographical
musings of one of the greatest practitioners of the discipline in the
twentieth century, namely, W. V. Quine.
I may have been nine when I began to worry about the absurdity of
heaven and eternal life, and about the jeopardy I was incurring by
those evil doubts. Presently I realized that the jeopardy was illusory
if the doubts were right. My somber conclusion was nonetheless
disappointing, but I rested with it…. Such, then, was the beginning
of my philosophical concern. Perhaps the same is true of the
majority of philosophers.5

Unfortunately Quine does not tell us why he found the idea of

heaven and eternal life absurd, but seems to assume it is obvious or
at least that he can take it for granted among his fellow
philosophers. Nor does he tell us why he thinks the majority of
philosophers were awakened from their dogmatic slumber by
similarly skeptical doubts about the reality of heaven. Regardless, it
is hard to exaggerate the di erence between seeing the quest for
heaven as the very goal of philosophy, and seeing the absurdity of
heaven as the dawning moment of philosophical awakening.

If Quine’s speculation is true that the majority of philosophers

were initiated into philosophical thinking by the absurdity of
heaven, or even if they share his view on the matter, we should
hardly expect to nd much recent philosophical literature on the
subject. And judging by that criterion alone, Quine’s speculation has
a lot to be said for it, for contemporary philosophers have given
scant attention to heaven, to put it mildly. Indeed, more
surprisingly, heaven has even lost its imaginative grip on many
people who would still profess to believe in it. In their history of
heaven, Colleen McDannell and Bernhard Lang observe that
‘although fundamentalists would discard the suggestion that heaven
no longer is an active part of their belief system, eternal life has
become an unknown place or a state of vague identity’.6

When we turn our attention to hell, we discover that it has

su ered perhaps even more from benign neglect. In the centuries
since the Enlightenment, hell has melted down from a hard-cast
theological conviction into an emotionally charged caricature of
revivalism. For many, hell exists primarily as a parody, surviving in
popular culture through numerous ‘Far Side’ cartoons and an
occasional episode of ‘The Simpsons’. In 1985, Martin Marty
published an article tellingly entitled ‘Hell Disappeared. No One
Noticed. A Civic Argument’.7 Marty’s article takes aim at those who
want to return prayer to schools as a means of restoring moral
values, and argue that we can do so without endorsing any
particular religion. Marty points out, however, that the belief that
God punishes evil behavior, not only now but hereafter, has always
been an essential component of any morally relevant view of God.
Unfortunately the doctrine of hell is no longer ‘culturally available’,
so there is no substantive sense in which God can be restored to
public schools and civic life. As part of his evidence that hell is no
longer available to contemporary culture, Marty pointed out that a
bibliographical search of recent literature turned up almost nothing
on the subject.8
 In the past several years, however, hell has returned with a
vengeance—or without a vengeance, depending on how the doctrine
is understood! Indeed, just six years after Marty’s article appeared,
U.S. News and World Report ran an article entitled ‘Hell’s Sober
Comeback’, in which it claimed that ‘hell is undergoing something
of a revival in American religious thought’, even among
theologians.9 Were Marty to do a bibliographical search today he
would nd plenty of material on hell. Indeed, he would discover a
lively debate on the nature of hell, not only among theologians and
biblical scholars but among philosophers as well. Moreover, there
are signs of renewed interest in heaven as well, although
philosophical literature on hell is far more prevalent to date,
perhaps because hell presents more urgent di culties.

No doubt there are several factors involved in the revival of

interest in heaven and hell, but for our purposes, one factor deserves
highlighting. The last two or three decades of the twentieth century
witnessed a remarkable resurgence in the eld of philosophy of
religion and philosophical theology. In addition to rst-rate work in
traditional areas of concern such as the problem of evil and theistic
arguments, Christian philosophers have gone beyond generic theism
and engaged distinctively Christian beliefs and doctrines as well.
Much of this work has challenged the sort of minimalist and
agnostic theology represented by liberal theologians such as
Kaufman, and more fundamentally it has examined and criticized
the philosophical and intellectual foundations of such theology.
Many of the alleged ‘irreversible changes’ that Kaufman and his
fellow liberal theologians hold sacred are no longer viewed as such
by many of the best Christian philosophers. These intellectual
changes suggest the possibility that serious belief in God as well as
the larger framework of Christian belief could become more
pervasive in our society to the degree that even hell could again
become ‘culturally available’. Let us now turn our attention more
carefully to the doctrine of hell.

TRADITIONAL VERSUS CONTEMPORARY DEFENSES OF HELL

Hell has always posed a challenge for theologians and Christian

philosophers for fairly obvious reasons. If God is all powerful as well
as perfectly good and loving, it is hard to see why he would consign
anyone to eternal punishment. So put, the problem of hell is best
situated as part of the larger problem of evil and the project of
theodicy. Indeed, as Alfred Freddoso has observed, the di culty of
reconciling even the possibility of anyone’s going to hell with the
existence of God is ‘ultimately the most troublesome form which the
problem of evil can take for the orthodox Christian’.10

Traditionally this problem has been met, at least among western

theologians, by defending eternal hell primarily as a matter of
retributive justice that God imposes on deserving sinners. The
essential argument for understanding hell in this fashion can be put
simply as follows.

(1) Any sin against God is in nitely serious.

 (2) An in nitely serious sin deserves a proportionate punishment,
which must also be in nite.

(3) God is perfectly just.

(4) Therefore, God must punish any sin that is not atoned for
with in nite punishment.
Among classical theologians who have defended hell on these
grounds are Augustine, Aquinas, Anselm, and Jonathan Edwards.

Anselm’s version of this account of hell was formulated as part of

his famous explanation of the atonement of Christ, an observation
that illustrates that the doctrines of heaven and hell are organically
connected to the larger body of Christian doctrine, particularly the
doctrine of salvation. His explication of why the atonement is
necessary hinges on his perception of the gravity of sin, and on this
matter Anselm leaves us in no doubt that he sees it as in nitely
serious. He makes the point most vividly by contending that it
would be better to let the whole world perish, and come to nothing,
even an in nite number of such worlds, than do the smallest thing
against God’s will. Jonathan Edwards gave a distinct twist to the
argument by casting it in terms of God’s in nite loveliness,
honorableness, and authority. These attributes place us under an
obligation to return to him a proportionate amount of love, honor,
and obedience. If we fail in this obligation, then again, we merit
in nite punishment. There are other variations on this argument,
less famous than the formulations of Anselm and Edwards, but they
share in common the attempt to show that sin is in nitely serious.
Among these variations are the claims that sin harms God in nitely,
that sin harms oneself in nitely, that sin harms other creatures
in nitely, and that sin harms an in nite number of creatures.

The punishment that classical theologians pictured as the tting

retribution for such sin is horri c beyond description. In elaborating
the misery of the damned, they typically distinguished between the
‘pain of loss’ and the ‘pains of sense’. The former of these refers to
the pain that naturally occurs from being separated from God, while
the latter was typically understood to include re of agonizing
intensity. Indeed, all conditions of the damned are such as to
contribute to what we might call maximal misery. For instance, in
commenting on the composition of light and darkness in hell,
Aquinas contends that ‘The disposition of hell will be such as to be
adapted to the utmost unhappiness of the damned. Therefore both
light and darkness are there, insofar as they are most conducive to
the unhappiness of the damned.’11

The thought of such horror is even more excruciating when we

add the traditional claim that hell is eternal and that there is no
possibility of escape, either by repentance, or suicide, or
annihilation. Naturally speaking, such su ering could not be borne,
but God will supernaturally keep body and soul together, and thus
force the damned to bear, paradoxically, unbearable agony and
torment. Leaving aside vivid imagery, we can sum up the traditional
view of hell in three propositions.
 (TH1) Some persons will never repent of their sins, and therefore
will not be saved.

(TH2) Those who remain impenitent at death will be consigned

to hell, a place of consummate misery that is the just
punishment for sins they commit in this life.

(TH3) There is no salvation after death so there is no end to hell

by repentance nor is there an end to it by suicide or
annihilation.

This third claim can be taken in two distinct ways, either as

contingently true or as necessarily true. Taking it the former way, it
would be possible to repent and escape hell, or to be annihilated,
but as a matter of fact, none will ever do so. This could be
epistemically certain even though metaphysically contingent, if it
were, for instance, revealed by God to be true.12 On the latter
reading, it is impossible to escape hell by repentance, and likewise
impossible to commit suicide or be annihilated. Moreover, the
impossibility of escaping hell by repentance could be due either to
the fact that repentance is psychologically or metaphysically
impossible for the damned, or to the fact that God would no longer
(could no longer?) accept their repentance.13

Although the defenders of this conception of hell number among

the most in uential of western theologians, it has come under re
among recent and contemporary philosophers and theologians and
would nd relatively few defenders today. Even most of those who
would defend hell as a matter of retributive justice are not inclined
to hold that the punishment of hell includes literal re or must
cause consummate misery. They would accordingly amend (TH2) to
state that hell is a place of misery, without adding that it must be of
the consummate variety.

But more fundamentally, the notion of eternal hell as a matter of

just punishment has been largely rejected by contemporary
philosophers. The so-called proportionality objection to hell
contends that in nite punishment is radically out of proportion to
any sins that nite beings could commit. Much of the recent
discussion has been devoted, accordingly to attacking premise (1) of
the argument above. Now if premise (1) of the above argument is
rejected, the argument obviously fails. Moreover, the second and
third claims of the traditional view of hell are undermined as well.

Among contemporary philosophers who have subjected the

notion that our sins deserve in nite punishment to searching
criticism are Marilyn Adams, Jonathan Kvanvig, and Charles
Seymour.14 A broad consensus has emerged from this work that the
punishment model of hell is not the best way to defend the doctrine.
This is not to deny that there have been attempts to defend a
punishment model, but those who have taken this line have not
typically tried to make the case for in nite punishment. James Cain,
for instance, defends the claim that eternal punishment, even
punishment that causes intense su ering, is compatible with God’s
justice, but he o ers models to show that endless experiential
su ering could be nite.15

In view of the emerging consensus rejecting the traditional

punishment model of the doctrine, philosophers have turned to
other ways to show how eternal hell can be compatible with God’s
perfect love and power. The most common strategy has been to
appeal to libertarian freedom to explain how human beings may be
separated from a loving God, even forever, and may consequently
experience eternal unhappiness. This is the view I have defended,
along with several other philosophers.16

It is worth noting at this point that the traditional view of hell

sketched above has often been defended by theologians who were
determinists and consequently did not accept libertarian freedom,
most notably Augustine and Edwards, and perhaps Aquinas as well.
Now the denial of libertarian freedom exacerbates the problem of
hell, for the sins that merit eternal punishment on this view are sins
that the damned were determined by God to commit. Those who
hold this view have the added di culty of explaining how sinners
who were determined to commit the sins they did could be
deserving of such a fate. This combination of divine determinism
and the traditional punishment model of hell is a mix that generates
glaring moral di culties that have frequently served as a ready
target for critics of orthodox Christianity. The appeal to libertarian
freedom is accordingly a double bene t, for it not only provides an
alternative to the traditional punishment model of hell, but also
repudiates the doctrine of divine determinism often associated if not
identi ed with orthodox theology, at least of the Protestant variety.

The heart of the view of hell that invokes libertarian freedom is

that we have the capacity to reject God, and may persist in doing so
to the point that we are forever separated from him. C. S. Lewis
famously summed up this view when he wrote ‘that the damned are,
in one sense, successful, rebels to the end; that the doors of hell are
locked on the inside’.17 More recently, Richard Swinburne has
argued that our freedom gives us the ability, over time, to form the
sort of character that we no longer desire God and the good. The
essence of damnation, as he sees it, ‘is a loss of good, not an
in icted evil; and it is not so much a punishment in icted from
without as an inevitable consequence of a man allowing himself to
lose his moral awareness’.18 Put in terms of the traditional
distinction noted above, this view emphasizes the pain of loss rather
than the pain of sense. Indeed, Swinburne thinks the New Testament
is ambiguous on the matter of whether hell must include any pains
of sense, and suggests that perhaps annihilation is the appropriate
end for those who reject God.

In a similar vein, Kvanvig has defended what he calls ‘the issuant

conception of hell’,19 so named because he believes the doctrine of
hell should ‘issue’ from the same attribute of God as the doctrine of
heaven. The common claim is that hell issues from God’s justice
whereas heaven issues from God’s love. It is Kvanvig’s contention
that heaven and hell both issue from divine love. However, he holds
that all of us face the ultimate free choice of whether we will live in
relation to God or live independently of him. The latter choice is
actually an illusory one, for none of us can actually do so, and to so
choose is actually to opt for annihilation. But not all those who
reject God choose annihilation in a clear and settled way, and it is
precisely in his love that God allows them to remain in existence.20
Kvanvig appropriately characterizes his view as a ‘composite’ one
since it allows both eternal separation from God and annihilation as
possible nal ends for the damned.21

Seymour has also defended a conception of hell that emphasizes

libertarian choice, as is evident from the fact that he calls his
position ‘the freedom view’. Unlike some others who stress freedom,
Seymour believes it is important for hell to include pains of sense as
well as the pain of loss. This is re ected in his basic de nition of
hell, namely, ‘an eternal existence, all of whose moments are on the
whole bad’.22 The importance of freedom for his view comes into
focus in his rejection of the traditional arguments that the sins
committed in this life could be su cient to warrant eternal
punishment. Rather, according to Seymour, what keeps sinners in
the perpetual su ering of hell is that they continue to sin. In
principle, sinners could repent and escape hell, so if they remain in
hell, it is due to their free choice to persist in sinning.

The notion of continuing sin in hell has been clearly illustrated

by Michael Murray, who gives the example of a criminal who is
given a twenty-year sentence, but then commits further crime while
in prison. Although none of his crimes taken alone deserve a life
sentence, the cumulative sentence for his ongoing crimes is never
completed. Likewise, the ‘unchecked sinful desires’ of the damned
may ‘continue to lead them to sin even in hell and so continue to
mount penalties which are never satis ed’.23 This account of what
keeps people in hell, it is worth noting, is compatible with the
contingent version of (TH3). There is no end of hell for those who
go there, but it is due to their own choice.

The attempt to defend the contingent version of (TH3) by

appealing to the continuing sin model has been challenged by
Kenneth Himma.24 Taking Murray as his foil, Himma contends that
the proportionality problem is not solved if one holds that hell is
eternal for all who enter it and there is no escape for those who die
without salvation. Before developing his main argument, he points
out that whether the su ering of hell is disproportionate depends on
both the duration and the intensity of that su ering. Like most
recent philosophers, he agrees that if hell is intensely painful, as it
would be if it included literal re, as well as eternal, it would be a
disproportionate punishment for any sins we might commit in this
life. He also considers two distinct ‘separation’ conceptions of hell.
In the rst, the damned are unaware that God exists and therefore
unaware that they are separated from him. Their only punishment is
that they are deprived of the in nite good of communion with God,
though they do not know this. The other separation model includes
awareness of what is lost by those excluded from communion with
God. This knowledge increases the su ering of the damned because
they would regret their loss as well as the choices that led to their
condition.

Now Himma contends that it seems likely on the separation-

with-knowledge conception that some persons could escape hell
because not all would choose to persist in sin. Sincere advocates of
other religions who genuinely seek the truth, as well as ethical
unbelievers, for instance, are persons who would be likely to repent
if they discovered that their mistaken choices had led to their
separation from God. The separation-without-knowledge view does
not fare much better but for di erent reasons. While persons
without the relevant knowledge would not be likely to repent, it
seems unjust to keep them in hell because they have no realistic
chance to reform themselves. Thus, Himma concludes that the
continuing sin scenario does not provide an adequate defense of the
traditional view that there is no end of hell for those who die
without salvation. His argument does not lead to universalism, but it
does leave open the possibility, if not probability, that at least some
of the inhabitants of hell would eventually nd their way to heaven.
And if he is right about this, there is good reason to think (TH3) is
not even probably true, let alone epistemically certain.

In drawing this section to a conclusion, note that the views of

Himma, Kvanvig, Lewis, Seymour, and Walls agree in denying
(TH3), the third claim of the traditional view listed above. Kvanvig,
and perhaps Lewis, think it is possible to escape hell by making
choices that lead to annihilation. Moreover, Himma, Kvanvig,
Lewis,25 Seymour, and Walls think that sinners in hell can in
principle repent, and if they do so, they will be accepted by God. Of
course, this claim is also compatible with the contingent version of
(TH3). However, my own reason for denying (TH3) altogether is
that I believe it is likely that some persons who die without
receiving salvation will do so after death. My position hinges on my
argument that it follows from God’s perfect love and goodness that
he will give all persons ‘optimal grace’, the best opportunity they
can have to be saved, and only if they decisively reject grace will
they nally be damned. It seems apparent to me on empirical
grounds that many persons have not received optimal grace or made
a decisive choice in this life, so it seems likely that many will do so
in the next life.26

The point I want to emphasize now, however, is that if sinners

are forever lost, either because of annihilation or their persistent
refusal of grace, their fate depends on their continuing choices
rather than only upon choices they made in this life. There is thus
an appropriate match between their ongoing choices and their
ongoing existence in hell, or annihilation, which resolves the
proportionality problem.

CHALLENGES TO FREEDOM ACCOUNTS OF HELL

While the appeal to freedom provides ways to relieve the notion of

eternal hell of its most pressing moral di culties and to allow fresh
formulations of the doctrine, these new formulations have not
escaped criticism. Indeed, a growing number of Christian
philosophers and theologians are challenging the rst claim of the
traditional doctrine of hell (TH1) and advancing universalism, the
view that all persons will eventually be saved.

The topic of universalism has been on the table as part of the

contemporary discussion in the eld of philosophy of religion at
least since the publication of John Hick’s landmark work Evil and the
God of Love. In that work, in which Hick defended an Irenaean view
of theodicy over against the Augustinian view that has
predominated in western theology, he also argued for universal
salvation as a ‘practical certainty’. Given his view that human free
choices are undetermined, along with his view that ‘the thoughts
and actions of free beings are in principle unknowable until they
occur’,27 Hick could not go so far as to rule out the possibility that
some may never be saved. Thus, it is metaphysically and logically
possible that all may not be saved, and whether or not this is so
must remain epistemically uncertain, even for God. But given the
in nite resources of God’s love, his perfect understanding of his
creatures, and endless time to work on them, Hick says it seems
morally impossible that God could fail in his quest to save them all,
and therefore practically certain that he will succeed.28

More recently, philosophical advocates of universalism have

mounted stronger assaults against the barrier of human freedom in
making their case that all will be saved. One of these is Marilyn
Adams, who also has given signi cant attention to the issues of hell
and universalism as part of her larger work on theodicy. Judging
this debate to have grown stale, Adams, like Hick over a generation
ago, attempts in a recent work to o er new resources for theodicy
and to point it in some fresh directions. Among her more radical
suggestions is that sin is primarily a metaphysical problem rather
than a moral one. One aspect of this problem particularly pertinent
to our concerns is what she calls the ‘size gap’ that separates us from
God. Failure to appreciate the reality of this size gap causes us to
exaggerate the dignity of human nature as something so sacrosanct
that not even God may legitimately interfere with it.

Adams sees this tendency especially manifested in what she calls

‘mild’ versions of the doctrine of hell that emphasize our freedom to
reject God and the corresponding misery that is the natural
consequence of that choice. She complains that this view treats God
and human adults as moral peers, and objects that giving us the
right to damn ourselves is not the appropriate sort of respect for
God to pay to the likes of us. Whereas this account pictures the
relationship between us and God with the analogy of parents and
adult children, Adams suggests that a better model would be the
relationship between a mother and an infant or a toddler. On
Adams’s preferred model, there is no meaningful sense in which the
child is free or responsible, nor is there any basis for objecting to
interfering with his choices.
 It is easy to see how this undermines the freedom view of hell
and allows Adams to argue that God can save everyone. Given the
size gap between us and God on her picture, human freedom poses
no obstacle to universalism. If God needs to override our freedom
and causally determine things in order to prevent some of his
children from eternal damnation, this is ‘no more of an insult to our
dignity than a mother’s changing a diaper is to the baby’.29

While this move does indeed eliminate human freedom as a

barrier to universal salvation, the question is whether Adams has
paid too high a price to achieve this result in her willingness to
minimize, if not sacri ce, moral categories. It is also doubtful if her
mother-infant image of the divine human relationship does justice
to the biblical picture of that reality. It is not insigni cant that the
Bible often portrays the divine-human relationship with the image
of a husband and wife or a lover and his beloved, and sin is pictured
as unfaithfulness to this relationship. This image certainly suggests
that God desires and expects far more from us in terms of behavior,
including responsible moral behavior, than a mother could ever
expect from an infant in diapers or a toddler.30

Another philosopher who has mounted a sustained attack on the

doctrine of eternal hell is Thomas Talbott. In addition to
philosophical arguments, he has also given signi cant attention to
biblical exegesis in making his case for universalism.31 This is vital
to the overall case because for many believers, the matter is
decisively settled by clear biblical teaching. Typical of this line of
thought is Peter Geach, who took Christ’s teaching as ‘perfectly
clear’ that ‘many men are irretrievably lost’ and insisted on this
basis that ‘universalism is not a live option for a Christian’.32

By sharp contrast, Talbott wants to argue that universalism is the

only option for a Christian. His philosophical case for this strong
claim is largely focused on his argument that the notion that anyone
could choose eternal hell is nally incoherent because there simply
is no intelligible motive to explain such a choice. Now this is not to
deny that we can choose evil in the short run. Obviously we can do
so, typically under the illusion that this will bring us the happiness
we crave. Talbott insists, however, that God will eventually shatter
the illusion that motivates the choice of evil by bringing home to us
a hard-edged reality check. The more we choose evil, the more
miserable we will become. Under the impact of ‘greater and greater
misery’33 we will reach a point that we cannot bear it any longer
and turn to God, the true source of the happiness we are seeking.

Talbott’s argument for this conclusion hinges crucially on his

account of what is involved in the free choice of an eternal destiny.
The essence of such a choice is that it must be fully informed in such
a way that once the person making the choice gets what he wants,
he never regrets his choice. This requires that the choice must be
free from ignorance and illusion both initially as well as in the long
run. The person making it must fully understand what he has chosen
and freely persist in that choice.
 With this account of what is involved in choosing an eternal
destiny clearly on the table, Talbott thinks there is an obvious and
crucial asymmetry between choosing fellowship with God in heaven
as an eternal destiny, on the one hand, and choosing eternal misery
in hell, on the other. Whereas it is overwhelmingly clear that it is
possible to choose the rst option because there are readily
intelligible motives for it, the latter is not possible because it makes
no sense at all to say that a rational creature could knowingly
persist in the choice of eternal misery.

Next, in addition to his view of what is involved in freely

choosing an eternal destiny, Talbott’s argument for universalism
also depends on his view that the New Testament picture of hell is ‘a
forcibly imposed punishment rather than a freely embraced
condition’, that it is ‘unbearable su ering’.34 Indeed, Talbott chides
those who say hell is a freely embraced condition for holding a view
that ‘in e ect takes the hell out of hell, as least as far as the damned
are concerned’.35 The claim that hell is unbearable misery
underwrites his claim that it is unintelligible that anyone could
persist in freely choosing it forever. Rather, the misery of hell will
eventually move even the most hardened sinners freely to choose
heaven for their eternal destiny.

It is important to stress that Talbott a rms a libertarian view of

freedom at least in the sense that he agrees that ‘creaturely freedom
could never exist in a fully deterministic universe’.36 He has deep
reservations, however, about standard views of libertarian freedom
that assert two essential claims.

(1) a person S performs an action A freely at some time t only if

it should also be within S’s power at t to refrain from A at t;
and
(2) it is within S’s power at t to refrain from A at t only if
refraining from A at t is psychologically possible for S at t.37

Talbott challenges the second of these claims by drawing a

distinction between the power to do something and the psychological
capability of doing it. This distinction is a third component of his
argument for universalism. As he notes, in narratives of dramatic
conversions, such as those of St Paul and C. S. Lewis, ‘the nal act of
submission seems to occur in a context where the alternative is no
longer psychologically possible at all’.38 Another relevant
illustration of the di erence between power and psychological
ability is Augustine’s account of how the redeemed in heaven will
no longer even be tempted to sin or disobedience. Does this mean
they will no longer be free or that they lack the power to sin? No,
they will not lack the power to sin, even though they will no longer
be psychologically capable of sinning. Because they will see with
such perfect clarity that God is the true source of happiness and sin
the cause of misery, sin will lose all appeal and will no longer
remain a live option for them.39

To sum up Talbott’s view in relation to the traditional view of

hell sketched above, it is clear that he rejects both (TH1) and (TH3).
However, he a rms a version of (TH2). While he may not agree
that the su ering of hell is consummate misery, he does insist that it
is unbearable su ering that no one could freely choose to undergo
forever. Thus, for Talbott, the notion of eternal hell is deeply
incoherent, and universalism is not merely possibly true, nor even
practically certain as it is for Hick; it is necessarily true.

THE COHERENCE OF ETERNAL HELL

Talbott’s challenge to recent versions of the doctrine of eternal hell

is a powerful one that requires an answer, so let us consider it more
carefully. Let us begin with his distinction between power and
psychological ability, a distinction I am inclined to accept. Even
with this distinction clearly in hand, there are serious problems with
Talbott’s claim that all persons can be brought freely to repent in a
non-deterministic sense under the pressure of forcibly imposed
punishment. The central di culty with this claim is that the notion
of ‘greater and greater misery’, misery that does not have a distinct
limit, destroys any meaningful notion of a free choice. The reason
for this is that nite beings such as ourselves are not built with the
capacity to absorb ever-increasing misery. At some point in the
process of being dealt greater and greater misery we would
inevitably crack and would either be coerced to submit or be
psychologically and mentally shattered or perish altogether.40
Appealing to Talbott’s distinction will not help, for we have neither
the power nor the psychological capability to stand up under
constantly increasing misery, regardless of whether that misery is of
the physical or emotional variety. In other words, there is a limit to
how much pressure our freedom can bear. Where exactly that limit
lies is perhaps something only God could know, but clearly there is
such a limit.

Now this point is particularly interesting in light of a distinction

Talbott himself has drawn between two kinds of compulsion. The
right kind of compulsion, which he defends, is illustrated in
dramatic conversions such as that of C. S. Lewis, who had a sense of
God closing in on him in such a way that it seemed impossible for
him to do otherwise than submit.41 By contrast, he repudiates the
sort of compulsion advocated by Augustine, who was prepared to
use the sword to persuade the Donatists to return to the church. He
spells out the distinction as follows: ‘A stunning revelation such as
Paul reportedly received, one that provides clear vision and
compelling evidence, thereby altering one’s beliefs in a perfectly
rational way, does not compel behavior in the same way that
threatening with a sword might.’42 As Talbott goes on to note,
swords are not evidence, so they provide no rational reason to alter
one’s beliefs. Conversion at sword’s point is neither rational nor
morally acceptable, nor is it free in any meaningful sense of the
word.

Talbott’s distinction is another one I am inclined to accept, but

precisely this distinction poses problems for him, given his
insistence that hell is a matter of ‘forcibly imposed punishment’, and
‘unbearable su ering’. The traditional accounts of the su ering of
hell that Talbott professes to endorse surely include physical pain of
a rather intense variety. If he does not believe this, he should not
pretend to be espousing the traditional understanding of what
makes hell literally unbearable. But if he really does support the
traditional view, then he owes us some explanation of how forcibly
imposed punishment that produces unbearable su ering is not the
wrong kind of compulsion that he repudiates. Without this sort of
explanation, his view su ers from obvious inconsistency. This point
holds, incidentally, even if Talbott thinks the misery of hell is purely
psychological and spiritual. If conversion at sword point is
objectionable, then it is surely equally objectionable to compel
repentance by any sort of forcibly imposed misery, even if that
misery is ‘only’ spiritual and psychological.

But there is another di culty with Talbott’s account of the

proper kind of compulsion, namely, his appeal to ‘compelling
evidence’ that alters beliefs in a perfectly rational way. The
di culty is not in the claim that evidence can so alter our beliefs.
Indeed, it is arguably the very heart of rationality to be willing to
alter one’s beliefs in light of appropriate evidence, especially if that
evidence is staring you in the face! The problem is in the claim that
such evidence can ever be truly compelling, a problem that is very
much compounded by the fact that religious truth claims are highly
contested, including the claim that God even exists.
 This is not to deny that there is substantial evidence in favor of
Christian faith. Indeed, it is arguable that a certain amount of
evidence is necessary in order to keep faith from being an irrational
or arbitrary commitment. However, it is also arguable that the
evidence must fall short of being truly compelling for our response
to be free. Pascal represented such a balance when he wrote the
following: ‘But the evidence is such as to exceed, or at least equal,
the evidence to the contrary, so that it cannot be reason that decides
us against following it, and can therefore only be concupiscence and
wickedness of heart.’43

At the end of the day, for Pascal, neither belief nor unbelief is
simply a matter of evidence or recognition of certain facts. Belief
requires a heart that is rightly disposed, and unbelief exposes a
wicked heart. And if this is correct, then evidence alone can never
be compelling for genuine faith. The evidence can be compelling in
the more modest sense that it makes clear the disposition of our
hearts. However, neither evidence that is compelling in this sense
nor unbearable su ering can guarantee the sort of free response that
God desires from us. Evidence can never be compelling in the sense
that Talbott’s argument requires and unbearable su ering cannot
elicit the sort of repentance that is rationally motivated and morally
free.

Now in response to these sort of criticisms, Talbott has clari ed

his position in some important respects. First, he has made clear
that those who repent under the pressure of ever-increasing misery
 nally have no choice but to turn to God, so this choice is not free
in the libertarian sense. The choices we make to reach this point are
free in the libertarian sense, but the ironic result of these choices is
that they take us to a point where the only thing we can do is repent
of those choices.

Second, and more importantly, Talbott has explained just what

he means by the ‘unbearable su ering’ and the ‘forcibly imposed
punishment’ of hell. And it is now clear that what he means by these
terms is very di erent from how they are typically understood in
traditional accounts of hell, despite what some of his previous
writings suggested. To illustrate the sort of su ering he has in mind,
he invites us to consider the example, inspired by a well-known
movie plot, of a married man who foolishly has an a air with an
unstable woman. When the a air ends, she takes out her revenge by
murdering his wife and child (he does not mention the fate of the
bunny rabbit).44 We can imagine that the man’s subsequent guilt,
sorrow, and profound sense of loss would be an unbearable su ering
that God could use to move him to repent, and ‘insofar as God uses
the man’s su ering as a means of correction, or as a means to
repentance, we can again say that the man has endured a forcibly
imposed punishment for his sin’.45 He goes on to point out that ‘the
good in the worst of sinners—the indestructible image of God if you
will—can itself become a source of unbearable torment’. The ‘good’
Talbott has in mind is especially our moral nature, our conscience,
that causes us to feel guilt and regret and thereby spurs us to
repentance.

Now these clari cations are indeed helpful for gaining an

accurate picture of Talbott’s position but unfortunately, these very
clari cations expose further di culties for his case for necessary
universalism. The most glaring problem is that his explanation of
what he means by unbearable torment no longer sustains his claim
that all sinners must inevitably reach a point where resistance
would no longer be possible. This claim was tenable so long as we
took him to mean that the forcibly imposed punishment of hell is a
matter of ever-increasing su ering of such intensity that no one
could bear it inde nitely. Ironically, however, Talbott seems to have
tamed the bear of unbearable su ering and made it quite possible to
embrace.

Consider his example of the foolish philanderer. Granted that his

actions and the consequences that followed would surely cause him
great misery, it is still highly doubtful that his repentance would be
inevitable. No doubt Talbott is right: God could use his su ering as
a way to show him the folly of his actions and to move him in the
direction of repentance. But there is nothing in the case as described
that makes it inevitable that he will remain on the road to
repentance and truly come to see the error of his ways. Indeed, we
can easily imagine that rather than sincerely repenting and
reforming, he might become angry and embittered if he thought
God allowed the murder of his family in order to punish him for his
a air. Rather than viewing God as a just judge who rightly created a
moral universe where actions have consequences, he might come to
see God as a vengeful deity more deserving of scorn than worship
and obedience.

Similar points can be made about Talbott’s observation that the

good in sinners can become the source of unbearable torment.
Again, the question is whether there is any way to relieve this
torment, or whether it invariably must become greater and more
intense. Contrary to Talbott’s scenario of ever-increasing misery, it
is a well-known psychological reality that we do have the power to
dull and desensitize our consciences. One way we can do this is by
deliberately choosing to continue performing those very acts that
caused our guilt and remorse in the rst place. Moreover, we can
rationalize our actions or restructure our values in such a way that
our conscience no longer bothers us. Such choices require a
signi cant degree of self-deception, and indeed, the ability to
deceive ourselves may be an essential component in our freedom to
love God or not.

As Talbott himself has observed, we have the freedom,

‘expressed in thousands of speci c choices, to move incrementally
either in the direction of repentance and reconciliation or in the
direction of greater separation from God, and that freedom God
always respects’.46 Now if this is true, the question that begs to be
answered is why such incremental movement must inevitably arrive
at a point where our illusions would be shattered and repentance
would be unavoidable. Intuitively, it seems just the opposite would
occur. If we can move further and further from God, this would
mean we would become more and more hardened in sin,
correspondingly less and less aware of his goodness, and therefore
less and less likely to repent and be reconciled to him. Objectively,
such a person would become ever more miserable as he moved ever
further from a loving relationship with God, but subjectively his
hardness of heart could make the misery more tolerable.

Of course, God could intervene and circumvent the natural

hardness of heart that would develop as we moved ever further from
him. Moreover, he could impose other sorts of pain on us, such as
ever more intense physical agony with each incremental move away
from himself. Were God to impose such su ering on us, then at
some point everyone would reach a point where they would crack
and have to submit to God. But given Talbott’s distinction between
two kinds compulsion, he would presumably reject this as the wrong
variety.

What this means is that Talbott has a dilemma on his hands,

indeed, a hell of a dilemma. He must choose between his clari ed
account of divinely imposed misery and his claim that all sinners
must inevitably arrive at a place where resistance to God is no
longer possible. Given his clari ed account of forcibly imposed
punishment, there is no convincing reason why rebellious sinners
could not go on rejecting God forever. But if he reverts to the claim
that the su ering of hell induces repentance through the imposition
of pain and agony so intense that it is literally impossible for anyone
to bear it, then he will be endorsing the ‘wrong’ kind of compulsion.
Moreover, this view of the pains of hell is at odds with his claim
that God grants us the freedom to move ever further away from
him. In order to maintain his case for necessary universalism,
Talbott must choose the horn of the dilemma that involves
embracing the sort of compulsion that he would prefer to repudiate.
Otherwise, his case for necessary universalism fails.47

OBJECTING TO HEAVEN

As noted above, the doctrine of heaven has not received as much

attention from philosophers as hell has. Nevertheless, there have
been some interesting challenges raised against the Christian hope
of eternal joy. Ironically, one of these can be generated from the
argument just developed. Suppose it is not merely possibly true that
some will be lost forever, but, as many traditional theologians have
contended, some, perhaps many, surely will be. If this were true,
would it not detract from the joy of the saved? As stated by Eric
Reitan, the crux of the argument, which goes back at least to the
nineteenth-century theologian Friedrich Schleiermacher, ‘is that the
eternal damnation of anyone is incompatible with the salvation of
any, because knowledge of the su erings of the damned would
undermine the happiness of the saved’.48

Obviously, the argument assumes that the saved would not only
be aware of the damned but would also empathize with their
miserable condition. Since the saved would be fully transformed and
perfected in love, they would surely have a deep love for all persons
and desire their salvation. So if any remain separated from God,
none could experience the perfect happiness that is supposed to
characterize heaven. Indeed, the argument can be extended to claim
that not even God himself could be fully happy if some are forever
lost.49 Although this argument is usually cast as an argument
against eternal hell, it can also be used against the doctrine of
heaven since it says in essence that heaven and eternal hell are
incompatible.

There have been several attempts to blunt the force of this

argument by suggesting various scenarios in which eternal hell can
be compatible with perfect happiness.50 For instance some have
proposed that the blessed are not aware of the lost either because
they are in a state of such overwhelming bliss that they would never
think of them, or because the memory of them has been eliminated.
Others have argued that the saved will see the nature of evil with
such perfect moral clarity that they will not be disturbed by the
reality of damnation. They will clearly understand the damned have
freely chosen their fate, and if they will not repent, they cannot be
allowed to hold heaven hostage. The damned can refuse to accept
God’s invitation to joy but they cannot have the right or power to
spoil the happiness of those who have accepted the invitation
simply because they will not.
 A second objection to heaven that is being advanced in
contemporary literature focuses on the historical origins of the
doctrine. In the spirit of Nietzsche’s suspicious reading of the
‘genealogy’ of morality, some recent historians have written
histories of the afterlife that purport to show that doctrines of
heaven (and hell) were formulated with the fuel of dubious
motivations.51 J. Edward Wright, for instance, argues that Jewish
and Christian visions of heaven were often prompted by a desire to
prove God’s special favor for their group. And while these visions
did indeed inspire hope for members of the chosen group, they also
created ‘a dangerous sense of superiority over outsiders by
dehumanizing or demonizing them’.52 Similarly, Alan Segal has
contended that doctrines of the afterlife not only mirror the values
of the societies that hold them, but also ‘have bene ted a particular
social class and served to distinguish the purveyors of the idea from
their social opponents’.53

While it is no doubt true that many persons who have held and
promoted beliefs about heaven have been at least partially
motivated by the dubious sort of motivations cited by Wright and
Segal, there is more to the story, even as told by them. As Wright
notes, the idea of the afterlife in Judaism originally grew out of
bedrock convictions about the ultimate reality of God’s goodness
and justice. The treacherous injustice of this life poses an obvious
challenge to these convictions and they could be sustained only if
this life is not the last word.54 Now if the deepest roots of belief in
heaven are tied up with convictions about God’s moral nature, then
the hope of heaven at its best is motivated by our highest and best
aspirations for goodness and justice. This part of the story provides
resources to critique dubiously motivated belief in heaven where it
occurs as well as to challenge histories of the doctrine that unduly
emphasize such motivation.

The nal objection to heaven I shall mention is a poignant one,

with more than a touch of irony. In a famous article entitled ‘The
Makropulos Case: Re ections on the Tedium of Immortality’,
Bernard Williams contended ‘that an endless life would be a
meaningless one, and that we could have no reason for living
eternally a human life’.55 As Williams sees it, the hope of eternal
happiness is actually an incoherent one, because no matter what
delights heaven might hold, inevitably it would eventually become
boring. Whereas previous generations have looked forward to
heaven as the profound cure for what ails us in this life, Williams
says there is no cure. The boredom and meaninglessness that haunt
this life now threaten heaven as well, so perhaps the best we can
hope for is to die ‘shortly before the horrors of not doing so become
evident’.56

A number of responses to the boredom challenge have been

o ered in the recent literature.57I am inclined to think, however,
that the crux of the issue comes down to whether or not we nd the
idea of God intelligible, whether or not we believe that a being
whose very nature is to be ecstatically happy has existed from all
eternity, and will do so forever. If there is such a God, the hope that
he might share his in nite happiness with us is a rational one, even
if it hard fully to imagine in our current situation. Indeed, perhaps
the fact that we fear boredom, even in heaven, is a telling mark of
how far we have fallen from our intended relationship with God.
The fact that Williams suggests that we should abandon the hope of
heaven and nd the meaning of our lives in a timely death before
boredom completely sets in is, moreover, a stark indicator of how
far our hope for happiness and meaning have fallen.

Recall Quine’s autobiographical musing in which he recalled that

the loss of his juvenile faith in heaven was attended by a sense of
disappointment. This is hardly surprising. Anyone who has had even
a glimpse of the beauty that heaven might hold or has felt even a
murmur of the hope of eternal joy stirring in his heart could hardly
fail to be disappointed if he came to believe it was not real. While
heaven has received scant overt attention in recent philosophical
literature, it does sometimes enter the conversation in some
backhanded ways that betray this disappointment. This comes
through most tellingly perhaps, in the various secular substitutes for
heaven that secular philosophers have o ered. Indeed, what is
perhaps most striking is how secular philosophers sometimes
advance the extraordinary claim that they actually have something
better than heaven, a claim more notable for its bravado than its
plausibility.58
 Once the concept of heaven has entered our hearts and minds,
there is no alternative to disappointment if it is not real. Heaven
holds out the hope of perfect happiness, a depth and intensity of joy
beyond what we can fully imagine, that is literally without end.
Never to know such joy, to deny it could even exist, to claim that
nite joy with a timely death is the best possible outcome, all of
these are disappointing to the highest dreams for happiness we have
dared to dream. Only heaven makes possible an understanding of
life not nally resigned to disappointment.

NOTES

1. The Book of Heaven, ed. Carol and Philip Zaleski (New York:
Oxford University Press, 2000), 5.

2. See J. Edward Wright, The Early History of Heaven (New York:

Oxford University Press, 2000); Alan F. Segal, Life After Death:
A History of the Afterlife in Western Religion (New York:
Doubleday, 2004).

3. For a recent explication of this idea that draws on classic

resources, see John Saward, Sweet and Blessed Country (Oxford:
Oxford University Press, 2005), 15–55.

4. Cited by Kenneth L. Woodward, ‘Heaven’, Newsweek (27 March

1989), 54.

5. W. W. Quine, The Time of My Life (Cambridge: MIT, 1985), 14.

 6. Colleen McDannell and Bernhard Lang, Heaven: A History (New
York: Vintage, 1990), 351–2.

7. Harvard Theological Review 78: 3–4 (1985), 381–98.

8. ‘Hell Disappeared’, 393.

9. ‘Hell’s Sober Comeback’, U.S. News and World Report (25

March 1991), 56. Nine years later, hell was the cover story in
the same magazine. See Je ery L. Sheler, ‘Hell Hath No Fury’,
U.S. News and World Report (31 January 2000), 44–50.

10. Alfred J. Freddoso (ed.), The Existence and Nature of God (Notre
Dame: University of Notre Dame Press, 1983), 3. Although hell
has not been de ned with ‘orthodox’ precision in anything like
the way Christology has been de ned in the Councils of Nicea
and Chalcedon, the doctrine of eternal hell is nevertheless a
matter of broad consensus among orthodox Christians of all
the major traditions.

11. Summa Theologiae, Suppl. to III q. 97 a. 4.

12. This assumes, of course, that God knows future contingents.

13. Charles Seymour analyzes Aquinas’s arguments for the claim

that the damned are ‘frozen in their wickedness’ and unable to
repent. He contends that although Aquinas’s argument appears
at rst to be metaphysical, it rests on the moral premise that
the damned deserve eternal punishment for past sins. See his A
Theodicy of Hell (Dordrecht: Kluwer Academic, 2000), 167–70.
14. Marilyn McCord Adams, ‘Hell and the God of Justice’, Religious
Studies 11 (1975): 433–47; Jonathan L. Kvanvig, The Problem
of Hell (Oxford: Oxford University Press, 1993), 25–66;
Seymour, A Theodicy of Hell, 37–94.

15. James Cain, ‘The Problem of Hell’, Religious Studies 38 (2002),

355–62.

16. Jerry L. Walls, Hell: The Logic of Damnation (Notre Dame:

University of Notre Dame Press, 1992), 113–38.

17. C. S. Lewis, The Problem of Pain (New York: Macmillan, 1962),

127.

18. Richard Swinburne, Responsibility and Atonement (Oxford:

Clarendon, 1989), 182. See also ‘A Theodicy of Heaven and
Hell’, in The Existence and Nature of God, 46–52.

19. Kvanvig, The Problem of Hell, 112.

20. Eleonore Stump has also defended a variation of the view that
God in his love sustains the damned in existence. See ‘Dante’s
Hell, Aquinas’s Moral Theory and the Love of God, Canadian
Journal of Philosophy 16 (1986), 181–98.

21. Kvanvig, The Problem of Hell, 151–9. Kvanvig notes,

incidentally, that sometimes C. S. Lewis’s comments about hell
implied that the damned could be annihilated: see 121–3.

22. Seymour, A Theodicy of Hell, 161.

23. Michael J. Murray, ‘Heaven and Hell’, in Michael J. Murray

(ed.), Reason for the Hope Within (Grand Rapids: Eerdmans,
 1999), 293.

24. Kenneth Himma, ‘Eternally Incorrigible: The Continuing-Sin

Response to the Proportionality Problem of Hell’ Religious
Studies 39 (2003), 61–78.

25. In The Great Divorce, Lewis’s fantasy novel of sinners in hell

who take a bus ride to heaven, one of the characters explains
that if any choose to remain in heaven and leave the grey
town (hell) behind, then it is purgatory. See The Great Divorce
(San Francisco: Harper, 2001), 68.

26. See Walls, Hell, 83–105. Seymour also takes this view: see A
Theodicy of Hell, 167. See also Andrei A. Buckare and Allen
Plug, ‘Escaping Hell: Divine Motivation and the Problem of
Hell’, Religious Studies 41 (2005), 39–54.

27. John Hick, Evil and the God of Love (San Francisco: Harper &
Row, 1966), 343. Gordon Knight has recently argued that
‘open theists’, those theists who hold that God cannot have
infallible knowledge of our future free choices, should
embrace ‘contingent universalism’. See Gordon Knight,
‘Universalism for Open Theists’ Religious Studies 42 (2006),
213–23.

28. Hick, Evil and the God of Love, 343–4. If God is necessarily
good, and something is morally impossible, then it should be
metaphysically impossible as well, for God’s nature limits what
 is metaphysically possible. For further criticism of Hick, see
Walls, Hell, 70–81.

29. Marilyn McCord Adams, Horrendous Evils and the Goodness of

God (Ithaca: Cornell University Press, 1999), 157.

30. See Katherin A. Rogers, ‘The Abolition of Sin: A Response to

Adams in the Augustinian Tradition’, Faith and Philosophy 19
(2002), 69–84.

31. Talbott lays out his biblical argument in Robin A. Parry and
Christopher Partridge (eds.), Universal Salvation? The Current
Debate (Grand Rapids: Eerdmans, 2003), 15–52. This volume
also includes critiques of Talbott’s universalism by several
authors representing di erent disciplines, including
philosophy. Talbott’s biblical arguments are critiqued by I.
Howard Marshall and Thomas F. Johnson. The book also
includes Talbott’s replies to his critics.

32. Peter Geach, Providence and Evil (Cambridge: Cambridge

University Press, 1977), 123–4. Geach thought it less clear that
the ultimate fate of the lost is eternal misery rather than
destruction.

33. Thomas Talbott, ‘Freedom, Damnation and the Power to Sin

with Impunity’, Religious Studies 37 (2001), 420.

34. Ibid. 417.

35. Ibid. 429–30.

36. Ibid. 426.

37. Ibid. 426.

38. Ibid. 426.

39. Thomas Talbott, ‘On the Divine Nature and the Nature of
Divine Freedom’ Faith and Philosophy5 (1988), 13.

40. Some traditional theologians have held that God supernaturally

sustains the damned in such a way that he causes them to
endure what they could not otherwise endure in order to
maximize their misery. Of course, Talbott assumes no such
thing and would surely reject such a notion.

41. Talbott quotes passages from Lewis’s spiritual autobiography to

make this point. See C. S. Lewis, Surprised by Joy (New York:
Harcourt Brace Jovanovich), 224, 229.

42. Talbott, ‘Freedom’, 427.

43. Blaise Pascal, Pensées, trans. A. J. Krailsheimer (London:

Penguin, 1966), no. 835.

44. Presumably he has in mind the movie ‘Fatal Attraction’,

although the wife and child do not share the same fate as the
bunny rabbit in this lm.

45. Talbott, ‘Misery and Freedom: Reply to Walls’ Religious Studies

40 (2004), 218.

46. Ibid. 221–2.

47. This section reiterates and summarizes some of the key points
in two papers in which I criticize Talbott’s views. For further
 argument of these points, see my ‘A Hell of a Choice: Reply to
Talbott’ Religious Studies 40 (2004), 203–16; and ‘A Hell of a
Dilemma: Rejoinder to Talbott’, Religious Studies 40 (2004),
225–7.

48. Eric Reitan, ‘Eternal Damnation and Blessed Ignorance: Is the

Damnation of Some Incompatible with the Salvation of Any?’
Religious Studies 38 (2002), 429.

49. I have responded to this line of the argument in Hell, 106–10. I

argue, moreover, that if God can be perfectly happy despite
eternal hell, then the saved may share his perspective and also
be perfectly happy. For Reitan’s critique of my argument, see
‘Eternal Damnation and Blessed Ignorance’ 432–4.

50. In the article just cited, Reitan surveys and criticizes several of
these arguments. It is worth noting, incidentally, that this
problem is one that would not have bothered many classical
theologians. Indeed, the likes of Augustine and Aquinas argued
that contemplating the misery of the damned would provide a
contrast that would actually enhance the joy of the saved. See
D. P. Walker, The Decline of Hell (Chicago: University of
Chicago Press, 1964), 29.

51. For Nietzsche’s own thoughts along this line, see The Birth of
Tragedy & The Genealogy of Morals, trans. Francis Gol ng (New
York: Anchor, 1956), 178–85.
52. J. Edward Wright, The Early History of Heaven (Oxford: Oxford
University Press, 2000), 202; see also 137, 157, 163, 177.

53. Alan Segal, Life After Death: A History of the Afterlife in Western
Religion (New York: Doubleday, 2004), 699; see also 11, 16,
68, 243, 292, 344.

54. Wright, The Early History of Heaven, 158, 191–2. See also Segal,
Life After Death, 271–2.

55. Bernard Williams, ‘The Makropolus Case: Re ections on the

Tedium of Immortality’, in John Martin Fischer (ed.), The
Metaphysics of Death (Stanford: Stanford University Press,
1993), 81.

56. Williams, ‘Makropolus’ 92.

57. Garth Hallett surveys six proposed solutions to this di culty in

‘“The Tedium of Immortality”’, Faith and Philosophy 18 (2001),
279–81. See also John Martin Fischer, ‘Why Immortality is Not
So Bad’ International Journal of Philosophical Studies 2: 2
(1994), 257–70; Jerry L. Walls, Heaven: The Logic of Eternal Joy
(New York: Oxford University Press, 2002), 193–200; Arthur
O. Roberts, Exploring Heaven: What Great Thinkers Tell us About
Our Afterlife With God (San Francisco: Harper, 2003), 112–28.

58. I have documented a number of these secular substitutes for

heaven. See my Heaven, 177–85.
 CHAPTER 23

THE EUCHARIST: REAL PRESENCE AND

REAL ABSENCE

ALEXANDER R. PRUSS

1. INTRODUCTION

THE Eucharist has traditionally been the center of Christian liturgical

life. Typically the liturgy includes Jesus’ words from the Last
Supper: ‘This is my body’ and ‘This is my blood.’ Until the
Reformation, most Christians took these words at more or less face
value—the Eucharist is Christ’s body and blood, which is joined
with the whole of Christ, and so in the Eucharist Christ nourishes his
followers with himself.

The Eucharist lies at the center of a vast array of deep,

philosophically charged theological questions. The Eucharistic event
appears triggered by certain physical actions by the priest, on the
Catholic and Orthodox views, or by the congregation and its
presider on some Protestant views. Prior to these actions what is
present is just bread and wine, devoid of intrinsic spiritual
signi cance, but afterwards there is ‘the body and blood of Christ’.
Are the actions of the human agents given a supernatural power of
producing such an e ect or does God produce the e ect entirely by
himself on the occasion of these actions? Likewise, is the reception
of the Eucharist a cause of the occurrence of grace, or does God
simply happen to choose to provide grace on the occasion of the
receiving of the Eucharist? And if it is a means, then in what way
does this causality work?

What does it mean to ‘eat’ and ‘drink’ in general and what

signi cance is to be found in the idea of Christ giving himself to us
to be eaten and drunk? How does the Eucharist cause both physical
and spiritual nourishment, and potentially even drunkenness? Do
Christ’s body and blood become a part of the physical body of the
believer, do they revert to ordinary bread and wine just prior to
being digested, do they cease to be present, or are Christ’s body and
blood transubstantiated again, this time into the esh and blood of
the recipient?

When Christ’s words are spoken in the Eucharistic liturgy, who

counts as their speaker? To whom, if to anyone, does ‘my’ refer in
‘This is my body’? 1 Is it an indexical? How does the apparently
demonstrative ‘this’ gain reference to the invisible reality here?

In the Catholic tradition, the Eucharist is seen as a sacri ce,

ful lling the prophecy of Malachi that in messianic times a sacri ce
will be o ered from the rising to the setting of the sun. Yet
according to the Letter to the Hebrews, Christ’s sacri ce is the only
sacri ce in messianic times. Catholic theology attempts to reconcile
these two claims by saying that the sacri ce of the altar and the
sacri ce of Calvary are one and the same sacri ce. What, then, are
the identity and individuation conditions for sacri ces? Is there on a
deep level a single act of self-giving that Christ undertook, and if so,
how is it related to the events of the altar and those of Calvary? Are
they perhaps manifestations of that act? Are they parts of it?
Catholic devotion talks of being present at Mass as a way of being
present at Calvary. Can this be literally true, space-time being
bridged in a supernatural way? Or does the Eucharistic liturgy
simply represent Calvary, and if so, what philosophical account can
be given of the nature of this representing—is it conventional or in
some way natural, for instance?

Next come the ontological issues surrounding the question: What

actually happens that makes it true to say that ‘the body and blood
of Christ’ comes to be present?

The ontologically simplest answers are ones that take this

presence to be non-literal. Thus, one might simply stay on a
naturalistic level and say that Christ’s body and blood are ‘present’
in the congregation’s thoughts, and are represented by the bread
and wine. Or one might say that at communion, God gives the
recipient of the Eucharist graces that ultimately ow from the
sacri ce of Christ’s body and blood on the cross, and so the body
and blood are ‘present’ through their e ects.

There are, however, serious theological di culties with these

two solutions. The most obvious is that, as far back as we can trace
it, Christians have generally taken it that the ‘presence’ is to be
understood in a more substantive way, and have made the Eucharist
a central part of their Christian worship, as is already seen in the
New Testament (see e.g. Acts 2: 42 and 1 Cor. 10). If one believes
that the Christian church is guided by the Holy Spirit, at least in the
central aspects of Christian life, this creates at least a strong
presumption in favor of a more substantive interpretation.

Furthermore, the New Testament overall has a strong emphasis

on the reality of Christ’s body and blood, in contrast to gnostics who
saw the esh as something unbecoming, and also contains Christ’s
promise to abide with Christians. It would be tting indeed for this
abiding also to be bodily in some way.

Thus one should take seriously the idea of Christ’s body and
blood being present in a non-metaphorical way, ‘really present’. The
doctrine of ‘real presence’ presents several questions. First, we may
wonder about the sense of ‘present’ here. While we have taken
‘presence’ as not metaphorical, there may still be multiple senses of
presence. Is Christ’s body and blood ‘spatially present’ in the same
sense in which the bricks of the church building are ‘spatially
present’? Or is there some other non-metaphorical way of being
present that is applicable? How can Christ’s body and blood be
simultaneously present in multiple, disconnected places? Is a part
here and a part there, or is the whole present in each place?

A parallel question concerns what happens to the bread and

wine. It certainly appears as if bread and wine are present after
consecration. Some take this appearance at face value, and insist
that not only is Christ’s body and blood present, but so are bread
and wine. This is ‘consubstantiation’. Others insist that the
appearance alone is present, and bread and wine are really absent.
This conjunction of the real presence of Christ’s body and blood and
the non-existence of bread and wine is, according to Pope Paul VI’s
1968 ‘Credo of the People of God’, at the core of the doctrine of
‘transubstantiation’.

If consubstantiation holds, we have two options. First, by

analogy with the incarnation, we could have ‘impanation’. Just as
one and the same person is both a human and God, one and the
same entity is both bread and body, and likewise for wine and
blood, or maybe one and the same entity is both Christ and bread,
as well as both Christ and wine. Or one might have co-presence, in
which case bread and body are in the same place, and wine and
blood are in the same place. The co-presence version is subject to
the objection that ‘this’ in ‘This is my body’ would seem to more
appropriately apply to the visible of the two substances, namely
bread, if there were two substances there.

If, on the other hand, transubstantiation holds, we have several

further questions. Is there a real connection between the bread and
wine and the body and blood, with, say, the bread and wine literally
becoming transformed, or do bread and wine simply cease to exist,
being followed by the coming-present of the body and blood? What
makes it be the case that bread and wine are present? Is an illusion
miraculously caused in the minds of the people present? Or is it that
the causal powers of the bread and wine are somehow sustained, so
that light bounces o just as it did before? If so, what are these
causal powers grounded in? Are they now the causal powers of
Christ’s body and blood? Are they the causal powers of God? Are
they self-standing causal powers, present in the same place as the
body and blood? Or had bread and wine received a power of
a ecting future events at a time at which they no longer exist?

To discuss even half these questions would take a book. For

present purposes, then, I shall focus simply on the question of
whether the doctrine of the real presence of Christ’s body and blood,
and likewise the doctrine of the real absence of bread and wine, can
be defended philosophically. I shall argue for an a rmative answer,
and I shall do so by considering a variety of metaphysical models,
including that of Aquinas. It will appear, thus, that
transubstantiation is a philosophical possibility. If it is possible for
two substances to be in the same place at the same time,
consubstantiation will be a philosophical possibility as well. Of
course, the question of actuality is a theological one.

For brevity, I will usually speak of the body rather than the
blood. There is here yet another question that I shall not address,
which is whether the presence of the body brings along with it the
presence of the blood and vice versa. The Catholic tradition answers
this a rmatively, and in the 13th Session of the Council of Trent
adds that the soul and divinity are present along with the body. On
Aquinas’s view, the blood, soul, and divinity are present on account
of their union with the body which is eucharistically present where
the bread used to be (Summa Theologiae III q. 76 a. 1).

2. REAL PRESENCE

2.1. The Problem

According to the doctrine of the real presence, Christ is present

wherever the Eucharist is validly2 celebrated. By any plausible
criteria of validity, the Eucharist is validly celebrated at the same
time in multiple places. Therefore, the doctrine of the real presence
seems to imply that Christ is present in multiple places at once.
Moreover, according to orthodox Christianity, Christ currently exists
bodily in heaven. If a body can exist only in a place, and no earthly
place is also a heavenly place, then we can conclude that Christ is
present in multiple places at one time from his being present on
earth in the Eucharist and in heaven.

Assuming a body is a material object, it seems that the doctrine

of real presence leads to:

(M) A material object can be simultaneously wholly located in

more than one place.
This multilocationist thesis will need to be examined. But before we
do that, we should examine strategies for holding on to the claim
that Christ’s body is really present in the Eucharist without being
committed to (M).
2.2. First Anti-Multilocation Strategy: Paptial Presence

Leibniz, while not himself a Catholic, for ecumenical purposes

and/or to gain a wider following for his views, attempted to show
that his metaphysics was compatible with the Catholic doctrine of
transubstantiation. To that end, in his correspondence with Des
Bosses (Leibniz 1989: 197 .), Leibniz claimed that his monadic
metaphysics was compatible with the idea—to which he was not
committed—that a material substance required the existence of a
‘substantial chain’. The basic item in Leibniz’s ontology is the
monad, which is a substance that has no parts. The soul of a human
being is a monad, and all material reality is in some sense built up
out of monads as well.

Leibniz then suggests extending this ontology by introducing

substantial chains. These are entities of an ontological type di erent
from a monad, that tie together multiple monads. More precisely, to
each substantial chain there corresponds a collection of monads that
are its members, i.e. which stand in a special member-of relation to
the chain. Chains can lose and acquire members, and monads can
come to be and cease to be the members of chains. Thus chains can
survive replacement of their parts, in the way in which we think
organic bodies can. A material substance, then, is at any given time
constituted by both a substantial chain and its member monads. If a
and b are material substances, at the same or di erent times, then
they are identical if and only if they have the same chain. A good
way to think of the chain is as a Leibnizian analogue to a scholastic
substantial form, but perhaps without the form’s causal e cacy.

Leibniz’s proposal, then, was to say that in the Eucharist the

monads of the bread and wine become members of the chain that
constitutes Christ’s body. Thus, Christ’s body grows to include the
monads that were previously the monads of the bread and wine.
This metaphysical event, we may note, has no observable
consequences in Leibniz’s thought, because the relations between
chains and their monads are not observable.

Leibniz’s account shows a way to understand real presence that

blocks the argument from real presence to (M) by a rming that
Christ’s body is present in more than one place but denying that it is
wholly located in more than one place. Instead, a part of Christ’s
body is in heaven and another part of his body is on earth. Because
the monads of the bread that is consecrated in di erent places are
distinct, one part is present at one Eucharistic liturgy and another at
another.

We need not accept the full monadic system to see the

attractiveness of Leibniz’s account. The account requires only a
primitive metaphysical relation that constitutes something as a part
of a real whole. This might be a relation to other parts, or to a form,
or to some central and irreplaceable part. What happens at
consecration, then, is that the parts of the bread acquire such
relations as make them be parts of the body of Christ.
Does this account do justice to real presence, or does real presence
require the presence of all of Christ’s body? Plausibly, it is not
necessary for all of a person to be in a room for the person to count
as really present in the room. If a witness to a marriage were
required to be in the church in which the marriage happened, and
due to the crowd the witness stood in the church near the doorway,
and as it happened had stuck his foot outside the doorway during
the exchange of the vows, we would not, I think, deny that he was
present in church for the vows. In fact, if due to the crowd he was
able to have only his head in the church, I think we would say that
he was present.

On the other hand, if the part of the witness present in the

church were the tip of his little nger, we would be rather less
con dent that he was in the church. It seems that we count someone
as present in a place provided that a signi cant portion of her is
there, such as a head. What portion is signi cant may be contextual
—for the presence of a witness in a place it may be the part bearing
the eyes and ears, or perhaps the eyes or ears themselves, while for
the presence of a runner receiving an award on a podium a foot
might be enough.

But if the parts of the bread become parts of Christ’s body, they
do not seem to become organically important parts. As far as we can
tell, they do not engage in any organic interaction with Christ’s
heavenly body. Thus on the Leibnizian account, what is present in
the Eucharist seems to be an insigni cant portion of Christ’s body. If
there is no organic interaction there, the part present seems to be
more like a hair than a heart. Moreover, it is a part that Christ could
easily do without, since moments earlier he did not have it, and
nonetheless had a fully complete, indeed glori ed, body.

Thus this kind of partial presence would not only not be doing
justice to the idea of a real presence, but would also not match the
devotional tradition’s focus on Christ’s self-giving to us in the
Eucharist—here one thinks of the common image in Christian art of
the pelican legendarily feeding its young with blood from her breast
—since his giving us his hair to eat would not seem to have the kind
of depth of self-giving that is ascribed to the Eucharist. Moreover, it
would appear that on the Leibnizian view, Christ should have said
‘This is a part of my body’ rather than ‘This is my body.’

A di erent partial presence approach that avoids multilocation

would be that di erent parts of Christ’s heavenly body come to be
present instead at di erent Eucharistic liturgies, still connected to
the whole of the body of Christ via the relations that constitute
them as parts thereof, and these parts are then not present in
heaven. Thus, the more consecrated hosts there are, the less of
Christ there is in heaven, unless the parts left behind in heaven
grow to compensate.

This would lead to the odd question of which part of Christ

comes to be present at which liturgy, di erent parts being present at
di erent ones. Perhaps more worrisomely, a view on which the
parts of the heavenly body of Christ left heaven and came back to
earth would not seem to do justice to the triumph of the
resurrection, through which Christ received a complete glori ed
body, and reigns in that body in heaven.

A further absurdity that results from any partial presence view is

that although St Paul warns about gorging at eucharistic liturgies (1
Cor. 11: 21–2), on these views one really does receive more of
Christ the more one eats and drinks at the liturgy, and so Paul’s
warning would end up being a warning not to receive too much of
Christ.

The existence of partial presence views nonetheless shows that

the argument from real presence to (M) requires that the doctrine of
real presence be taken as saying that Christ’s body is wholly located
where the eucharistic bread used to be.

2.3. Second Anti-Multilocation Strategy: Placeless Presence

A common phrase among Catholic theologians is that Christ is

present ‘sacramentally’ in the Eucharist. For this to block the
argument for (M), we need the claim that a sacramental presence is
not only not the same as presence in a place, but does not even
entail it. Someone who thinks that sacramental presence is just a
special case of standard spatial presence when this happens to occur
in the context of a sacrament can agree that Christ is present
‘sacramentally’, but cannot use ‘sacramental presence’ to escape the
inference to (M). Nonetheless, the predominant view among
Catholic theologians, including St Thomas Aquinas, has apparently
been that sacramental presence does not entail ‘local presence’, i.e.
presence as in a place. This view is not magisterially binding on
Catholics, but it is important and must be examined.

Let us, then, consider the family of ‘placeless presence’ views of

the real presence—views on which Christ’s body in the Eucharist,
while real, is not locally present as in a place and the inference of
(M) is blocked. On a placeless presence view, Christ’s body is really
present in the Eucharist on account of standing in a presence
relation R to some anchoring entity y. It is essential to the account
that the anchoring entity y either be the place of the Eucharist or be
unproblematically present there, and then it is on account of the R-
relatedness to y that Christ counts as really present. The anchoring
entity y might be the accidents of the bread, or the congregation, or
even a place in space, depending on the details of the view.

We require, then, an explanation of the presence relation R and

of the anchoring entity y which makes it clear why it is that x’s
standing in relation R to y is a real presence of x in or at y. I shall
say that a relation R such that x’s standing in R to y is a real
presence of x in or at y is a ‘real presence relation’.

Examples of relations that are not real presence relations include

‘symbolic presence’ where x is symbolically present at y provided
that a symbol of x is present at y as well as ‘mental presence’ where
x is mentally present at y provided that x is thinking about being at
y. In fact, we can see that these cannot be real presence relations
because (a) an unreal entity can be symbolically present in a real
place, Sherlock Holmes being these days symbolically present, by
virtue of a statue, outside the Baker Street underground station, and
(b) a real person can be mentally present in an unreal place, C. S.
Lewis arguably having been mentally present in Narnia.

Say that a relation R is ‘weakly locational’ provided that x’s

standing in R to y entails that either y is not anywhere or x is
spatially present where y is. Here, a strongly locational relation
would be one where the rst disjunct can always be dropped. An
example of a weakly but not strongly locational relation is the
relation of a trope to the entity having the trope: when the entity is
spatial, we can say that its tropes are where it is, but when the
entity is non-spatial, this would make no sense.

The defender of a placeless presence view must nd a real

presence relation R that is not even weakly locational. For if the
relation in question is weakly locational, then it seems we can
derive (M) just as well as in the case of standard local presence. For
the anchoring entity y had better be where the Eucharist takes
place, or else we cannot say that Christ’s body is really present in
the Eucharist, and so Christ’s standing in R to the anchoring entity
would entail Christ’s being locally present where the Eucharist is
being celebrated if R were weakly relational.

It is not immediately clear that there are any real presence

relations that are not weakly locational. To see the di culty in
nding such a relation, consider Aquinas’s discussion of God’s
omnipresence. Unlike Hud Hudson in this volume, rather than
ascribing local presence to God everywhere, Aquinas says: ‘He is in
all things by His power insofar as all things are subject to His
power. He is in all things by His presence insofar as all things are
bare and open to His eyes. He is in all things by His essence insofar
as He is the cause of being for all things, as has been explained’
(Summa Theologiae I q. 8 a. 3, trans. Freddoso). If we take
omnipresence in each of Aquinas’s three senses to be a real presence
everywhere, then we now seem to have three examples of real
presence relations that are not weakly local.

However, what counts as a real presence in the case of one

entity, say God, might not count as a real presence in the case of
another entity. According to a causal interpretation of Newtonian
physics, distant astronomical objects are capable of instantaneously
moving other objects. The sun would thus be present in the earth by
(gravitational) power, if all it means to be present by power in y is
that one can or does a ect y. But this way of speaking seems
metaphorical: we generally could not say that the sun is really
present in the earth without being quite misleading, even if
Newtonian physics happened to be true. Likewise, no matter how
much I observe a room from the outside, I am not really present in
it. Thus it seems that either being present by power and observation
are simply not real presence relations, or they are real presence
relations in the case of God but not in the case of a material object
such as a star, a living human (who may be more than material but
is also material), or a body.
 The case of presence by ‘essence’ requires some more discussion.
We could simply take Aquinas to be saying that when x causes y to
exist, then x is present by essence at y. If so, then this will generally
not be a real presence when x is an object in space. There is a sense
in which Michelangelo is present in the David, but he is not really
present. But Aquinas appears to be referring not just to God’s
causing entities to come-to-be, but God’s continual creation of
entities—his e caciously willing them to continue in existence.
Creatures are more intimately related to God qua creator than the
David is to Michelangelo. The continual dependence seems crucial
here. But if so, then this form of presence would simply be a more
hallowed case of being present somewhere by continuing immediate
action. However, that will in general not be a real presence in the
case of a material object. The sun continually acts on the earth
gravitationally, and even if this action were immediate as in
Newtonian physics, it would not make the sun be really present in
the earth. Hence this option does not help, either.

The task of nding a relation that is not weakly locational but

nonetheless a real presence relation in the case of a material object
like a body is prima facie di cult. Moreover, there is some prima
facie reason to suspect that no such relation can be found. For while
we may be willing to admit with Aquinas that God (or an angel) is
present where he is acting, and we may even take this to be a real
presence, the reason we might take it to be a real presence is that
there is no way for a being that is in no way material to be locally
present anywhere, and presence by action is a natural analogue for
such a being. But in the case of a material object, which is capable
of being spatially present somewhere, it seems likely that we should
say that it is really present in any location or in any located thing
only if it is there spatially. That seems to be the lesson of looking at
the three modes of omnipresence above.

However, this argument is based on a defeasible expectation that

no relation that is not weakly locational will be a real presence
relation. Aquinas appears to provide a counterexample to this
expectation (Summa Theologiae III q. 76 a. 5). According to Aquinas,
ordinary material substances have dimensions that they occupy. The
dimensions of an object consist in the inner geometric con guration
of the parts of the object, and can be thought of as basically the size
and shape, or extension (in the Cartesian sense), of the object. We
can intuitively visualize the dimensions of an object as the object’s
three-dimensional lled-in outline, which changes as the object
changes its shape and/or size, but remains unchanged as the object
moves rigidly through a at space. But we need to be careful with
this visual image, because we cannot say that the dimensions are a
material object—they are, in the ordinary case, just an accident of
some substance. For a substance to have such-and-such dimensions
is for it to be shaped and sized thus and so.

Ordinarily, a substance, say a frog, has its dimensions, i.e. the

dimensions are an accident (roughly, a non-essential trope) of it. In
virtue of the substance’s having the dimensions, we can say that the
substance is in these dimensions. What makes it be the case that a
material substance is found in a place P is that the substance is in
the dimensions D, and D is in turn located in P. There are ordinarily
thus two steps in locating a substance: Find its extension, and then
see where that extension lies. If we think of the frog’s lled-in
outline, which is the shape and size of the frog and moves with the
frog, then it makes sense to say that the frog is in its lled-in
outline, and that lled-in outline is in a place.

There are, then, two things that the substance is really present
in: it is directly present within the dimensions D, and thereby
mediately present in P. The substance is then said by St Thomas to
be ‘substantially’ present in D and ‘locally’ present in P.

We can nd plausible the claim that the presence of a substance

in a place is mediated by its dimensions if we accept that the size
and shape of an entity are intrinsic properties of the entity (we shall
see in sect. 2.4 that if we deny this claim, we get another account of
real presence). An intrinsic property is one such that an entity’s
having that property is not a matter of anything beyond the entity,
but simply of how the entity itself is. If the dimensions are intrinsic,
plausibly they are not dependent on location. On the other hand,
what counts as the location of an entity depends on the entity’s
shape and size—if I grew to gigantesque proportions I would no
longer be wholly in this room, for instance. This makes it plausible
that the location depends on the dimensions. Moreover, to locate
something in a place we need only relate its lled-in outline to the
place in such a way that the relation makes it true that the object’s
lled-in outline is there, i.e. we need to locate the object’s
dimensions in the place. Hence, it is plausible that it is the
dimensions that are directly locally present.

Aquinas’s move in the case of the Eucharist, then, is to insist that

while the substance of bread has turned into the body of Christ, the
accidents of bread—including most signi cantly its dimensions—
have remained in existence when there is no longer any subject that
they are the accidents of, the body of Christ not having these
accidents. (It is, of course, ontologically quite surprising that an
accident could survive the perishing of its object—this is an issue
that will be discussed in sect. 3—but for now let us grant this
possibility.)

What we then have are the dimensions of bread that are directly
present on the altar as in a place, and then the body of Christ is
substantially present in these dimensions, in the way in which the
substance of bread had been present in them. The body of Christ,
however, is not present locally on the altar, whether directly or
indirectly. It is not present directly locally, because material
substances are only directly present in their dimensions—it is only
through the dimensions’ presence that they are present anywhere.
And the body of Christ is not indirectly present locally, because to
be indirectly present locally in P, a substance must have dimensions
that are directly locally present in P. But the dimensions in question
are not ones that Christ’s body has, though it is contained in them
‘substantially’.

To summarize, Aquinas is rst analyzing the ordinary notion of

the presence of a substance in a place through two presence
relations: a substance’s being substantially present in dimensions
and these dimensions being present in space. He then claims that
one way for a substance to be present in its dimensions is just for
these dimensions to be its dimensions. But, he claims, that this is
just a special case of the more general relation of ‘substantial
presence’, and the body of Christ stands in that more general
relation to the dimensions that the bread had, without actually
having these dimensions.

The presence of Christ’s body in the dimensions is real, since it is

supposed to be the same kind of presence that an ordinary material
substance has in its dimensions, and that had better count as a ‘real
presence’ if it is to mediate the ordinary substance’s real, local
presence in a place. The substantial presence of the body in the
dimensions, thus, would count as a real presence. The account also
has the merit of explaining why it is that when the eucharistic host
is moved, Christ’s body continues to be present there. If it were
simply an invisible presence directly in some location in space, it
would take an additional miracle for the minister’s hand to relocate
it. But the body is not relocated when the host is moved, because it
is not present in a place, but present in the dimensions, which
continues to be the case when these dimensions are relocated by the
minister’s hand. Moreover, the account ts with the Catholic and
Orthodox tradition that when the accidents of bread cease to exist,
e.g. due to rot or digestion, Christ’s body ceases to be present.

There are, nonetheless, at least two di culties with Aquinas’s

account.

1. The idea of accidents existing without a substance is

problematic. This will be discussed in sect. 3.

2. Can something be substantially present within dimensions that

are not its own dimensions, and in a way that does not give rise to a
local presence? Consider a putative case of being present within the
dimensions of something else: I am present within the dimensions of
the room. But this is not a substantial presence. I am present within
the dimensions of the room by virtue of my own dimensions and the
spatial connection between them and those of the room. But
substantial presence is not mediated in this way. Moreover, my
presence in the dimensions of the room entails a local presence.

Consider now this argument: ‘substantial presence’ either is or is

not meant in the same sense when we say that x is substantially
present in the dimensions it has as when we say that x is
substantially present in the dimensions that are not its dimensions.
If the two are not meant in the same sense, then it does not seem we
have good reason to agree that substantial presence is a real
presence relation. For while we agree that being substantially
present in one’s own dimensions is an instance of real presence—it
has to be since it mediates local presence for substance—we can no
longer conclude from this that the second kind of substantial
presence, where something is present within dimensions that are not
its own, is a real presence relation. Moreover, we no longer have a
grasp on this second kind of substantial presence. We might be able
to do something if there is an analogy between the two kinds of
substantial presence, but it would be an analogy di cult, I suspect,
to setup.

Suppose, on the other hand, that ‘substantial presence’ is meant

in the same sense in the two cases. Then it seems that substantial
presence is a strongly locational presence relation. For it seems that
what makes it be true that an ordinary substance is found in a place
is the pair of facts that it is in its dimensions and that these
dimensions are in the place. But the same would be true, and in the
same sense, even if these were not its dimensions, as long as it was
substantially present in them, and hence it seems that the same
inference about being found in a place could be made.

But perhaps what makes it be true that an ordinary substance is

found in a place is not some more general relation of ‘substantial
presence’ in virtue of which it is in its dimensions which are in that
place, but more simply that it has the dimensions which are in that
place, whereas Christ’s body stands in the more general relation of
substantial presence to the dimensions of the bread but does not
have these dimensions. This is what Aquinas would say. But at this
point, I fear, it is quite unclear what it means for a substance to be
‘in’ its dimensions in the mode of ‘substantial presence’. We
unproblematically have a concept of ordinary substances having
dimensions. But if their being in the dimensions is anything other
than their having these dimensions, it seems very di cult to
understand what their being in these dimensions consists in.

These di culties are not knock-down arguments. But what they

suggest is that Aquinas might be better o allowing that Christ’s
body is locally present on the altar, in virtue of being substantially
present in the dimensions that are locally present on the altar. This
would seem to force an acceptance of (M). However, this acceptance
should not be problematic to someone friendly to Aquinas’s view,
since it does not seem to be any more di cult to be in more than
one place than to be in more than one set of dimensions. Such a
modi ed view that denies (M) would be, it seems, fully consonant
with the Catholic tradition of transubstantiation.

Furthermore one might think that shape and size are not intrinsic
properties of an entity, but instead depend on the location of the
parts, with location being a relational property (this will be
discussed further in sect. 2.5). If one takes this view, then one will
not see locatedness as mediated by being within dimensions, and
Aquinas’s view will not be attractive. However, one can still take
one of the views of the next two sections.

2.4. Third Strategy: Curved Space

The argument from real presence to (M) makes another assumption

that has not yet been noted. It assumes that if Christ is wholly
located on one altar and wholly located on another altar or in
heaven, then Christ is wholly located in two places. For validity, this
inference requires the claim that the same place cannot be on two
altars or on an earthly altar and in heaven at the same time.

This claim seems quite plausible, but is not clearly true. Observe
rst that a person might be wholly located in two buildings at once.
Thus, the Royal British Columbia Museum contains a wooden house
ceremonially belonging to the descendants of Chief
Kwakwabalasami, and someone within that house would be in two
buildings at once. Moreover, then the place where the person would
be would itself be in two houses. But if a place could be in two
buildings at once, it seems plausible that a place could be on two
altars at once. Of course that would require a special spatial
relationship between the two altars. Imagine, for instance, a large
altar whose surface seen from above was a wooden rectangle
surrounded by a larger stone surface. It might turn out that the
wooden rectangle had its own independent legs, and was originally
consecrated as a separate altar, and then became a part of a larger
altar without losing its initial consecration. Then something lying on
top of the wooden rectangle would by lying on two altars at once—
the larger one of stone and wood, and the smaller one just of wood.

This response may seem ippant—analytic (or, I think

equivalently, scholastic) philosophy at its worst. But actually there
is a serious point here. To make the argument for (M) go through
we need substantial claims about the structure and arrangement of
altars and heaven.

For consider one plausible way of making the argument go. We

might note that the eucharistic liturgy can be simultaneously
celebrated in two non-overlapping buildings, and conclude that
therefore Christ’s body will be present in two non-overlapping
buildings, and hence in two places. However, there is an ambiguity
in ‘non-overlapping buildings’. In one sense, two buildings are non-
overlapping provided the walls do not overlap. This cannot be the
sense meant, because a common place might be found in two
buildings that are not overlapping in this sense, if, say, one building
is entirely within the other, and so the argument for (M) would be
invalid.

In a second sense, two buildings are non-overlapping provided

the spaces within them are non-overlapping. Then we cannot fault
the inference that if Christ’s body is present wholly in one building
as well as wholly in the other, then it is wholly in each of two non-
overlapping places. But we now need a second argument to show
that the liturgy can be celebrated in two buildings that have no
common place. On the face of it this seems obvious empirically:
surely if the liturgy is celebrated at St Peter’s in Rome and in the
National Shrine in Washington, it is celebrated in two buildings that
have no common place.

But how would we argue that St Peter’s and the National Shrine
have no common place? The best way I can see—and one whose
failure will be illustrative of the failures of other ways—would be to
note that a straight line segment can be drawn between any pair of
points one of which is wholly in St Peter’s and the other wholly in
the National Shrine, and the line segment will have positive length.
This argument, however, presupposes the premise that when a line
segment of non-zero length can be drawn between point A in space
and point B in space, then A and B are distinct. This premise is
empirical, and we do not at present actually know if it is in general
true. For all we know, the space of the universe has a geometry such
that if you go far enough in any one direction you will come back to
where you started.

Of course, all I am saying continues to look as though it is in the

worst tradition of analytic nit-picking. After all, granted, it may be
that if you go billions of light years in a straight line you will come
back to where you started, but we know the topology near the earth
is roughly euclidean, and going a mere several thousand miles will
surely not get you back to where you started.

But remember that we are considering a situation that is,

professedly, a miracle. A miracle can involve violations of the laws
of physics, though not of the laws of logic. The topology of space, as
long as it is logically consistent, is a matter of physics—this is the
foundation of relativity theory, at least when generalized to space-
time. Thus unless one can show that there is no logically consistent
topology that could allow the same place to be found on an altar in
Italy and on an altar in the United States, the argument that real
presence implies (M) has a serious gap.

And in fact it does seem quite possible to give such a topology by

using a mathematical construction known as a quotient space. A
quotient space construction ‘identi es’ together points of the
original space to form a new space. For instance, as a topological
space, take the surface of a long rectangular strip running from left
to right. We might identify together corresponding points on the left
end of the strip and on the right end, identifying the top left point
with the top right point, and so on. The result of such identi cation
is a new topological space which is no longer a rectangle but a loop.
As you keep on going leftward, you come to, as it were, the points
on the left edge. But these got identi ed with points on the right
edge. And since those points had something to the left of them
(namely, points near the right edge of the original strip), you can
continue going leftward, and come around in a loop.

We can likewise identify together the points in the space

occupied by the preconsecration wafer in Italy and the points in the
space occupied by the preconsecration wafer in the United States,
and thus construct from our standard, non-miraculous, semi-
euclidean space a quotient space in which there is a place
surrounded by both Italy and the United States. We could more
easily visualize this if our space were two-dimensional instead of
three. Just bend space until the wafer-shaped place in Italy lines up
with the wafer-shaped place in the United States, and glue the two
places into one. It appears this can be made mathematical sense of.3
But then God could miraculously ensure that after consecration our
world’s space is just like the quotient space, and that matter
continues to behave as expected. A bit more work would be needed
to account for the four dimensionality of space-time, but there does
not appear to be any serious obstacle there.

The same quotient space move could be done with more than
one wafer, as well as with the wafers and Christ’s body in heaven.
Christ’s body would come to be present on the altar, then, in the
sense that the points in the space just around the eucharistic host
would come to be neighbors of points in heaven. It would become
literally true that a little piece of heaven is on earth. The only thing
one needs to do to form the quotient space is to have a one-to-one
correspondence between each set of points where a wafer is before
consecration and the points in Christ’s heavenly body.

Of course a number of miracles would be needed to keep particle

movements as observed. And one may think this story implausibly
di cult. But of course it makes little sense to say that one thing is
more di cult to an omnipotent God than another.

One might, however, have a general preference for theological

accounts that involve fewer invisible miracles. This preference
would have to take the form of the claim that if there are two
accounts of how some doctrinal claim could be true, then the
account involving fewer miracles should be chosen. If so, then this
account has plausibility if one with fewer miracles cannot get o the
ground. But whether there is another account or not, the existence
of this account shows that an argument against the real presence
from the denial of (M) cannot be made to work. For the real
presence indeed is important enough that God would bend space
itself to make heaven be present on earth if that were needed to
make it possible. Hence if none of the other solutions work, this one
is available. The Kingdom of Heaven would then be present among
us in a more literal sense than one might have initially thought
possible.

2.5. Bilocation

But suppose that Aquinas’s ‘substantial presence’ solution is held

indeed to involve a local presence, contrary to Aquinas’s words, that
no better sense of real but placeless presence can be found, and that
we do not want to allow space to have strange topologies. In that
case, the defender of real presence must accept (M). Is the
acceptance of (M) absurd? I shall argue in the negative.

There are, after all, coherently describable cases of bilocation. As

Koons (2005) points out, time travel is one such case—if one were
contemporaneous with one’s teenage self, one would be bilocated,
once as an adult and once as a teenager. Another case given by
Koons is Feynman’s supposition that all particles of the same type
are one, multilocated particle.

The primary di culty is with making sense of what it could

mean for something to be wholly located in more than one place.
We need to distinguish a case where x is wholly located in places P
and Q from a case where x is located in both P and Q simply
because x is big enough so that some of it is in P and some of it is in
Q (cf. the Leibnizian model of the real presence).

Readers familiar with the concept of an object’s being wholly

present in a region of space-time may want to observe that this
concept is to be distinguished from the concept of being wholly
located in a place P at a time t. The concept of being wholly present
has received signi cant attention in the recent literature (an
excellent survey is provided by Crisp and Smith 2005). The main
metaphysical use of the concept of being wholly present is to
distinguish between two views of how an object persists over time.
On the endurantist or three-dimensionalist view, all of an ordinary
material object is present at every time at which the object exists,
and hence ordinary material objects are three-dimensional. On the
perdurantist view, objects are four-dimensional—they extend
through space and time. At any given time t at which the object
exists, it has a special part known as a stage, which exists only at t,
and is a complete three-dimensional cross-section through the four-
dimensional object. Only that three-dimensional cross-section of the
ordinary object exists at t. Thus, the endurantist holds that ordinary
material objects are wholly present at every time at which they exist,
while the perdurantist denies this. The concept of an object x’s
being wholly located in a place P at a time t, however, is meant to be
neutral between endurantism and perdurantism, and is compatible
with the idea that x might have parts existing at other times which
are not present at time t. Being wholly located in a place at a time is
meant simply to capture the intuitive notion of all of the object at
that time being there. If endurantism is true, then to be wholly
located and wholly present somewhere at a time is one and the
same thing. If perdurantism is true, then it may be that being wholly
located in a place at a time is the same as for the relevant stage to
be wholly present in that place.

It is tempting to de ne something as wholly located in a place P

at a time t provided that some of the parts of the object at t are in P
and none of them are outside P. If this were true, (M) could not
hold. However, one can reasonably object to such a de nition by
insisting that the question of whether x is wholly in P should depend
only on x and on what happens in P, not on what does or does not
happen outside P, e.g. whether any of the parts of x are found there
as well.

An obvious de nition that does justice to this intrinsicness

requirement (cf. Koons 2005) is:

(W1) An entity is wholly located in P if and only if either it is

simple—i.e. without any parts—and present in P or each of
its parts is present in P.

Since the whole is where all the parts are, we can then say that the
whole is in P.
 A di culty with this de nition is an ambiguity in ‘all the parts’.
If we mean all the parts that exist at the given time t, then
multilocation is only possible when the object has the same parts in
all one’s locations. In particular, the kind of multilocation that is
involved in time travel will be ruled out, because the adult time
traveler who visits her teenage self has cells that the teenager does
not and vice versa. If, on the other hand, we mean all the parts that
are found in P, then it follows that we are trivially wholly located
wherever any of our parts is present, so that being wholly located
reduces to being present, which seems implausible.

There is a large literature attempting to de ne the concept of

being wholly present in a way that avoids analogues of these
di culties, and if one wanted to come up with a de nition of being
wholly located one could draw on that literature. However, there is
no need for that. For even though (W1) is perhaps not a good
de nition of being wholly located in P, it yields a perfectly ne
su cient condition:

(W2) An entity is wholly located in P at t if either it is simple—

i.e. without any parts—and present in P at t or each of the
parts it has at t is present in P at t.

This is all we need in the case of the Eucharist because there the
possibility that an entity may have more parts at one of its multiple
locations than at another may not come up. Rather, we can simply
say that each of the parts that Christ’s body now has is found in
each of the multiple locations of the Eucharist.
 A more serious di culty with bilocation is the possibility of
di erences in intrinsic properties. If I am present in more than one
place at a time, I might be red all over as found in one place and
green all over as found in another, which would imply that I am
simultaneously red all over and green all over. One response,
defended by Koons 2005, is that the problem here is no bigger than
the problem of intrinsic change for an eternalist, i.e. for someone
who believes that the past and future are equally as real as the
present. The problem of change, raised by Lewis 1986: 204 is
generated by the fact that an entity might have contradictory
properties at di erent times, say, being round at one time and
square at another.

There are two familiar positions on the problem of change. A

perdurantist believes that I am present at a time in virtue of a
temporal slice of me being present there. In the bilocation case, I
might suppose that something like an organically complete slice is
present in one location and another slice is present in the other. In
the case of the Eucharist, though, this would lead to a possibly
unsatisfactory variant on the partial presence view of Leibniz: for
one slice of Christ’s body is present in heaven, then, and another on
the altar.

An endurantist, on the other hand, believes I am wholly present

at each time at which I exist, and thus he is already accepting that I
am wholly present in disjoint regions of space-time. But if so, then
why can’t two of these locations in which I am wholly present
happen to be simultaneous, in which case I would be wholly located
in two disjoint regions of space? As Koons points out, anything the
endurantist eternalist might say about the problem of change can be
said here.

The problem of change for the endurantist is, basically, that an

object that is wholly present at each of t8 and t9 might be both
positively charged all over and negatively charged all over, at t8 and
t9 respectively, which seems absurd. However, an endurantist might
resolve the problem by indexing properties to times: e.g. being
positively charged at t8, and being negatively charged at t9. Well, for
bilocation one simply indexes them not to times but to times and
positions: being positively charged at t8 as located in P and being
negatively charged at t9 as located in Q. Or the endurantist might say
that what seemed to be intrinsic, non-relational properties such as
being electrically charged in fact do include relations to points in
time. Thus, the object might stand in a positive-charge-at relation to
t8 but in a negative-charge-at relation to t9. Again, the bilocationist
can likewise make properties be relational to positions and points in
time: an object might stand in a positive-charge-at- …-and-as-located-
in relation to t8, P but in a negative-charge-at-…-and-as-located-in
relation to <t9, Q>.

(And when we realize that, given General Relativity, times are

just maximal spacelike hypersurfaces, it might well seem not to be
any more problematic to index properties to subsets of these
hypersurfaces (a subset of a spacelike hypersurface being an
encoding of both a place and a time) or to make them relative to
such subsets, than it is to index or relate them to whole
hypersurfaces.)

But actually it seems plausible that these complications are not

needed. These complications were introduced to take care of the
problem of intrinsic properties and multilocation. But actually,
according to the standard view of transubstantiation, it is not at all
clear that we need to allow for the body of Christ to have di erent
intrinsic-like properties as located on the altar and as located in
heaven. It is tempting to say that the body of Christ has, say,
whiteness and roundness as found on the altar but not as found in
heaven. But on the standard view, these are not properties of
Christ’s really present body. It is not Christ’s body that looks white
and is round—there is just an appearance (the Latin term used by
the Council of Trent is species) of whiteness and roundness, but this
is not an appearance of the body of Christ. In general, then, it may
well be that Christ’s body does not directly acquire new intrinsic
properties by virtue of consecration (there may be indirect
acquisition, e.g. because Christ’s human brain might be a ected by
his divine knowledge of the fact of consecration). It simply comes to
be additionally present in a new place. If this is right, then the
problem of di erent intrinsic properties at di erent locations of a
multilocated entity simply does not arise in the case of the
Eucharist.
 If shape and size, however, count as intrinsic properties, then
this solution would be di cult. For Christ’s heavenly body is surely
larger than a wafer. Moreover, Christ is wholly located in the
Eucharist—he does not, for instance, invisibly stick out outside the
host. It seems to follow that if Christ is spatially present, he is
present as round and circular. Aquinas, however, would block this
inference by saying that Christ is present in the dimensions of the
bread, but that he does not have these dimensions, and in fact,
nothing has them, because the one thing that had them—the
substance of the bread—is no longer in existence. How successful
this would be would depend on whether one can make sense of
being present in dimensions that one does not have.

But in fact shape and size are not intrinsic properties. We learn
from Einstein that shape and size depend on reference frame, for
instance. We might also argue that the shape and size of an object
supervenes on the positions of its parts. If so, then we might take
the shape and size as constituted by facts about the positions of the
parts. But the position of an entity may well be relational, consisting
in the relation between the entity and a point in space on an
absolutist view of space, or the relation between the entity and
other entities on a relational view. (For other arguments for the
non-intrinsicness of shape, see Skow 2007.)

Consider the following quick argument against the shape and

size of an object being intrinsic properties. Consider an object that is
a large circular ring, one light day in radius (i.e. it takes light one
day to travel from the center to the edge). The diameter of the ring,
then, is two light days: this is the distance between two opposite
edges. But suppose that God creates ex nihilo a massive star right in
the middle of the ring, and consider the situation ve minutes later.
The ring is not intrinsically changed by the creation of the star in its
middle, because physical e ects, including gravitational ones,
propagate at most at the speed of light, and it would take about a
day for light to travel from the star to the ring. About a day later,
the ring will be intrinsically changed, but it is as yet unchanged.
However, the ring’s size has changed. For the insertion of the star
has warped space-time near the star, and so the distance between
two opposite edges of the ring is no longer what it used to be.
Therefore, the size of the ring is not an intrinsic feature of it. It
depends on the structure of space-time beyond the space-time
occupied by the matter of the ring itself. Likewise, if the star pops
into existence o -center, the ring will no longer be circular in the
classical sense of a circle as a set of points equidistant to a center. It
appears that size and shape are not intrinsic.4

One di culty here is that there seem to be intrinsic properties of

an organic entity that directly depend on its shape and size; but how
can an intrinsic property directly depend on an extrinsic one? The
heart would not work if it were as thin as a wafer, for instance.
There is a connection between the spatial con guration of parts and
their causal interconnection. However, this is not a problem given
divine omnipotence if the dependence is nomic and not logical. For
it may be that the causal interconnections between the multilocated
parts of Christ’s body depend only on the spatial interconnections
that the parts have as found in heaven, and not as found in the
Eucharist.

It may also be the case that we can distinguish an internal shape

and size of an object, de ned by spatial relations between parts of
the object where these relations in an appropriate sense ‘do not
leave’ the object (in the way in which the relations between the
opposite parts of the ring ‘left’ the ring by entering into the space in
the middle) and the external shape of an object, de ned by spatial
relations between the parts of the object and things outside the
object (either space itself or other objects).

(See also McDaniel 2003.) If so, then it might be that the internal
shape and size is something intrinsic but the external one is not.
One might be suspicious of even the idea that the internal shape and
size is intrinsic (e.g. the arguments of Skow 2007 would still apply),
but even if we grant the intrinsicness, it is not clear that a problem
for multilocational accounts of real presence results.

For the main di culty is, presumably, that Christ is, let us
suppose, ve feet tall while the Eucharistic host is, let us suppose,
an inch in diameter, and there is a disparity of shape. However, we
can imagine a space where the inside of an area is in some sense
bigger than the outside, and whose internal geometry is di erent
from what one might expect from the outside. Imagine an in nite,
at two-dimensional space. Think of this as a rubber sheet. Draw a
small circle. Now keep the edges of the circle xed, but stretch the
rubber within the circle into a large bubble whose diameter is much
larger than that of the circle (for visualizing this, we stretch the
sheet into the third dimension; to avoid making use of an extra
dimension, we could simply modify the metric). The resulting space
is at, except for a circle that pinches o a large balloon. Then there
is a real sense in which the inside of the area is larger than the
outside: the outside is a small circle, but the diameter of the bubble
within is larger. Moreover, the geometry inside may be quite
di erent from what we expect from the outside.

We can now give two model metaphysical views of location on

which we can get real presence by multilocation. On the rst model,
space (or space-time—the same things can be said in a more
relativistic framework, but for simplicity I just deal with space)
consists of points. There is also a primitive relation L that can hold
between an extended entity x and a non-empty set P of points in
space at any given time. We can read LxP as ‘x is wholly located at
P’ Given this, what happens in the real presence on the present
model is that the internal causal relations and intrinsic properties of
Christ’s body remain as they were, but Christ’s body comes to be
additionally L-related to the area in space to which the bread was
previously related. Location is de ned by a primitive relation
between an object and a nonempty set of points, and while normally
a material object is thusly related to only one set of points,
miraculously it can come to be thusly related to more than one set.
This seems to be a coherent story about locatedness and of what it
would mean for Christ’s body to come to be locally present in more
than one place without changing in internal relations and intrinsic
properties.

On the second model, space is constituted by a network of

relations between things in space. The exact nature of these
relations will depend on the details of the relational view. One
option is the following: We have a relation Rxyd which means that
the distance between objects x and y is d. It might, further, make
sense to suppose that a pair of points can stand in more than one
such relation if space is non-euclidean: thus there might be more
than one distance between two points. For consider the two
di erent ‘straight line’ distances between Washington and the North
Pole (assuming for simplicity the earth is a perfect sphere): one is
obtained by measuring the length of the trip directly north from
Washington to the North Pole, and the other is obtained by
measuring the length of the ‘straight line’ (or more precisely ‘great
circle’) trip south from Washington to the South Pole and then up to
the North Pole on the other side of the globe. Thus it might well be
that two points have more than one distance between them: for
instance, a distance ‘the short way’ and a distance ‘the long way’.5

The location and shape of an entity x might well then be

constituted by all the Rzyd relations where z ranges over the parts of
x and y ranges over entities other than the parts of x. On this model,
real presence is again no di culty to an omnipotent being who can
bring it about that if Rxyd held with x a part of the preconsecration
bread and y something outside it, then now Rx*yd holds where x* is
some part of Christ’s body, and that every part of Christ’s body
enters into one of these new relations. It appears possible to make
this coherent.

Observe that Christ’s body then comes to be at a positive

distance from itself: if the distance between one eucharistic host and
another is 100 meters, then it will be the case that R (Christ’s body)
(Christ’s body) (100m). But there is no di culty in something being
a positive distance from itself: an object on a perfect sphere is both
at zero distance from itself and at a positive distance (just walk
straight in any direction and you’ll come back where you started).

Thus it seems possible to produce coherent metaphysical models

of how a multi-location of Christ’s body might work in a eucharistic
context. Until we decide what is the right metaphysics of location,
we cannot say which model is correct. But at least we have reason
to think that there is nothing in any way obviously contradictory
about the hypothesis of a real multilocation.

Two further remarks are in order. The rst is that Catholic

teaching holds that Christ whole and entire is present in each part.
This requires massive multilocation, where in fact Christ comes to be
wholly located in the place where any part of the eucharistic bread
was. But this in turn simply requires more L or R relations on the
above models.
 The second remark is that we are now in a position to see that
there are prima facie two kinds of bilocation. The rst we might call
robust bilocation. This is the kind of bilocation that would obtain if
you went back in time and were simultaneous with your past self.
The robust bilocation of a person would be a doubling of mental
capacity and a doubling of centers of consciousness. In robust
bilocation, the entity at each location can have what at least
intuitively seems to be di erent intrinsic properties (though they
may turn out to be relational). Robust bilocation is what generates
the problem of intrinsic properties, as well as the challenge of
coming up with an account of ‘wholly located’ that allows a
di erent mix of parts to be in each of two places at each of which
the entity is wholly present.

But if the eucharistic real presence is a bilocation, it is probably

not a robust bilocation. One way to see this is to note that the
cessation of the existence of a person as wholly located in one place
in a case of robust bilocation is very much like a death. If you go
back in time to meet your younger self and then you die while in
the past, then you have died, though oddly enough you continue to
be alive at the location of your younger self. But the consumption of
the Eucharist is not seen by the Christian tradition as this closely
akin to the death of Christ, given that Christ is now glori ed, though
it might well be seen as representing that sacri cial death.

In non-robust bilocation, on the other hand, the same entity

comes to be present in multiple places, but has the same intrinsic
properties in each of those locations (except perhaps shape and size
if we are forced to concede shape and size to be intrinsic). If the
entity is a person, centers of consciousness and mental capacity are
not multiplied, but all that happens is that the person becomes
related locationally to multiple places. If you need to be active at
meetings of two di erent committees at the same time, you need
robust bilocation—you need a di erent set of mental capacities to
be available to each committee, for instance. If you just need to ll
out the quorum of each of two di erent committees but do not need
to participate in both discussions, non-robust bilocation will do. Un-
bilocating after the two meetings in the non-robust case will not be
like a death—you will just be once again present in one place
instead of two.

Non-robust bilocation seems to have signi cantly fewer

philosophical di culties than robust bilocation. And if the
eucharistic real presence needs bilocation at all, it only needs the
non-robust kind.6

3. REAL ABSENCE

According to Catholic dogma, not only is Christ’s body really

present, but bread is really absent. All that remains of it are the
‘appearances’ (‘species’) of bread. These have the same e ect on
measuring instruments as bread does, but the substance is no longer
there.
 Aquinas’s metaphysical view was that what remains are the
accidents of bread. These accidents are not the accidents of any
substance any more. They are not the accidents of bread since there
no longer is any bread. But neither are they the accidents of Christ’s
body or of any third substance such as the air (note that the
accidents remain interrelated: in fact all the accidents other than the
dimensions inhere in the dimensions according to Summa Theologiae
III q. 77 a. 2). The idea of accidents not attached to a substance
seems to play havoc with Aristotelian categories on which accidents
are precisely what exists in a substance. However, Aquinas modi es
that traditional de nition to say that ‘it belongs to the quiddity or
essence of accident “to have existence in a subject”’ (Summa
Theologiae, III q. 77 a. 1 ad. 2). The analogy of the four-leggedness
of a sheep might help. It belongs to the essence of a sheep to have
four legs. But a particular sheep might have only three. ‘Essence’ in
this context does not refer to what the entity must have, but to that
which it in some sense should have given what it is like, but might
in fact lack. Thus an accident is the sort of thing that should have a
subject, but God can miraculously bring it about that it does not.

Aquinas’s argument for this possibility is that substances cause

their accidents, and yet God is closer to each thing than its non-
divine cause, so he can directly preserve the being of the accidents
even absent the substance (Summa Theologiae III q. 77 a. 1).
Accidents, thus, cannot be seen as properties—for only something
can have a property. Rather, Aquinas sees accidents as real beings,
albeit ones dependent on a substance in a deep way.

We can perhaps do better. Suppose one believes the past is real,

in the sense that Alexander’s warhorse Bucephalus exists, though
not in the present-tense sense of exists’ but in a tenseless sense. We
can introduce this tenseless sense of exists’ by two examples. When
we say that a solution to the equation x2 + x = 9 ‘exists’, we are
not saying anything about the time of its existence—the word exists’
is really tenseless (even though grammatically in English it looks
like the present tense). Likewise, when we say that the root of
World War II lies’ in the discontent of the German people after
World War I, we are not implying that this discontent is something
present—we are using lies’ in a tenseless sense, that does not imply
pastness, presentness, or futurity. If this is right, then likewise the
bread and wine exist, though pastly, after consecration, and the
conclusion that the accidents of bread exist without a subject can be
blocked without a rming that the accidents are accidents of
Christ’s body or the air or suchlike. For why cannot an accident
have its subject exist pastly if the past is real and past existence is
bona de existence? Certainly an appearance can survive that which
it is the appearance of. One hears a lightning strike after the strike is
already over and it is quite possible to see the light from a distant
star that in some reference frames has long ago perished. If the
sound of the lightning strike and the sight of the star count as
accidents, then it seems accidents can exist while their subjects no
longer do.

Aristotelians do think of accidents as e ects of substances. If we

accept that e ects need not be simultaneous with their causes, then
seeing the accidents as continuing to exist as e ects of the bread
does not seem all that absurd.

Alternatively, one might identify accidents with certain

dispositions. Thus, the accident of being positively electrically
charged might be identi ed with the disposition to attract negative
charges and repel positive ones in an appropriate way. Seen this
way, the accidents of bread can persist after the cessation of the
bread if God arranges some other way for these dispositions to
subsist. (It might not be the numerically same causal powers,
depending on the criteria of individuation of dispositions. But it is
not clear that it is necessary to insist on numerical identity here.)

If it is possible for there to be bare dispositions without a

substance, so that there might be a disposition in a region of space
to bend light in a particular way, there is no di culty here at all. Or
it might be that God directly e ects everything these dispositions
did, so that all the same counterfactuals remain true as were
previously true. If all we mean by saying that there is such-and-such
a disposition is that certain counterfactuals (such as that negative
charges would be attracted in a particular direction) hold and have
a truth-ground of the fact of their holding, then the dispositions can
indeed be said to continue existing. It is also philosophically
possible, though there might be theological di culties here,7 that
some third substance comes to be the ground of these causal
powers. Finally, perhaps there could be dispositions grounded in the
power of a pastly existing substance, like the pastly existing bread.

The possibility of the accidents of bread inhering in Christ or

Christ’s body, on the other hand, would seem to require something
like robust bilocation, with its attendant problems of intrinsic
properties. And if these problems were solved, the view might
nonetheless be theologically problematic, if it be thought un tting
that the glori ed Christ become small and round (this was
apparently a standard medieval objection to certain views).

Interestingly, Aquinas distinguishes between the accidents of

bread and the properties that follow from its essence, such as its
ability to nourish us. It is only the accidents that remain. The ability
to nourish, then, must be miraculously supplied by God. Here, I
think, we might make a theological improvement. Why not say that
Christ’s body gains the ability to nourish us? If so, then we can
understand why Christ says in the Gospel of John (6: 55) that his
esh is true food.

The main philosophical challenge of the doctrine of the real

absence of bread and wine is to explain how it is that the
appearances of bread and wine are there, with the same physical
interactions happening as happened prior to consecration. As we
saw several explanations are available.
 At the same time, one might ask a non-metaphysical question:
are not the appearances of bread and wine deceptive absent the
substance of bread and wine? Aquinas notes that our senses are not
deceived. They see the accidents, and the accidents are there. A
similar thing can be said on dispositional views—it really is true
that there is a disposition to de ect light rays in the way bread does,
and so on. Of course an inference to the existence of a subject for
these accidents would be incorrect. But faith can guard the Christian
from such an inference. And we need the appearances of bread and
wine ‘because it is not customary, but horrible, for men to eat
human esh, and to drink blood. And therefore Christ’s esh and
blood are set before us to be partaken of under the species of those
things which are the more commonly used by men, namely, bread
and wine’ (Summa Theologiae III q.75 a. 5).

4. CONCLUSIONS

There are no metaphysical di culties with symbolic presence. We

have seen that it is possible to make sense of transubstantiation,
understood as the conjunction of real presence and real absence.
Granted, to make such sense we need to make metaphysical
assumptions. But there is more than one set of metaphysical
assumptions that will do. And clever thinkers can no doubt nd
others. It would also be possible to make sense of consubstantiation,
provided that we can make sense of impanation—Christ becoming
bread—or that it is possible for two distinct material objects to be
wholly located throughout the same place at the same time.

NOTES

1. These two questions arose in conversation with Terence Cuneo.

2. ‘Valid’ here is a technical term from Catholic canon law. An

apparently sacramental action is valid provided that it is done
in such a way, by such a person, with such intentions, and
under such circumstances as guarantee, given God’s plan, the
sacramental e ect—in this case, Christ’s real presence. The
exact conditions believed to be necessary for the validity of a
eucharistic celebration di er between denominations.

3. Here is one way to do the construction mathematically. It is

only going to yield topological and not metric properties, but
with some work metric properties can be handled. First take
some one-to-one correspondence between points within the
wafer in Rome and points within the wafer in Washington, and
de ne the relation ∼ on points in space as follows: x ∼ y if
and only if either x = y or x and y correspond under the
aforementioned one-to-one correspondence. A topological
space consists of a set S of ‘points’ and a set O of subsets of S,
known as ‘open sets’, where O is closed under nite
intersections, and in nite unions, and contains S. Let S be our
space and O its topology. De ne a new space S* containing all
the points of S minus those found in the wafer in Washington.
 Let U be a set in O. De ne U* as the set of all points x in S*
that stand in relation ∼ to some point of U. We can now let
O* be the set of all sets of the form U* where U is in O. It is
easy to see that U* is also closed under nite intersections and
in nite unions, and contains S. The desired quotient space,
then, consists of S* and O*.

4. For mathematically more sophisticated discussions of

intrinsicness and relativity, see Bricker 1993 and Butter eld
2005.

5. If we want to be precise, we can say that Rxyd provided that

there is a geodesic (a generalization of a straight line) between
x and y of length d.

6. This also allows an additional escape from a worry that Hud

Hudson 2005:113–16 raises for the multilocationist. If
multilocation is possible, is it not a more parsimonious theory
of the world to suppose there is only one particle in existence,
multilocated at 1080 regions, and doing di erent things in
di erent regions? However, this would seem to require robust
multilocation. In any case, I do not think Hudson’s worry is
very pressing. The theory that the universe consists of a single
multilocated particle would surely generate absurd
consequences for the moral life, which is surely good enough
reason for rejecting it, once we grant, as we should, that we
have genuine moral knowledge of our duties (we have the
right to be more con dent about the wrongness of torturing
 persons for fun than about just about any scienti c theory).
For instance, it is unclear how there could be multiple
embodied persons in such a world.

7. One worry is that ‘this’ gets its reference from the visible
accidents when these are attached to a substance. If there were
some third substance in view, then ‘This is my body’ might
seem to refer to that substance, which would be incorrect.

REFERENCES

AQUINAS, THOMAS (1920). Summa Theologica. Transl. the Fathers of

the English Dominican Province,
<www.newadvent.org/summa>, accessed 28 May 2008.

——— (undated). Summa Theologiae, trans. A. J. Freddoso,

<http://www.nd.edu/∼afreddos/summa-
translation/TOC.htm>, accessed 28 May 2008.

BRICKER, PHILLIP (1993). ‘The Fabric of Space: Intrinsic vs. Extrinsic

Distance Relations’, Midwest Studies in Philosophy 43: 271–93.

BUTTERFIELD, J. (2005). ‘Against Pointillisme about Geometry’,

<http://philsci-
archivepitt.edu/archive/00002552/01/APG1.pdf>, accessed 28
May 2008.

CRISP, THOMAS M., and Smith, Donald P. (2005). ‘“Wholly Present”

De ned,’ Philosophy and Phenomenological Research 71: 318–44.
HUDSON, HUD (2005). The Metaphysics of Hyperspace. Oxford:
Clarendon.

KOONS, ROBERT C. (2005). ‘On Being in Two Places at the Same

Time’, unpublished.

LEIBNIZ, G.W. (1989). Philosophical Essays, ed. and trans. R. Ariew

and D. Garber. Indianapolis: Hackett.

LEWIS, DAVID (1986). On the Plurality of Worlds. Oxford: Blackwell.

MODANIEL, KRIS (2003). ‘No Paradox of Multi-Location’, Analysis 63:

309–11.

PAUL VI (1968). ‘Credo of the People of God’,

<http://www.vatican.net/holy_father/paul_vi/motu_proprio/do
cuments/hf_p-vi_motu-proprio_19680630_credo_en.html>,
accessed 28 May 2008.

SKOW, BRADFORD (2007). ‘Are Shapes Intrinsic?’ Philosophical Studies

133: 111–30.
 PART V

NON-CHRISTIAN PHILOSOPHICAL
THEOLOGY
 CHAPTER 24

JEWISH PHILOSOPHICAL THEOLOGY

DANIEL H. FRANK

IN 1783, at the height of the Enlightenment, the Jewish savant

Moses Mendelssohn wrote Jerusalem, or On Religious Power and
Judaism. It is an apologetic work, attempting to make the case for
inclusion of Jews and Judaism in modern European social and
political life. Jews had long been debarred on account of their
strange ways and practices, antithetical to Christianity. Seemingly
irrational and anachronistic, Judaism needed to undergo reform, if
ever Jews were to be full participants in the modern age. What form
should Judaism take to shed itself of its past? What precisely is
‘traditional’ Judaism that stands in the way of reform? Minimally
we can say that traditional Judaism is at least a certain set of
practices supportive of anti-idolatrous worship, of a strict
monotheism that requires obedience to the one God, creator of
heaven and earth. Such practices are quotidian, from daily prayers
to digestion of food to washing of hands to sexual practices. All
these activities have the e ect of sanctifying the mundane, and
support the notion that Jews are a ‘holy’ people in the service of the
one God.
 Given this, one might well suppose that the requisite reform of
Judaism would entail an abrogation of its ‘outdated’ practices. Jews
should commence to eat both with and like their non-Jewish
brethren, to work on the seventh day, etc. As one peruses the
history of Jewish life in the post-Enlightenment period, it is
certainly the case that many western European Jews, and somewhat
later, Jews in the Americas, reformed their religion in just these
ways. Full inclusion in civic life, and civil society, seemed to require
a level of assimilation to contemporary ways that necessitated the
abandonment of traditional practices. And so Jews and Judaism
reformed themselves. At this point one may begin to wonder what
beliefs are underwritten by such revised practices. As noted, a very
strict monotheism is underwritten by traditional practices. A strong
dichotomy between secular and divine, creation and creator is
mandated. What happens to such binary oppositions when
traditional practices, whose sole purpose is to support these
oppositions, are undercut? What one sees in much of Jewish reform,
especially in its initial stages in the mid-nineteenth century, is a
diminishing of the strict monotheism of earlier times. The via
negativa of Maimonides, which more than anything underscores the
unfathomable otherness of God and our epistemological fallibility,
gives way in some in uential modern quarters to an understanding
of God as an ethical ideal.1 Whatever one may make of this gloss, it
indexes God to our moral imagination, and worse, from a traditional
vantage point, it makes God a partner in collective action. It tends
to understand God as caring about what we care about, and even
su ering along with us.

In addition, of course, there have also been deeply anti-religious

reforms of Judaism, of which political Zionism, secular Jewish
nationalism, is the best example. Political Zionism (there are also
religious versions of Zionism) has as its goal the establishment and
continuing existence of the Eretz Israel, the land of Israel. The
nation-state replaces God as the goal of human action, and with it
the binary opposition between secular and divine completely
collapses.

So, there are many ways of Jewish reform, but interestingly

Mendelssohn’s brief on behalf of Judaism does not call for a reform
of Jewish practice. Famously he writes,

Among all the prescriptions and ordinances of the Mosaic law, there
is not a single one which says: You shall believe or not believe. They
all say: You shall do or not do. Faith is not commanded, for it accepts
no other command than those that come to it by way of conviction.
All the commandments of the divine law are addressed to man’s
will, to his power to act. In fact, the word in the original language
that is usually translated as faith [emunah] actually means, in most
cases, trust, con dence, and rm reliance on pledge and promise …
Commandments and prohibitions, reward and punishment are only
for actions, acts of commission and omission which are subject to a
man’s will … Hence, ancient Judaism has no symbolic books, no
articles of faith. No one has to swear to symbols or subscribe, by
oath, to certain articles of faith. Indeed, we have no conception at
all of what are called religious oaths; and according to the spirit of
true Judaism, we must hold them to be inadmissible.2

This is a remarkable passage. In it Mendelssohn defends Judaism

at the bar of the Enlightenment, but not as we might expect. He
does not call for a reform of religious practice and ritual. Rather, in
understanding that the olive branch o ered to Jews is grounded in a
commitment to religious toleration, Mendelssohn responds in kind,
that Judaism (as a tolerant faith) does not command belief, and that
the (eternal) commandments are injunctions to act in certain
speci able ways. For Mendelssohn, that Judaism knows of no
articles of faith entails that Jews can be (practicing) Jews and loyal
citizens of modern states. Jews are not bound by ‘religious oaths’,
which would render them servants to an authority other than the
state. Their ‘obedience’ to God takes the form of living a certain way
of life, with supra-political goals. But these goals in no way
compromise or lessen loyalty to the state. For Mendelssohn, Jewish
religious practice and ritual is eternal, the goals of the
commandments and prohibitions. By contrast, beliefs are not, and
cannot be, commanded, and this point of empirical psychology
entails for Jews considerable latitude in belief, even as they practice
the religion of their forebears. The demand upon Jews to express
their loyalty to the secular state is feasible for Mendelssohn since
there is no con ict of interest. One is reminded here of a similar
discussion when John Kennedy, a devout Catholic, was a candidate
for the Presidency. A worry arose about divided loyalties; from
whom would he take his orders—from Rome or from the
Constitution of the United States? Analogous with Mendelssohn,
JFK’s response was that his ability to uphold the Constitution and be
a solid citizen was compatible with being a Catholic. In both cases, a
bifurcated existence, living a life whose trajectories are at once
political and supra-political, is not only a possibility, but a necessity,
if ever certain prejudices were to be overcome.

If Mendelssohn is right that his ancestral religion is incompatible

with ‘orthodoxy’, then prima facie Judaism would appear to be
unpromising territory for (normative) theological speculation.
Indeed, ‘Jewish philosophical theology’, the very topic at hand,
would seem to be a non-starter. Often Judaism is characterized as
‘orthopraxic’, this in distinction to Christianity, whose commitment
to articles of faith and creeds are clear. The emphasis in Judaism is
on practice, actions that are commanded and prohibited, not on
recti cation of belief. The perennial (anti-Semitic) charge of the
‘unspirituality’ of Judaism seems to point in the same direction,
supposing that Jews go about their daily lives in an utterly
‘bloodless’, emotionally lifeless way. Detractors of Judaism might
try to capture this point by reference back to Mendelssohn’s remarks
about the Hebrew term emunah. In wishing to establish the case for
‘freethinking’ in Judaism, and in so doing to indicate the
compatibility of Judaism with the Enlightenment ideal of
unencumbered thought, Mendelssohn is at pains to gloss emunah
(faith) as trust and con dence. In the place of faith and articles of
faith, Mendelssohn substitutes rm trust grounded in promises. Jews
made a promise at Sinai to keep commandments, and this promise
and pledge provides the eternal foundation for their living in a
certain way. The psychological point being made here about the
direct object of imperatives is not new with Mendelssohn. It can be
found in Locke’s Letter on Toleration (1689) and in medieval Jewish
circles in the late fourteenth century anti-Maimonidean, Hasdai
Crescas. The point is simply that belief cannot be commanded, and
it is quite pointless to found a society on a set of beliefs, and to
demand of the citizenry articles of faith. Rather, the goal should be
goodness and rectitude of actions, and prima facie this latter is
compatible with a certain doxastic freedom.

As noted, such an act-centeredness might be construed as an

emotionally dead, somnambulistic going through the motions.
Certainly Mendelssohn never entertained this thought, wishing
instead to present Judaism and Jews as open-minded and modern.
But the issue is now joined, about the undogmatic nature of
Judaism, and its presumed inhospitality to dogmatics and even
theological speculation. What are we to make of Jewish
(philosophical) theology?

A MEDIEYAL PROLEGOMENON
If we turn back to the medieval period, when Jews and Judaism
were not vying for a spot at a (modern) secular table, we might
imagine we would see Judaism in an unapologetic mode, not intent
on (ex hypothesi) e acing dogmatic, credal commitments. With a
view to maintaining group loyalty, we might expect a certain
presumption of orthodoxy and credal commitment. The rst great
work of medieval Jewish philosophy, The Book of Doctrines and
Beliefs, written in the rst half of the tenth century in Baghdad by
Saadya Gaon, is a work of Jewish kalam (dialectical theology).3 It
presents in considerable detail discussions of a standard set of issues
such as creation of the world, the nature and unity of God, divine
justice, freedom and determinism, and divine reward and
punishment. In its own way this work is an apology, a defense of
traditional Judaism, but the defense is of a certain way of
understanding Judaism, namely as committed to a distinct set of
philosophical theses, noted above. Saadya’s project is manifestly not
a defense of (what Mendelssohn would call) ‘articles of faith’.

Saadya has well been called ‘the revolutionary champion of

tradition’.4 The ‘revolutionary’ nature of his book lies in its
underlying assumption that Judaism—traditional (rabbinic) Judaism
—is in fact (reducible to) a set of doctrines and thus amenable to
systematic theological speculation. It was due to Saadya that a
defense of Judaism took the form of a Mu‘tazilite (Greek-inspired)
theological discussion of such issues as creation ex nihilo, the
absolute oneness of God, free will, and divine reward and
punishment. What might have been, for the religious
fundamentalist, a reactionary call to cease philosophizing becomes
in Saadya’s hands a rich philosophical fare of argument and
counterargument. Saadya’s book is a full participant in
contemporary intellectual life, while it defends traditional Judaism.
Jewish philosophical theology is born.

The very title of Saadya’s philosophical magnum opus clari es

his innovation. The Book of Doctrines and Beliefs was written in
Arabic under the title Kitab al-Amanat wa’al-I‘tiqadat. It was rst
translated into Hebrew in 1186 by Judah ibn Tibbon as Sefer ha-
Emunot ve-ha-De‘ot. Focusing on the title, both in Arabic as well as
Hebrew, reveals much about the purpose of the book as a whole.
Amanat (Hebrew emunot) are beliefs held on the basis of scriptural
authority. Contrarily, i‘tiqadat (Hebrew de‘ot) are the very amanat
subjected to rational re ection and critical scrutiny. The purpose of
Saadya’s book is thus ‘to enable the reader to reach a stage where
the Amanat (“doctrines,” i.e. of Judaism) become the object of
I‘tiqadat (“conviction,” i.e. faith based on speculation)’.5 In the
introduction to the ten chapters, or treatises, that make up the text,
Saadya indicates the dynamic of the book when he says that ‘the
believer who blindly relies on tradition will turn into one basing his
belief on speculation and understanding’.6

Viewed this way, the trajectory of Saadya’s project reminds one

of Aristotle’s general (dialectical) way of proceeding
philosophically.7 Both Aristotle and Saadya commence with the
status quo, generally the untutored beliefs and customary actions of
the neophyte, and proceed to transport the ‘student’ to the point
where he begins to understand the grounds for those very beliefs
and actions. In his own way, Saadya wishes to turn (mere) belief, or
even confusion, into rationally grounded conviction. As we shall
soon see, we should likewise understand Maimonides’ own
philosophical project in The Guide of the Perplexed. It is Aristotelian
in the way just noted, since it is addressed to a traditional Jew,
described as ‘perfect in his religion and character’ (introduction to
the Guide), who, on account of naively having accepted traditional
beliefs, has nally become perplexed by the ‘externals of the law’
(its literal meaning), by ‘having studied the sciences of the
philosophers’.8 Traditional, unre ected-upon beliefs square o
against natural science and philosophy, with apparently disastrous
e ect. And so it becomes Maimonides’ grand project in the Guide, as
it was Saadya’s some centuries earlier in his book, to clarify the
tradition in such a way that its ‘philosophicality’ is revealed, and the
addressee’s perplexity is removed. In fact, it is just this unearthing
of the ‘philosophicality’ of Scripture that so annoyed Spinoza in the
Tractatus Theologico-Politicus (1670).9

What is to be noted right at the outset of Jewish theology in the

medieval period is its penchant for philosophical argumentation, for
advancing an understanding of Judaism on philosophical grounds.
This is not to suggest that philosophical theology did not have its
detractors, those who viewed such speculation as suspect, as leading
to possible antinomianism. Rather the suggestion is that credal
formulation, the development of a kind of catechism, is not the
desideratum of theologians. At this point we might recall
Mendelssohn’s comment that ‘ancient Judaism has … no articles of
faith’.

I have spoken above of the medieval philosophical project as one

of ‘unearthing’ the ‘philosophicality’ of Scripture. The metaphor
should be pressed. The medieval philosophers are, in their own way,
archaeologists. They wish to root out the deepest layer of truth,
beneath the surface. Beneath the literal meaning of biblical text,
likewise underlying spatio-temporal phenomena, are deep truths
and regulative ideals. And these truths are to be teased out by
philosophical acumen. They are manifestly not emunot, dogmatically
held beliefs; rather they are the grounds for belief, and this is why
Aristotle is so dear to them and why natural theology, grounding
belief in science, is pronounced.

It is important now to turn at some length to the very greatest of

the medieval Jewish philosophers, Maimonides (1138–1204). The
reason for so doing is simple. He is taken to have commenced
Jewish dogmatics, the enunciation of principles (‘foundations of
faith’), acceptance of which de nes a Jew as such, and brings about
salvation.10 In his commentary on Pereq Heleq, the tenth chapter of
the Mishnaic tractate Sanhedrin, Maimonides outlines thirteen
foundations of the law. The thirteen foundational principles are
belief in God’s existence, unity, incorporeality, ontic priority
(creation ex nihilo), that only God may be worshipped, that
prophecy exists, that Mosaic prophecy is unique, that the Torah is
revealed, that the Torah is eternal, that God has knowledge of
human acts, that God rewards and punishes, that the Messiah will
come, and nally that there will be resurrection of the dead. The
actual list and number of principles is less important than the rather
simple terminological point that Maimonides’ refers to them as
qawa‘id al-i‘tiqad (foundations of belief). It is important to recall that
this precise terminology is that used by his predecessor Saadya,
when the latter laid out his own philosophical defense of Judaism.
The point I am now stressing is that Jewish dogmatics is not what
we might think it is, an unmediated belief in a set of doctrines,
acceptance of which de nes a Jew as such. On the contrary, as we
can infer from Saadya himself, belief is the conclusion of rational
speculation, hard mental work. And the reward is salvation. Indeed,
Maimonides understands love of God as a function of knowing
him,11 and such knowledge is (scienti c) knowledge of his
(bene cent) creation. In this regard, Maimonides and his great
detractor, Spinoza, join hands. Both understand the summum bonum
as an amor Dei intellectualis.

At the end of the medieval period, Isaac Abrabanel agreed with

Maimonides on this, and his thought on the matter clari es the issue
at hand. Abrabanel asserts, ‘Maimonides did not count as a positive
commandment the form of the belief and its truth, but, rather,
knowledge of those things which bring one to acquire beliefs’
(Principles of Faith [Rosh Amanah], trans. Kellner, p. 155). On this
passage Menachem Kellner comments, ‘According to Abrabanel,
when Maimonides codi es as commandments the obligation to
accept that God exists, is one, and is incorporeal, he means that we
are commanded to apply ourselves to the study of physics and
metaphysics so as to become convinced of the truth of the claims
that God exists, is one, and is incorporeal.’12 Kellner’s gloss here on
Abrabanel is a good one and provides a way for precisely
understanding Maimonides’ own position. When Maimonides
commands belief in divine existence, unity, etc., he commands the
undertaking of a process of study by which one will be led by the
light of reason to see that God is a necessary existent, is one, etc.
Maimonides does not command belief in the sense of enjoining one
to stop what she is doing at the moment and to commence believing
something. Recall the point made earlier by Mendelssohn that belief
cannot be commanded. So Maimonides too does not command belief
ex nihilo, but rather an activity (of scienti c inquiry) that will result
in rmly established convictions.

This very process has practical implications. Maimonides’ project

can be understood as suggesting that something more than mere
behavior counts as normative. Like Aristotle, he sets the bar for
virtue high, requiring an understanding of reasons or principles of
action. For Maimonides, the law itself has reasons, and to achieve
human excellence, one must ascertain those reasons to the extent
that a human can. Short of this, one is not a Jew in a normative
sense. Maimonides knows better than anyone that being a Jew de
facto does not require the arduous intellectual program just
outlined. In this latter sense, one needs only to be born of a Jewish
mother or to have converted according to traditional practice, but
ever the intellectual elitist, Maimonides demands more. He requires
that Jews go beyond the given, beyond easy acceptance of the law
and behavior commensurate with it. He demands that we come to
see why we do what we do, and to see what the scienti c and
cosmological foundations for our actions are. The project here is not
di erent from Maimonides’ grand project of discovering ta‘amei ha-
mitzvot, the reasons for the commandments. And this latter project is
itself commanded.

TA’AMEI HA-MLTZVOT (THE REASONS FOR THE COMMANDMENTS): A CASE

STUDY IN JEWISH PHILOSOPHICAL THEOLOGY

In what follows I present a case study of Jewish philosophical

theology, in which I show how Maimonides explicates the reasons
for the revealed commandments. Prima facie some of the
commandments appear to be quite arbitrary and irrational, and we
shall see how Maimonides deals with this. Further, and this is a very
important point, this ‘theoretical’ discussion in legal philosophy
about the reasons for the commandments has manifestly practical
implications, speci cally aretaic implications about the inculcation
and establishment of certain dispositions.
 We moderns tend to share with Kant the notion of morality as a
struggle, a struggle against those impulses that would overwhelm
us. To the victor belongs the praise. In this regard, the rabbinic
sages are no di erent from us. According to them, l‘fum tza’arah
agrah [the reward is according to the pain] (Avot 5. 23). In
re ecting upon this rabbinic dictum, however, Maimonides is struck
by its apparent inconsistency with the philosophical, Aristotelian
view, which explicitly accords the self-controlled person (the
enkrates), the one who successfully battles against the pull of
contrary impulses, merely second place to the virtuous (temperate)
person, this latter being one who has no bad appetites to overcome
and who takes pleasure in the doing of virtuous acts.13

In facing this apparent inconsistency between the respective

philosophic and rabbinic evaluations of virtue and self-control,
Maimonides, in Shemonah Peraqim (Eight Chapters, the introduction
to his (early) commentary on Mishnah Avot) 6, brings about, in his
own words, ‘a wonderful reconciliation’ between the two positions.
Maimonides harmonizes the philosophic and rabbinic positions by
distinguishing between two types of law, generally accepted laws
and the ceremonial and ritual laws labeled huqqim. The rabbinic
praise of self-control, he notes, has reference solely to the ritual laws
such as sha‘atnez (prohibiting the wearing of certain mixed fabrics)
and the eating of meat with milk. But as for the universal and
customary prohibitions, such as those against murder and theft,
both the philosophers and the rabbis agree that the virtuous person,
without wayward desires, is to be ranked higher than the one who
succeeds in controlling his desires.

Yet a problem lurks beneath this apparent ‘reconciliation’. The

argument depends upon a radical dualism between laws universally
accepted for their intrinsic rationality and those accepted for no
reason other than their origin in a divine command. This radical
distinction, however, cannot be squared with the view that
Maimonides presents in Guide 3. 26, that all the laws, including the
huqqim, derive from God’s wisdom, not from his will alone.14

What is to be done about this inconsistency? In overcoming one

inconsistency, Maimonides appears to be ensnared in another. To
get our bearings on the nuanced discussion in the Guide, we need to
take a look back at Shemonah Peraqim 6. In focusing on the
reconciliation that Maimonides there e ects between the
philosophers and the rabbis, we tend to overlook what motivates the
inconsistency in the rst place. The whole problem between the
philosophers and the rabbis arises because the latter have rank-
ordered self-control above Aristotelian virtue in some cases. Why?
An obvious answer might be that praise of self-control, of successful
struggle against contrary impulses, underscores in a stark and vivid
way the heteronomous character of the law (speci cally, the
huqqim), and obedience to it celebrates the duty to obey the divine
will in spite of human disinclination.

In support of this gloss of the rabbinic ranking, one might point

to Maimonides’ explicit rabbinic references in chapter 6. As Rabbi
Gamaliel said, ‘Let a man not say, “I do not want to eat meat with
milk, I do not want to wear mixed fabric … but [let him rather say]
I want to, but what shall I do—my Father in heaven has forbidden
me.” ‘15 This rabbinic statement might be understood as magnifying
God, his law, and obedience to it, all at the expense of humans, and
their nature. This gloss, however, is one-sided.

Second to no one in pointing out the unbridgeable gulf that

separates the human from the divine (the very thrust of his
celebrated ‘negative’ theology), Maimonides is not making this point
here. Far from glorifying God, the rabbis praise humankind. In
presumably agreeing with the rabbinic praise of self-control and the
rank-ordering of it over virtue in some instances, Maimonides is eo
ipso countenancing the struggle in which the self-controlled person
engages and emerges victorious. Again, in context, the rabbis
forbade a person to say in respect of, for example, eating meat with
milk: ‘I would not naturally yearn to commit this transgression, even
if it were not prohibited by the law.’ The point stressed is
anthropocentric in its intent, a point about human beings and the
‘naturalness’, the ‘givenness’, of certain human inclinations. Better
to have recalcitrant desires (in certain situations) than no desires at
all, given that, according to the rabbis and Maimonides, ‘if it were
not for the law, they [the huqqim] would not be bad at all’. Since
there is no intrinsic rationality to the huqqim, it is quite natural for
human beings to jib at obedience to them. So, the Shemonah
Peraqim.
 As one turns from the Shemonah Peraqim to the Guide,
Maimonides’ philosophical magnum opus, I would emphasize that
even when in 3. 26 Maimonides re-evaluates his earlier (dualist)
position concerning the nature of the law, he maintains a view of it
that is attentive to the nature of humankind. Even as Maimonides
points out the general intelligibility of the law, contrary to the
Shemonah Peraqim, he is aware of the limits to which human nature
can be brought under the rule of reason.16 Maimonides’ entire
discussion concerning the reasons for the commandments strives not
only to elicit the general intelligibility of all the laws, but to
maintain some sort of distinction between the mishpatim (the
intelligible ones) and the huqqim (the ritual, ceremonial laws).

As much as countering the Ash‘arite (voluntarist) claim that all

the laws are the product solely of God’s will, as well as the
Mu‘tazilite view that only some of the laws are the product of divine
will, Maimonides is also concerned to forestall (what we might call)
‘hyperrationalism’, the rationalist philosopher’s attempt to discover
reasons for absolutely everything. In the present case, it is quite
pointless, according to Maimonides, to search for the reason why a
lamb rather than a ram was used for a certain sacri ce, or why a
certain number of lambs was sacri ced.17

Searching for the reason for every particular proves to be vain,

for there was never a reason for God’s choice in the rst place. The
quest for reasons here is a nonstarter. While Maimonides is clear
that historical knowledge (of time and place) is certainly helpful in
gaining an understanding of the reasons for the law,18 he is equally
insistent that ‘no cause will ever be found for the fact that one
particular sacri ce consists in a lamb and another in a ram and that
the number of the victims should be one particular number’.19
Although one might explain our ignorance by appealing to the
historical distance from Temple days, or to our incapacity as (mere)
mortals to fathom God’s plan, I myself think such explanations are
forced. They arti cially place the burden of responsibility for our
lack of understanding upon us.

Yet Maimonides holds explicitly that it is not any kind of human

limitation, biological or historical, that precludes our nding the
reasons for the particulars in the laws of sacri ce. Rather, the reason
is wholly non-empirical; indeed, it is one of logical necessity. The
reason why some number of lambs (seven) rather than another
number (eight) was chosen for sacri ce is simply that for the law (of
sacri ce) to be instantiated, that is, for there to be a law, some
choice had tobe made. But no other ‘reason’ is to be found. We may
wish to say that the choice was ‘reasonless’, without (rational)
preference, but we need to be clear that this lapse of rationality does
not limit the wisdom of God, nor does it covertly signal a return to
the kalamic (speci cally, the Mu‘tazilite) position regarding the law.

In sum, the fact that some laws are the product of reasonless
choice, enforced by logical necessity, does not compromise divine
wisdom. God is wise, and so is his law, even in those situations in
which he chooses, perforce, without preference. Given this,
Solomon’s traditional wisdom concerning grounds for the
commandments (ta‘amei ha-mitzvot) is con rmed: it rightly extends
to ‘the utility of a given commandment in a general way, not an
examination of the particulars’.20

The question remains: in a section devoted to a discussion of the

(general) intelligibility of the laws, why should Maimonides wish
also to emphasize their residual want of reason? As noted, one
reason is to forestall hyperrationalism, the desire to nd reasons for
every particular. Another reason, a possible consequence of the
foregoing, is Maimonides’ desire to forestall (possible)
antinomianism, that is, nonperformance of the commandments.21
Although there is a deep need and even duty for those able to
inquire into the reasons for the commandments to do so, there is an
equal need, Maimonides seems to suggest, to recognize limits to the
inquiry. Not to recognize limits is to court antinomianism. In the
case of the non-philosophical masses, an unbounded inquiry into
reasons might well lead to confusion and hence to a failure of belief
and a consequent disregard of the commandments. For the
philosophical elite too, the inquiry might well culminate in
antinomianism, for in not discovering a reason for some particular,
the intellectual might query God’s wisdom and, as a result, withhold
his assent to the commandment. For these reasons, then, there is a
need for Maimonides to state explicitly that the law is in part
inscrutable.
 When all is said and done, however, I think that there remains
another reason for Maimonides’ desire to emphasize that some of
the laws are in their particularity without reason, although still the
consequence of divine wisdom. Recall our earlier discussion
concerning the grounds for the distinction between generally
accepted laws and the huqqim (the ceremonial and ritual laws). We
noted the problematic mapping of this distinction onto Maimonides’
approving of the rabbinic praise of self-control, his desire in e ect to
allow for human nature and inclination. With the rabbis of old,
Maimonides urges an awareness that certain human inclinations are
natural and, although to be overcome by obedience to the law, not
to be scorned. This doctrine found in the (early) Shemonah Peraqim
is not entirely absent in the (late) Guide, even though the Guide
presents itself as a considerably more ascetic, immoderate, work
than the earlier one.22

A relatively modest point buttresses Maimonides’ critique of

hyperrationalism and his correlative struggle against all forms of
antinomianism. Human beings need the law to achieve certain
political as well as spiritual and intellectual ends.23 The law
provides a teleological framework for the good life, but a good
human life, one in accord with human nature. I suggest that the non-
teleological aspects of details of particular commandments, their
‘matter’, are highlighted by Maimonides as paradigmatic of our
corporeality, that part of us that is ‘natural’ and cannot be brought
under the rule of reason. As matter provides limiting conditions on
the attainment of the ideal, so the huqqim, speci cally the
particulars of these commandments, in their indeterminateness and
recalcitrance to the otherwise teleological nature of the law, are
emblematic of our nite nature, even as they are no limitation upon
God’s.

Thus, there is in Maimonides’ discussion of the reasons for the

commandments a deep realization that we are, after all, merely
human. With the limits of human nature in mind, we must also
understand his remarks about the limits of the inquiry concerning
the grounds for the commandments. While Maimonides is second to
none in urging those capable to ground their actions in wisdom, he
is equally concerned to point out that some things just are: they
have no reason. And we should act accordingly.

SOME GENERAL (AND TENTATIVE) CONCLUSIONS ABOUT JEWISH

PHILOSOPHICAL THEOLOGY

Are there any substantive conclusions we can draw from this brief
excursus through the thoughts of some major Jewish philosophers?
One conclusion might be a healthy revision of the very notion of
philosophical theology. I think that the subject is generally
conceived to be a highly theoretical one, having little to do with
actions and living one’s life. The (philosophical) theologian is the
person who wonders about the nature of God, divine power and
knowledge, the creation (or eternity) of the world, and the
implications of this on divine freedom, etc. The list goes on and on,
and there can be no doubt that the medieval Jewish philosophers
from Saadya onwards address these issues. But perhaps what is
overlooked is how these ‘theoretical’ discussions appear to support
some rather practical, moral, and even political considerations. At
the very beginning of this chapter, we noted Mendelssohn’s claim
that the only commandments that Judaism knows are commands to
act and to refrain. He stresses as strongly as he can, and for the
apologetic reasons we noted, that Judaism has no ‘articles of faith’,
no commands to believe.

From this ‘orthopraxic’ perspective we can at least draw the

conclusion that un-mediated belief, ‘orthodoxy’ in a very literal
sense, will have little purchase. And in turning to two giants of
Jewish philosophical theology, Saadya and Maimonides, we noted
that for them, the most important intellectual task, for those able to
engage in it, is to ground the requisite beliefs in reason. This is
nicely captured by Saadya, who indicates that the goal of his
philosophical magnum opus is to lead one from unre ective to
grounded belief. Saadya, and Maimonides as well, do not wish to
impose ‘correct’ belief, for just the very reason that Mendelssohn
suggests. It cannot be done. Thus, the command laid upon one
wishing to achieve the summum bonum is to engage in natural
philosophy (science), the study of the divine creation, and to gain
from this endeavor such knowledge of God that it is possible to
obtain. And, as we have seen explicitly in the case of Maimonides’
philosophy of law, there is an injunction to ascertain the reasons for
the commandments, at least in their general form. The telos is not to
follow the law blindly, any more than one is commanded to believe
that the world is created, but rather to ponder the very grounds for
the proposition in question.

The attainment of this wisdom has itself practical rami cations.

In both Saadya’s and Maimonides’ philosophical works, the focus on
particular ‘theoretical’ points in semantic theory, epistemology,
cosmology, prophetology, legal theory, and so forth tends to obscure
the overall non-theoretical telos of their respective works. Consider
the Guide. However conventional the literary form of the work is, an
epistle to a beloved student, troubled by a deep existential crisis,
one must take the topos seriously. Maimonides is not writing for
himself, nor is the Guide to be understood as a patchwork of
theoretical minitreatises on the aforementioned topics. This is not in
the least to de ate or to overlook the brilliance of Maimonides’
contributions to a variety of deep and di cult philosophical and
scienti c topics, but rather to contextualize those discussions in the
appropriate way. Everything in the Guide subserves the end of
showing the addressee that, properly understood, his religion—his
traditional way of life, one circumscribed by halakhic (legal) norms
—is philosophically defensible. And this latter point is not a
theoretical one, o ered for its own sake, but manifestly a moral and
even political one. In responding to the addressee’s worries, his
perplexity, the Guide repositions him squarely within the
community, from which as a deeply troubled ‘intellectual’ he is
temporarily estranged.

The very order of presentation of philosophical topics in the

Guide is another way one may see the overall (practical) trajectory
of the project. The Guide is a tripartite work, commencing with
philosophy of (divine) language and a nitist epistemology
consistent with that semantics, followed by cosmology,
philosophical psychology, and metaphysics, and nally concluding
in the third part with legal theory and moral and political
philosophy. This very procession maps on precisely to the ordering
of the edited Aristotelian corpus, and as importantly to the practical
thrust of (post-Aristotelian) Hellenistic philosophy generally. From
Epicurus to Sextus Empiricus, the point of doing science and
philosophy is to achieve a state of repose. In its own way Spinoza’s
Ethics is a part of this grand tradition of engaging in deeply di cult
study for the sake of happiness.

The suggestion here is not that the Jewish philosophical

theologians we have been interrogating are aiming at ataraxia
(unperturbedness), but rather that their ‘theoretical’ discussions
subserve practice. The project of understanding the reasons for the
commandments is undertaken so as to better ground a communal
life, much as Aristotle’s Ethics is o ered in the service of one well on
her way to e ective citizenship. Further, I think that if we read
sensitively between the lines of Maimonides’ celebrated discussion
of divine attributes in the rst part of the Guide, we shall discover
that his semantic theory is o ered inter alia to de ate a certain
impetuosity, a naive epistemological optimism about the (unlimited)
scope and powers of human knowledge. Such ‘pride’, intellectual
presumption, ends in perplexity. For a full picture of what
Maimonides is up to, the delineation of the bounds and limits of
human wisdom (and laying out the semantic correlate to this
epistemological nitism) must be harnessed to the aretaic point
about character reformation. In sum, for Maimonides, the doctrine
of divine attributes entails a certain humility as its desideratum.

So, it can be seen, at least on the basis of the few authors we

have discussed, that Jewish philosophical theology muddies the
grand dichotomy of theory and practice. Study is commanded for
the sake of moral and social reform. And the law itself becomes
most e ective in a human life when obedience follows on
consideration of its grounds.

NOTES

Primary source material for Jewish philosophical theology,

including philosophical exegesis of Scripture, may be found in:
D. Frank, O. Leaman, and C. Manekin (eds.), The Jewish
Philosophy Reader (London: Routledge, 2000); M. Walzer, M.
Lorberbaum, N. Zohar (eds.), The Jewish Political Tradition
(New Haven: Yale University Press, 2003–); C. Manekin (ed.),
Medieval Jewish Philosophical Writings (Cambridge: Cambridge
University Press, 2007). Further secondary reading in Jewish
 philosophical theology, and its history, include: C. Sirat, A
History of Jewish Philosophy in the Middle Ages (2nd edn.)
(Cambridge: Cambridge University Press, 1990); D. Frank and
O. Leaman (eds.), History of Jewish Philosophy (London:
Routledge, 1997); S. Nadler and T. Rudavsky (eds.), The
Cambridge History of Jewish Philosophy: From Antiquity through
the Seventeenth Century (Cambridge: Cambridge University
Press, forthcoming); D. Frank and O. Leaman (eds.), The
Cambridge Companion to Medieval Jewish Philosophy
(Cambridge: Cambridge University Press, 2003); K. Seeskin
(ed.), The Cambridge Companion to Maimonides (Cambridge:
Cambridge University Press, 2005); M. Morgan and P. Gordon
(eds.), The Cambridge Companion to Modern Jewish Philosophy
(Cambridge: Cambridge University Press, 2007).

1. This view is most prominently associated with the neo-Kantian,

Hermann Cohen, whose major Jewish philosophical treatise is
Religion of Reason out of the Sources of Judaism (1919).

2. Jerusalem, or On Religious Power and Judaism, trans. A. Arkush

(in D. Frank. O. Leaman, and C. Manekin (eds.), The Jewish
Philosophy Reader (London: Routledge, 2000), 349).

3. Saadya Gaon, The Book of Doctrines and Beliefs, abridged edn.

A. Altmann, with new introduction by D. Frank (Indianapolis:
Hackett, 2002).

4. R. Brody, The Geonim of Babylonia and the Shaping of Medieval

Jewish Culture (New Haven: Yale University Press, 1998), 235
 .

5. Saadya Gaon, Book of Doctrines, 20.

6. Ibid. 30.

7. As Tom Flint reminds me, the trajectory of Saadya’s project

stands comparison with Anselm’s a century later.

8. Maimonides, The Guide of the Perplexed, abridged edn. J.

Guttmann, trans. C. Rabin, with new introduction by D. Frank
(Indianapolis: Hackett, 1995), 41.

9. See Final note.

10. A comprehensive treatment is found in M. Kellner, Dogma in

Medieval Jewish Thought: From Maimonides to Abravanel
(Oxford: Littman Library of Jewish Civilization, 1986).

11. Guide 3. 51.

12. M. Kellner (trans.), Isaac Abrabanel, Rosh Amanah [Principles of

Faith] (Oxford: Littman Library of Jewish Civilization, 1982),
79.

13. Nicomachean Ethics [NE] 1. 8 (1099a7–12); 1. 13 (1102b27); 2.

3 (1104b3–8); 7. 9 (1151b32–1152a6).

14. Guide 3. 26.

15. Sifra to Leviticus 20: 26.

16. Guide 3. 8; 3. 33.

17. Ibid. 3. 26.

18. Ibid. 3. 49: ‘The fact that there are particulars the reason for
which is hidden from me and the utility of which I do not
understand, is due to the circumstance that things known by
hearsay are not like things one has seen …’

19. Ibid.3. 26. For a possible source for this latter point, see NE 5.
7 (1134b20–3).

20. Ibid.

21. For a full discussion of this point, see J. Stern, ‘The Idea of a
Hoq in Maimonides’ Explanation of the Law’, in S. Pines and Y.
Yovel (eds.), Maimonides and Philosophy (Dordrecht: Nijho ,
1986), 92–130.

22. For remarks on the reasons for this ‘developmental’ scheme in

Maimonides’ ethical theorizing, see H. Davidson, ‘The Middle
Way in Maimonides’ Ethics’, Proceedings of the American
Academy for Jewish Research 54 (1987), 31–72;D. Frank,
‘Anger as a Vice: A Maimonidean Critique of Aristotle’s Ethics’,
History of Philosophy Quarterly 7 (1990), 269–81.

23. Guide 2. 40 Final note: In large measure Jewish philosophical

theology presumes the ‘philosophicality’ of Scripture, the view
that Scripture is amenable to philosophical reconstruction. A
good way to enter this debate is to read Spinoza’s stinging
critique of the philosophical-theological project (Spinoza,
Tractatus Theologico-Politicus, chs.1–7), followed by
Maimonides’ defense of it in The Guide of the Perplexed.
 CHAPTER 25

ISLAMIC PHILOSOPHICAL THEOLOGY

OLIVER LEAMAN

ISLAM as a distinct religion started in the seventh century CE when

the Prophet Muhammad received a series of messages from God,
according to Islam. He in turn communicated those messages to the
local Arabs in his part of the Arabian peninsula and after some
di cult periods managed to establish a community of believers, the
Muslims. They rapidly went on to conquer the neighboring areas,
and converted the local inhabitants to the new religion. Islam,
however, does not really see itself as a new religion, but as a
culmination of earlier religions, based as they are on earlier
revelations.

As Islam spread, there was growing contact with centers of

civilization in the Arab world and Persia that were in one way or
another imbued with Greek culture. This led to a erce reaction by
many Muslims, rejecting that foreign and unbelieving culture for the
more familiar and authentically Islamic sciences of grammar, law,
the traditional sayings of the Prophet and his Companions, and
theology. Yet the very real material advances of Greek culture, in
particular in medicine and science, impressed many Muslims, and
there was evidently much enthusiasm for what could be learned
from the developed cultures that were rapidly incorporated into the
Islamic empire. The caliph al-Ma’mun, who reigned from 813 to 833
CE, was a signi cant advocate of Greek thought, and founded an
institution in Baghdad whose main purpose was translating Greek
texts into Arabic. Often these translations came about through
translation of a Greek text rst into Syriac, and then into Arabic.
Early translations included works of Aristotle, commentaries on him,
summaries of many of Plato’s dialogues, and later Greek
elaborations of their work, and many works in the standard
Neoplatonic curriculum.

This translation project was itself controversial, with some

arguing that logic is only a tool employed by philosophy, and has no
other implications. Other thinkers hostile to philosophy argued that
logic is inevitably contaminated with it and so should not be used.
Al-Ghazali (AH 450–505/1058–1111 ce), a steadfast critic of
philosophy, displayed his independence by both accepting logic as a
useful technique and rejecting the philosophy then current which
based itself very much on that logic. He rmly took the view that
logic is independent of philosophy and could, in fact, be derived
from scripture itself rather than a Greek source. There was also a
controversy about whether logic is independent of language,
obviously a not unrelated issue. Right at the start of Islamic
philosophy there was a celebrated debate in Baghdad between al-
Sira and Ibn Bishr Matta on whether knowing the grammar of a
particular language is more useful than knowing logic that takes
itself to form the foundation of all language itself. By the nineteenth
century Peripatetic philosophy had become part of the Arabic Nahda
or Renaissance, often seen as a symbol of modernity and the
growing identi cation with the non-Islamic world. After all, the
philosophical thought of Islam found a signi cant role in Europe in
the medieval period, and is often argued to have played an
important part in the European explosion of science and intellectual
thought in general. The works of Ibn Rushd (Averroes) and Ibn Sina
(Avicenna) were in the curriculum of the Christian and Jewish
worlds for many centuries, there were many translations of even the
same text and they seem to have played a signi cant role in the
development of philosophy and theology in Europe as a whole.

There are three main schools of thought in Islamic philosophy,

and they all have theological consequences.

PERIPATETICISM (FALSAFA)

Peripateticism or mashsha’i philosophy in the Islamic world is very

much based on Greek thought, and in particular Neoplatonism. This
was really established at the time of al-Kindi (d. 252/866) and is
often said to have come to an end with Ibn Rushd, who represented
the height of Peripatetic thought in al-Andalus, the Iberian
peninsula under Muslim rule.
 Al-Farabi (257–380/870–950), commonly referred to as the
second teacher (after the rst in signi cance, Aristotle, or
sometimes Ibn Sina), played a de ning role in establishing
Peripatetic philosophy in the Islamic intellectual curriculum. He
identi ed philosophy with the deep grammar of language, and as
decisive in dealing with intellectual issues as compared with the
other Islamic sciences. This led to two signi cant problems. It
downplayed the role of the religious scholars in the competition for
who had the most powerful conceptual equipment, always a potent
issue in religious thought. The other was that philosophy threw up
conclusions that are apparently in opposition to accepted
understandings of Islam within the community of believers. We shall
see how al-Ghazali came to list precisely these conclusions and
challenge the role of philosophy accordingly.

Ibn Rushd (520–95/1126–98) was unusual for his rejection of

mysticism, which by contrast was something that was adopted by
most of the other Peripatetic thinkers. He also rejected much
Neoplatonism, which he clearly saw as incompatible with
Aristotelian thought, and attempted to represent in his philosophy a
more genuine form of Aristotelianism. He had few followers in the
west of the Islamic world. The attack his style of philosophy had
received at the hands of al-Ghazali had swung the intellectual
debate away from philosophy in the style of the Peripatetics towards
a more theologically centered form of work. It is di cult to know
whether al-Ghazali’s work was really so in uential, because it did
not have much e ect on the survival of Peripateticism in the Persian
world, where that style of Peripatetic thought continued to play an
important part in the philosophical curriculum. This was despite the
fact that al-Ghazali, like so many Islamic philosophers, was Persian
and was held in high regard in general in philosophical circles in the
Persian world.

Ibn Rushd responded to al-Ghazali’s Incoherence of the

Philosophers with his Incoherence of the Incoherence, but there can be
no doubting that al-Ghazali managed to identify some very di cult
issues for the Peripatetic tradition. There are three propositions on
which the Peripatetics tend to agree, and they all constitute novel
and inappropriate views from an Islamic point of view, according to
al-Ghazali. They are all also invalid, he argues, even if one accepts
their premises and their philosophical principles. We shall return to
this point, but rst let us look brie y at what he called the heretical
and logically invalid views, and to which he took such exception.

The rst is the eternity of the world, a thesis apparently found in

Aristotle and generally shared by the philosophers. The world is
eternal because time is a function of motion, in that if there were no
motion, the concept of time would not make sense. Before the world
existed there was no motion and so no time, and so the world could
not have been created at a particular time. For a similar reason it
could not be destroyed, since then motion would cease and so
would time, and there cannot be a last time. The picture of creation
was very much one of emanation, where a constant over owing of
grace from the divine produces lower forms of existence, the earth
being the lowest.

The second thesis al-Ghazali highlights is that of the immortality

of the soul, and only the soul, rather than the soul and body. The
sort of immortality that the philosophers tended to defend was
spiritual and far from the religious account in the Qur’an and the
hadith. Philosophers had di culty in seeing how the matter of
bodies could be reconstituted after death, whereas the mind, with its
links to abstract issues and higher realms of being, could be
understood to have some form of immortality. This is because of the
principle that when one knows something, the knower, the known
object and the act of knowing are all the same, a view the Islamic
philosophers attributed to Aristotle. So if one comes to know
something that is permanent and unchanging, that part of one that
does the knowing might also lay a claim to share in its immortality.

The third thesis that al-Ghazali found so threatening to Islam is

the idea that God cannot know the ordinary events that take place
in the world. This means, as he says, that God would not know that
Muhammad prophesied, or whether one deserved to be rewarded or
otherwise in the next world. The thesis also goes right against a
constant theme of the Qur’an, that God knows everything that
happens. For the philosophers God’s knowledge of ordinary events is
a problem because if he knew of changing things, then this means
he changes, or at least his thoughts do, and this goes against the
idea that he is unchanging. It also makes him too much like us, in
having something like a sensory system. After all, how could he see
what happens in the world if he has no eyes? Now, of course they
suggest that a more sophisticated understanding of what it is for
God to know needs to be adopted to understand the nature of divine
knowledge, something that al-Ghazali seems to be denying. When
the Qur’an talks of God knowing what happens in the world, the
philosophers argued, the meaning is not that he knows in much the
same way that we know, but rather that he knows in a special and
unique sort of way. But al-Ghazali is surely quite correct in thinking
that there has to be in any system of thought that calls itself Islamic
some access of God to the world of generation and corruption. The
attack on the Peripatetics’ account of divine knowledge is perhaps
his weakest criticism, though, since they of course accept that God
knows what happens in the world, albeit via a more sophisticated
route. Al-Ghazali also advocates a more sophisticated route to
interpreting these three basic aspects of religion, and so can hardly
complain that the philosophers suggest something similar. He also
accepts that one should not always understand literally what one
reads in the Qur’an, and the anthropomorphic language to be found
there should not be taken to suggest that God is really like us. What
is really at issue is two distinct approaches to understanding how to
make sense of the Qur’an, and what implications those di erent
approaches possess.

One of al-Ghazali’s main targets is Ibn Sina (370–428/980–

1037), who formulated the main features of philosophy in the
Islamic world during the early centuries of the discipline. He
presented the Neoplatonic line, albeit with many variations, and
there are features of Neoplatonism that t in very easily with Islam.
The emphasis on the One, for example, links up nicely with the
monotheism of Islam. Islam emphasizes throughout the signi cance
of belief in one deity, someone who must not be linked with
partners, a being who is the One from which everything else comes.
For the Neoplatonists the chief problem is explaining how there
came to be many things in the world, given that only One thing
exists originally entirely by itself. They provide a theory in
accordance with which it thinks about itself, since there is nothing
else worth thinking about as far as it is concerned (anything else
being less perfect), and it is this that gets us to the existence,
eventually, of many other things. What keeps the universe in
existence is the constant thought that emanates from the One and
has the various e ects lower down the ontological scale that
eventuate in our world and in us. It is not di cult to see how the
One could be linked with the God of Islam, nor how this sort of
theory can be taken to express our links with what is higher than us
in a perfectly acceptable way from a religious point of view. God is
connected with everything that exists through producing thought
that subsequently produces everything else. So he is connected to
the world, but in a way that does not interfere with his perfect unity
and independence of that world. Variations of this Neoplatonic
model were employed by Islamic thinkers for a long time by those
working within all traditions of philosophy, and retained their role,
as we shall see, in ishraqi and Su thought right up to now, at least
to a degree. Ishraqi thought is based on the concept of light and uses
this as the leading idea on which to base its metaphysics, while
Su sm is Islamic mysticism. They both see themselves as
transcending Peripatetic thought. The notion of divine grace
emanating through the levels of being is rich in its implications for
explaining the relationship between a transcendent deity who
nonetheless is seen as involved in our level of existence in some
way, while also beyond that existence and entirely independent of
our world.

MYSTICISM

Mysticism or Su sm in Islamic philosophy is employed as a

philosophical technique very generally. Some thinkers such as Ibn
Sab‘in (614–69/1217–70) even went to the lengths of arguing that
philosophy that was analytical could not be of value since reality is
basically one, and dividing it up to examine it bit by bit is to
misunderstand it profoundly. (Of course, this itself is an analytical
argument.) The mystical thinker Ibn al-‘Arabi (560–638/1165–
1240) pursued this line and represented himself as burying the old
Peripatetic form of thought when he carried the bones of Ibn Rushd
back to al-Andalus for burial on the back of a donkey. But most
philosophers combined mysticism with Peripatetic philosophy,
arguing that these are just two di erent philosophical
methodologies, and mysticism goes deeper into the nature of reality.
 The application of Su sm to philosophical issues can be seen
today in the work of Seyyed Hossein Nasr. He argues that mysticism
represents a constant thread in philosophy and religion and a way of
thinking we need to recapture if we are going to be able to
appreciate the world as a spiritual as well as material realm. The
trouble with analytical thought, he argues, is that it denies the role
of God in the world, and sees it primarily as a material realm to be
understood by science. Nasr uses traditional ideas based on Su sm
to discuss such issues as ecology, ethics, and the nature of
philosophy itself.

ILLUMINATIONISM

Illuminationist (ishraqi) thought comes from the term ishraq, which

is linked with the idea of the East, and represents something of a
combination of Peripatetic thought and mysticism. One of the
limitations they thought existed in Aristotelian ways of doing
philosophy is its tendency to break down concepts into simpler
components as a way of understanding them better. The ishraqis
argued against the principle that reasoning starts with de nition in
terms of genus and di erentia, a process of explaining something by
reducing it to its smaller parts and so de ning the concept at issue.
Illuminationist thinkers such as al-Suhrawardi (549–87/1154–91)
argue that this is to explain the unknown in terms of something
even less known. That is, if one does not understand the concept
itself, then there is no reason to think that one will understand its
constituent parts any better, at least until they are in turn broken
down into yet smaller parts and so on ad in nitum. They also seek
to demote deductive knowledge, the sort of knowledge one gets
from using the principles of syllogistic reasoning, with knowledge
by presence, which is knowledge so immediate that it cannot be
doubted. That is, the radiance in one’s mind from the item of
knowledge is so strong that one cannot doubt it. An example often
given is of one’s awareness of the self that underlies all one’s other
experience. One cannot doubt that it is there since it makes such a
strong claim on one’s attention. Light comes into the picture since
the idea is that such knowledge is lit up in a way that makes it
impossible to doubt or ignore, and this is because light ows
through the universe and brings to existence and awareness a range
of levels of being. The di erences between things can be described
in degrees of luminosity or light, not in terms of their essences. God
is often aligned with the Light of Lights, the light that is the source
of all other light and that does not itself receive light, rather like
Aristotle’s unmoved mover, that other things move around but does
not itself move. This form of philosophy was particularly popular in
the Persian world, where it was combined in varying degrees with
aspects of mysticism and even Peripateticism. Mulla Sadra (979–
1050/1572–1640), whose thought has dominated the Persian
philosophical curriculum since his time, used all three forms of
philosophy in his writings, as became very much the style of much
Persian philosophy.
 What were the leading issues of Islamic philosophical theology?
Here we will concentrate on the Peripatetic tradition, since it is
clearest what its implications are for issues in religion, and it often
established the basic presuppositions of Su sm and Illuminationist
thought also.

ETHICS

There was an involved debate about the objectivity or subjectivity

of ethical rules, in particular over the issue of whether an action is
just if and only if God says that it is just, or whether it is just in
itself, both issues that arise in similar forms in Islamic theology. The
Mu‘tazilites argued that God’s actions are based on an objective
notion of justice, and so God has to act to reward the innocent and
punish the evil. Their opponents, the Ash‘arites, argued that it is a
mistake to insist that God has to act in any way whatsoever, and
that whatever he does is right. It is right because he does it, not
because he follows some objective rules of morality. This point
comes out in a very nice story by al-Juwayni (419–78/1028–85)
which is often used to show the shallowness of the Mu‘tazilite
theory. This involves three people who die. One is evil and goes to
hell, one would have become evil, so God kills him as a child and he
goes to the barzakh, the realm in between heaven and hell, so
avoiding hell, while another who has lived well goes to heaven. The
child complains that he missed the opportunity of going to heaven
by dying early, although God informs him that had he lived he
would have been evil and gone to hell. The sinner in hell complains
that God should have killed him while a child so that at least he
would have avoided hell and earned the barzakh. This example is
often taken to show that the idea that there is only one possible idea
of justice, that expounded by the Mu‘tazilites, and that it represents
the general principle on which God has to operate, is vacuous. God
could decide to intervene directly in our lives to prevent us from
developing in the ways in which we otherwise would do, but he also
might not. That is true of us also; we can often intervene to help or
hinder others, and we sometimes do and sometimes do not. There
are reasons for what we do, but those reasons do not have to be
seen as determining the action.

POLITICS

How to reconcile the social virtues that arise through living in a

community and the intellectual virtues that tend to involve a more
solitary lifestyle was a deeply felt issue, since so many philosophers
had di culties in being accepted by their local communities,
especially by the political authorities. Their means of expressing
themselves often set them against the ordinary believer. Why should
the philosophical thinker who can grasp the truth primarily through
the use of reason be part of the social and religious activities of the
community? They often argued that philosophy represents the truth
in a pure form, whereas religion is a working of that truth in a way
that makes it digestible and comprehensible to the community in
general. The implication might be that the philosophers really
regarded religion as signi cant only for the masses and not for the
intellectual elite, and so were not sincere in their apparent
adherence to religion. During a period when religious and social
identity were so closely connected, this gave rise to questions about
the real religious beliefs of the philosophers, and their sincerity was
often questioned. Although the philosophers were undoubtedly keen
for prudential reasons to link up with the practices of the
community, this is clearly a di erent form of attachment than was
normal, and gave rise sometimes to suspicion concerning their real
beliefs.

Political philosophy looked to Greek thinkers for ways of

discussing the nature of the state, and often combined Aristotelian
and Platonic political ideas with notions from the Qur’an and other
Islamic texts. This was not di cult since it enabled them to argue
that the state ought to be concerned with both the material and the
spiritual welfare of its inhabitants. The philosopher would be the
best ruler because he could understand dispassionately what is in
the general interest, and ensure that religion is used to teach the
community in general how to behave so that its welfare is
enhanced. This sort of elitism was widely shared by a range of
philosophies. Even Su and Illuminationist thinkers were largely of
the opinion that only a limited group of people could understand
precisely how the state ought to operate, and that traditional
religion was an important source of information for the people as a
whole.

THE SOUL

The nature of the soul, the thinking part of human beings, was a
particularly important issue from a religious point of view. Many
Peripatetic thinkers followed Aristotle in regarding the soul as the
form of a person, as the way in which we are organized and not
something distinct that is part of us. This implies that once the body
or matter dies, the soul or form of the matter no longer exists since
it then has nothing to inform. Yet Islam has a strong notion of an
afterlife, graphically described in the Qur’an and the traditional
sayings as a physical afterlife, and the soul and body would in some
sense be regarded as eternal in that afterlife. It might be suggested
that the Qur’anic view is largely allegorical, so that our actions in
this life have consequences that extend further than this life, and a
good way of communicating this to people who think that matter is
important is through talking about having an afterlife in the sense of
eternal souls. Other thinkers tended to use a Platonic account of the
soul as something eternal and immaterial, and this also seems to
contradict the Qur’anic account of the afterlife as a physical sort of
place, as heaven or hell or what lies in between. Many of the
Peripatetics (falasifa) argue that the religious language is physical
because for most people that is what is important in their lives. It is
a way of explaining to everyone why it is important for us to behave
well, whereas a more spiritual grasp of the links between this world
and the next one is achievable only by a few intellectuals or
spiritually advanced individuals, and should not be widely
broadcast.

LOGIC

Logic became a particularly lively topic of debate between those

who saw it as just a tool to be used, and others who regarded it as
bringing an ideology with it. For the former, even poetry was taken
to have some sort of a logical structure, since poetry is writing that
is expected to have a conclusion, perhaps the eliciting of an
emotion, and it sets out to achieve this conclusion by a careful and
logical organization of language. In fact, each type of writing has a
logical structure that describes how it is supposed to operate, and
what the appropriate rules are. These rules will di er for each area.
The logic of theology is dialectical, for example, and takes a
particular proposition as true, because it occurs in a text that is
accepted as true, and works out what the logical implications of that
text are. But since the text could be wrong, it is only accepted
within a particular context; the status of the conclusion that one
derives is always a bit dubious. Law and many of the other Islamic
sciences are dialectical in this sense, and philosophers often have an
account of knowledge that does not include those sciences as
representing the very highest type of knowledge. The main
di erence between philosophers and ordinary people is that
philosophers are capable of understanding the structure of these
arguments better than everyone else. Ordinary believers tend to be
very in uenced by imagination and their senses; they are emotional
and reliant on tradition, so they are not really able to understand
arguments that come from those who are able to see behind
ordinary experience to what exists at a deeper level, analytically or
spiritually, or indeed both. The philosophers are unique in having
access to what might be called the gold standard of argument, the
demonstrative syllogism, where one proceeds logically from
de nitions to entirely certain conclusions. The rest of the
community, including the religious scholars, only have a vague
grasp of how reasoning really operates.

THE DOUBLE TRUTH ISSUE

One of the shocking implications that was derived from Ibn Rushd
in Christian Europe was the idea that religion and reason are
entirely separate areas of thought. A proposition may be true in one
and false in another, a view often attributed to Ibn Rushd. Although
he certainly did not argue in this way, he did suggest that there are
di erent routes to the truth, to the same truth, and these routes may
be quite distinct from each other. This is an important topic for the
falasifa as a whole, since for them many issues and ideas can be seen
from a variety of perspectives, each of which represents where the
individual is coming from, each of which links up with the truth,
and all of which are di erent. For example, al-Farabi points out that
the religious understanding of the prophet is that he is someone of
excellent moral character chosen by God to transmit a message.
From the philosophical point of view he is someone of equally
sound moral character whose intellect is in line with the active
intellect and so knows how to persuade an audience of a particular
point of view. By the active intellect is meant the highest level of
conceptual thought that is accessible by human beings. The
philosopher and the believer both listen to the same message, that
provided by the prophet, and analyze what they hear di erently,
but both are right. The philosopher understands the rational basis of
the prophetic message, while the ordinary believer is impressed
with its emotional power. Does this mean that the former has a
better grasp of the meaning of the message than the latter?

It would be wrong to see the philosopher’s view as primarily

intellectual and abstract and the other as merely ‘user-friendly’,
since the sort of language that surrounds religious ideas such as
prophecy is hardly simple and perspicuous. Often theology is more
complex than philosophy, and the views of the ordinary believer
more complicated than those of the thinker who can organize his
thought into clear and distinct categories, or thinks he can. This was
very much the point of the falasifa, the Islamic philosophers in the
Peripatetic tradition, that each way of talking was valid, although
they surely also thought that the philosophical account was more
accurate than many of the alternatives. What is worth noting here is
that the idea of there being di erent routes to the truth takes
seriously the principle that the way we frame our beliefs ts in with
who we are and what we think we know, and so for di erent people
di erent ways of framing those beliefs are appropriate.

Some thinkers who could not express their thought in the

politically correct way paid the ultimate price, and the example of
Socrates and his fate at the hands of the Athenians had more than a
historical resonance for many in the Islamic world. Al-Farabi
comments that Socrates did not have signi cantly di erent views
from those of Aristotle, but Socrates was executed while Aristotle
remained in political favor throughout his life. The mistake that
Socrates made was to express himself clearly, in ways that anyone
could understand, and that led to people thinking that he was being
subversive. Aristotle prudently expressed his ideas in di cult and
complex language, so only those capable of really understanding his
presentation could come near to grasping his views.

QUR’ANIC LOGIC

It is di cult to say when philosophical and theological speculation

started in Islam. From a very early stage there are reports of the
Prophet, and those close to him, being questioned on issues of
interpretation, and these reports have been codi ed as the hadith or
traditions. The theoretical level of these remarks is not high,
although they are of great interest, and very helpful in xing how to
carry out actions in a way that ts in with Islam. They often relate
to the practice or sunna of the Prophet, and his actions are often
taken to be exemplars for Muslims, since although he was
undoubtedly a human being, as the nal prophet he is as close to
perfection as a human being can be and well worth copying.

But it is the Qur’an itself that is generally taken to be the main

guide for both behavior and speculation and the Qur’an frequently
describes itself as clear (16. 103) and also as wonderfully or
miraculously constructed. Indeed, the miraculous nature of the text
is often cited as a proof of Islam itself, in that a human being could
not have produced anything like it, and so it must have originated
with God. This claim is made quite often in the Book itself, and has
been frequently reproduced afterwards, and it is an interesting and
unusual claim for a religion to make about its central text. Islam
challenges those of other faiths and of none to show how the Book
could have come about except directly from God.

The other claims made in the Book describe our relationship

with God, what God has done for us, what he expects in return, how
the world started and how it will end, and the nature of various
beings such as angels, jinn, human beings, and other natural
creatures. These claims raise theoretical di culties once we wonder
how to reconcile them with aspects of theory coming from
elsewhere, such as Greek philosophy, and we have looked already at
some of these debates. It is worth pointing out that in the Qur’an
itself there is a heavy emphasis on the signi cance of knowledge
and a constant demand that the hearer and reader consider and
ponder over what he or she nds in the Book. Islam makes a range
of theoretical claims backed up by comments made in the Book and
elsewhere, although whether the comments are supposed to prove
demonstratively that those claims are true is another matter. Like all
major religious books, the Qur’an is not primarily a theoretical
work, although it does contain ideas and arguments that are capable
of and have undergone theoretical elaboration.

Some of these Qur’anic ideas very quickly became contentious.

The rst concerns the nature of the succession to the Prophet as
head of the community. The debate over succession led to the split
that has persisted ever since between the Shi‘a (shi‘at ‘Ali, party of
‘Ali), who thought that the leadership should be restricted to the
family of the Prophet, and the Sunnis, who argue that a much wider
category of candidacy for leadership could be considered. Like a lot
of theoretical arguments, this has considerable practical
rami cations, and determines who the political authority should be
and how this sort of issue should be determined. The Shi‘ites have
continued to look to leadership from those closely linked with the
descendants of the Prophet, and the succession of imams or spiritual
leaders who are linked with ‘Ali. The majority Sunnis on the
contrary regard as the appropriate authorities those selected in some
other way, by a consensus of those quali ed to judge, perhaps, or
through some political process, depending very much on the
particular school of Sunni thought that is accepted.

Another theoretical debate that swiftly arose was over the nature
of determinism and freedom. Are human beings free to act, or are
their actions determined by God? Both sorts of language are found
in the Qur’an and the remarks of the Prophet and his Companions,
but the nal position is a matter of debate. This is a good example
of a theological debate that quickly became philosophical, in that it
started by looking at scriptural passages and trying to understand
their implications, and ended up by exploring the rational
foundations of each position. The determinist can nd plenty of
Qur’anic passages that suggest that God makes everything happen,
even the fact whether we believe in him or otherwise. This implies,
though, that one cannot be held responsible for one’s actions, yet
the Qur’an goes into long and highly descriptive detail of the day of
judgment and the afterlife of reward and punishment. How can this
life be a realm of testing if God determines everything that happens?
Yet the idea that people have free will has di cult implications
also, in that it could be taken to suggest that they are in charge of
things rather than God.

The nature of belief itself became a signi cant topic. What

makes a person a Muslim? What does he or she have to believe? Do
they have to do anything to provide evidence of that belief? Can a
sinner be a Muslim, or is such a person ruled out by virtue of
behavior? The early centuries of Islam saw the introduction of
creeds, statements that listed the basic beliefs that a Muslim has to
accept if he or she can really claim to be a Muslim. As one might
expect, these di er from each other and represent the con ict of
ideas that took place within the Islamic community over which
theological and political group best represented genuine Islam.

The role of theory in Islam has always been controversial, as in

other religions. There are on the one hand theoretical techniques for
resolving issues that come from within the Islamic tradition itself
and make up the Islamic sciences. These include Arabic grammar,
particularly important since the revelation was provided in Arabic.
If people are to understand the revelation, they have to know the
language and its grammar. Di culties of interpretation can be
resolved, sometimes through an investigation of language itself.
Then there is the analysis of the order in which revelations
appeared, since some later ones might abrogate the earlier, or be
linked with the occasion in which they were given for a more
comprehensive understanding of them. Some are Meccan, some
Medinan, some have passages from both places, and many are
related to particular historical events, and it is held to be important
to know the history to x the precise meaning of many of the verses.
The hadith or Traditions are helpful in providing even more history
on what important people at the time had to say on di cult issues.

TRADITION (TAQLID) AND ARGUMENT

There has been a long discussion in Islam about the signi cance of
following tradition or taqlid. Clearly any religion requires
interpretation, and there are some authorities that are likely to be
more reliable. They are the people to follow when it comes to
seeking an understanding of what the religion means and expects
one to do. In any area, it was argued, one should seek to follow the
expert, since if one knew what to do one would not need to ask an
expert in the rst place. But who is an expert? The Qur’an is full of
injunctions to its hearers and readers to consider, think, reason, and
that is presumably because Muslims are supposed to use their
rational faculties to work out what the text means and indeed what
everything means, insofar as they can. There are limits to what one
can work out oneself, however, and it is here that there is a need for
guidance.

Taqlid has an important role in that there are people who can
help one understand a text and its contemporary relevance if one is
unable to grasp this owing to a lack of knowledge of the whole book
and its context and other appropriate sources of information. It is
rational to follow authorities under such circumstances, and to do so
is not blind obedience; it is submitting oneself to the authority of
another when one has good grounds for doing so. Those good
grounds are established rationally.

The Qur’an is full of arguments, and many of them have as their

basis the idea that God, the ultimate creator, knows the nature of
everything in the world, and especially his human creatures, and
that he has brought about a type of creation that is speci cally
designed to be in the interests of his creatures. As one might expect,
the creator then establishes rules that he expects people to obey if
they are to get the best out of this world and the next. The text
constantly provides reasons for believing what it asserts. God wishes
to communicate with his creation, he does this through messengers,
and yet his creatures often reject what he has to say. This rejection
is not because of any doctrine of original sin, but because God has
made everyone di erently and so is not surprised at the di erences
in their responses to him. He could have made everyone the same,
and thus avoided this sort of diversity of opinion (2. 118; 10. 19; 16.
93) but chose not to do so. The di erences between people are
important learning opportunities, since in comparing ideas we learn
a great deal about our own beliefs.

Does the text really argue for the existence of God? Not very
seriously. God’s existence is taken to be an obvious and natural fact,
based on what we observe around us all the time, and nature, like
the verses of the Qur’an, is full of ‘signs (ayat) for people who
understand’ (2. 164). There is quite a bit of argument about why we
should believe in the existence of one God as opposed to many gods,
and why associating partners with God (shirk) is to be rejected
totally. There are also many arguments in the Book for resurrection,
a basic doctrine given the importance of rejecting materialism and
the belief that our deaths are followed by nothing more than the
collapse and decay of our bodies. These are based on the idea that
God can do anything, and on our experience of nature in which life
often follows death. The materialist who believes that all there is is
this life has no reason to believe in God and his message. There are
also many arguments for the prophethood of the Prophet himself,
and much of this clearly replicates the sort of to and fro of debate
that took place in Mecca and Medinah when Islam was becoming
established in the early years. The issue of who should succeed the
Prophet, the source of the Sunni/Shi‘i split, is of course discussed
widely.

The range of arguments to be found in theology vary in form.

The Peripatetic philosophers, the falasifa such as Ibn Rushd and his
predecessors such as al-Farabi, used the organon or system of
Aristotle as an example of how such arguments may vary in form
and in particular in the general nature of their conclusions. It was
widely accepted that there is a hierarchy of argumentative strength,
with demonstration at the top of the scale, where one operates with
true premises and uses them to arrive at validly derived and entirely
general conclusions. Then comes dialectic, where the premises one
uses are those supplied by the side with which one is debating, and
so one has no independent reason to think they are true. After this a
number of other argument types follow, rhetoric and poetry, for
example, where the point is to change people’s minds by the use of
imagination and appeals to the emotions, and where the validity of
the conclusion may be quite limited, restricted to a particular
audience within a speci c context. That does not mean that there is
anything wrong with the argument. It is a perfectly acceptable
argument of its type, but not up to the rigor of demonstration or
even dialectic. Philosophy was identi ed with demonstration, while
theology was more closely linked with dialectic, since it is based on
a book that is accepted by a particular community but not
necessarily universally. Religion for ordinary people follows weaker
rules of reasoning such as rhetoric and poetry, since it is designed to
move people emotionally, and does not rely on advanced reasoning
skills to be understood. The point of religion is to communicate with
everyone and so although the message should be based on sound
reasoning, the particular form that the general language takes will
be able to stimulate the imagination.

REASONS AND CAUSES

The fact that reasons may explain but not determine subsequent
actions was not accepted by al-Ghazali, advocate though he was of
the use of logic in theology and of its incorporation as a natural part
of Islamic thought. Al-Ghazali criticized thinkers such as Miskawayh
(320–421/932–1030) for believing that God had reasons to establish
certain rituals and rules that t in with human nature. Miskawayh
suggests that many of the rites of Islam have the purpose of
strengthening the links between believers, so that religion uses
social norms to encourage and strengthen religious observance. Al-
Ghazali argues on the contrary that God institutes rules just because
that is what he wants to do; there is no necessity for him to align
himself with our social instincts, nor even to employ them. Where
al-Ghazali goes awry here is thinking that because God may have
had reasons for what he did, those reasons would make his actions
governed by them. He would be forced by the reasons to act in a
particular way. The Mu‘tazilites did in fact argue in this way that
God has to follow certain principles of morality in his behavior; he
has to have justice in mind when he acts. Al-Ghazali rejected that
theory, arguing plausibly that it is an error to think that God must
have a particular purpose in mind that we can understand when he
does something. But he went too far in suggesting that God could
have nothing in mind when he does something. For al-Ghazali, all
God wants to do in telling us how to act is demonstrate his power
and the necessity on our part to obey him. However, having reasons
for action only constrains us if they inevitably result in particular
actions, but God cannot be forced to act in this way, and nor can
we.

One might think that the case for God would be di erent here
than it is for us. We are nite and changing creatures with a partial
view of the world and ourselves. None of this is true of God; he is
perfect, in nite and understands everything. Yet for him also
reasons do not determine. There are many things he can do and
might wish to do given the same set of facts. Rationality and
bene cence do not inevitably lead to particular actions. When the
angels are told by God to bow down to Adam (2. 30–4) and they
complain saying that Adam and his descendants will wreck the
earth if they get power over it, the angels have a point. But God
replies that he understands this and will send human beings a guide.
The notion of a guide is important here. A guide guides, he indicates
the right direction and the appropriate ways to determine the right
direction. The guide does not force people to obey him nor even to
accept him. The guide brings a message, but messages get distorted
and twisted, or are just quite honestly misunderstood. God provides
general advice and instructions to his creation, but there is a great
deal of freedom for people to make their own decisions and take
their own risks. This explains the signi cant role of theology and
philosophy. The reasons they accept as signi cant in controlling
their actions do not point to a set of clear and distinct propositions,
since if they did there would be no variety of understandings of
religion, or even of Islam. Since the Qur’an tells us sometimes to
respect diversity in humanity and forbids us from compelling
religious allegiance, the idea that reasons are not followed
inevitably by particular actions and beliefs is presumably well
taken.

Muslims, like those in other religious traditions, are interested in

nding rational explanations for what they nd in their faith. We
often overemphasize the role of faith in religion, but although it is
signi cant, many other issues remain. Faith cannot tell us how to
resolve apparent di culties in reconciling religion with what we
know or suspect to be true in other parts of our life. Some times we
believe something to be true on religious grounds but have good
scienti c or even personal reasons nonetheless to doubt its truth.
This sort of issue needs to be resolved, and the only means at our
disposal is the use of rationality and the principles of theology and
philosophy as applied to religion and what is related to it.
 That does not mean that how one resolves these issues is
obvious, since it will depend on the particular theory that is
accepted. Some theologians think that one should base any
di culties in interpreting religious texts on other religious texts, in
particular by interpreting the Qur’an in terms of the Qur’an. Others
argue that one ought to give a great deal of respect to the traditional
sayings of the Prophet and his Companions, as they have come
down to us. Among these sayings, the hadith, one has to distinguish
between those that are more reliable and those that are less so, itself
very much a rational investigation. Regarding the Qur’an, one also
has to consider which verses might be thought to abrogate which
others, and for that one has to have some theory of which verses are
later than others. One requires, in fact, a historical grasp of the
events associated with the revealing of the various Qur’anic
passages, so that one can understand better the context in which
they were given. One also needs an opinion on the role that
independent judgment might be called upon to play, together with
who in the Islamic community one recognizes as an authority, and
how far that authority extends. This merely skims the surface of
some of the hermeneutic issues that arise in trying to understand
Scripture, and as we have seen there is also a wide range of more
general philosophical techniques available to us.

A way in which the di erent theorists in Islam are often

characterized is in terms of their being either rationalists or
traditionalists or something else, but in fact it goes with the
commitment to theory that one uses reason to try to make clear how
one is resolving problems and why that way of doing it is the right
way. Islamic philosophical theology has always in the past been a
lively arena for rational discussion and controversy, and there is
every sign that it will continue to have the same sort of character in
the future.

BIBLIOGRAPHY

(A source of general bibliographical information can be found in ‘A

Guide to Bibliographical Resources’ in S. Nasr and O. Leaman (eds.),
History of Islamic Philosophy (London: Routledge, 1996), 1173–9),
which is followed by a list of general introductions to Islamic
philosophy.)

ABDUL- RAOF, H., ‘The Linguistic Architecture of the Qur’an’, Journal

of Qur’anic Studies, 2/2 (2000), 37–51.

AL- BAQILLANI, ABU BAKR MUHAMMAD, I‘jaz al-Qur’an. Beirut: Dar lhya’
al-’Ulum, 1994.

CRAGG, K., The Event of the Qur’an, 2nd edn. Oxford: Oneworld, 1994.

FAKHRY, M., A Short Introduction to Islamic Philosophy, Theology, and

Mysticism. Oxford: OneWorld, 1997.

GWYNNE, R., Logic, Rhetoric, and Legal Reasoning in the Qur’an: God’s
Arguments. New York: RoutledgeCurzon, 2004.
HAHN, L., et al.(eds.), The Philosophy of Seyyed Hossein Nasr. Chicago:
Open Court, 2001.

HA’IRI YAZDI, M., The Principles of Epistemology in Islamic Philosophy:

Knowledge by Presence. Albany: State University of New York
Press, 1992.

HOURANI, G., Reason and Tradition in Islamic Ethics. Cambridge:

Cambridge University Press, 1985.

IZUTSU, T., God and Man in the Koran: Semantics of the Koranic
Weltanschauung. Tokyo: Keio Institute of Cultural and Linguistic
Studies, 1964.

——— Ethico-Religious Concepts in the Qur’an. Montreal: McGill

University Press, 1966.

JOMIER, J., The Great Themes of the Qur’an. London: SCM, 1997.

AL- JUWAYNI, Kitab al-irshad ila qawati al-adilla usul al-i‘tiqad, ed. M.
Musa and A. ‘Abd al-Hamid. Cairo, 1950.

LEAMAN, O., Averroes and his Philosophy. Oxford: Oxford University

Press, 1990.

——— Brief Introduction to Islamic Philosophy. Oxford: Polity, 1999.

——— Introduction to Classical Islamic Philosophy. Cambridge:

Cambridge University Press,

2001.
——— Islamic Aesthetics: An Introduction. Edinburgh: Edinburgh
University Press, 2004.

——— (ed.), Biographical Dictionary of Islamic Philosophers. London:

Continuum, 2006.

——— (ed.), The Qur’an: An Encyclopedia. London: Routledge, 2006.

MUTAHHARI, M., Fundamentals of Islamic Thought, trans. R. Campbell.

Berkeley, Calif.: Mizan, 1985.

NASR, S., and LEAMAN, O. (eds.), History of Islamic Philosophy. London:

Routledge, 1996.

PETERS, F.E. (ed.), The Arabs and Arabia on the Eve of Islam.
Brook eld: Ashgate, 1999.

AL- QURTUBI, ABU ‘ABD ALLAH MUHAMMAD, al-Jami‘li ahkam al-Qur’an.

20 vols. Beirut: Dar al-Kitab al-‘Arabi, 1997.

RAHMAN, F., Major Themes of the Qur’an. Chicago: Bibliotheca

Islamica, 1980.

——— Islam and Modernity: Transformation of an Intellectual

Tradition. Chicago: University of Chicago Press, 1982.

ROBINSON, N., Discovering the Qur’an: A Contemporary Approach to a

Veiled Text. London: SCM, 1996.

AL-SUHRAWARDI, The Philosophy of Illumination, trans. John Walbridge

and Hossein Ziai. Provo: Brigham Young University Press, 1999.
AL- ZAMAKHSHARI, ABU AL-QASIM, al-Kashshaf. 4 vols. Beirut: Dar al-
Kutub al-‘Ilmiyyah, 1995.

ZEBIRI, K., ‘Towards a Rhetorical Criticism of the Qur’an’, Journal of

Qur’anic Studies 5 (2003), 95–120.
 CHAPTER 26

CHINESE (CONFUCIAN) PHILOSOPHICAL

THEOLOGY

JOHN H. BERTHRONG

PREAMBLE

The Voice of the Turtle

The sun’s rays glint rst on the mountains to the west, then,
moments later, touch the thatched roofs of the temples and pit
dwellings that follow the curve of the Huan. The river, still in
shadow at the foot of the earthen cli , winds to the southeast
between clearings of sprouting millet, on its way to merge with the
powerful He. The year is the eleventh of Wu Ding’s reign, the season
spring, the day xinwei, eighth of the week.

Filtering through the portal of the ancestral temple, the sunlight

wakens the eyes of the monster mask, bulging with life on the garish
bronze tripod. At the center of the temple stands the king, at the
center of the four quarters, the center of the Shang world. Ripening
millet glimpsed through the doorway shows his harvest rituals have
found favor. Bronze cauldrons with their cooked meat o erings
invite the presence of his ancestors, their bodies buried deep and
safely across the river, but their spirits, some benevolent, some not,
still reigning over the royal house and the king’s person. One is
angry, for the king’s jaw ached all the night, is aching now, on the
eve of his departure to follow Zhi Guo on campaign against the Pa-
fang.

Five turtle shells lie on the rammed-earth altar. The plastrons

have been polished like jade, but are scarred on their inner sides
with rows of oval hollows, some already blacked by re. Into one of
the unburned hollows, on the right side of the shell, the diviner Que
is thrusting a brand of aming thorn. As he does so, he cries, ‘The
sick tooth is not due to Father Jia!’ Fanned by an assistant to keep
the glowing tip intensely hot, the stick ames against the surface of
the shell. Smoke rises. The seconds slowly pass. The stench of
scorched bone mingles with the aroma of millet wine scattered in
libation. And then, with a sharp, clear puklike sound, the turtle, the
most silent of creatures, speaks. (Keightly 1983:1)

Centuries later, Master Kong (Kongzi 551–479 BCE, or as he

is known in the west, Confucius) is recorded to have observed in the
Lunyu , or the Analects as they are known in English, that

3.13: Wang-sun Jia inquired of Confucius, quoting the saying

‘It is better to pay homage to the spirit of the stove
Than to the spirits of the household shrine.’

‘What does this mean?’

 The Master replied: ‘It is not so. A person who o ends against
tian has nowhere else to pray.’

7. 23: The Master said, ‘Tian has given life and nourished
excellence (de ) in me—what can Huan Tui do to me!’

11.12: Zilu asked how to serve the spirits [gui ] and the gods [shen
]. The Master replied, ‘Not yet being able to serve other people,
how would you be able to serve the spirits?’ Zilu said, ‘May I ask
about death?’ The Master replied, ‘Not yet understanding life, how
could you understand death?’ (Translations from Ames & Rosemont
1998: 85, 116, 144)

2. 4: The Master said, ‘At fteen I set my mind on learning; at thirty,

I took my place in society; at forty, I became free of doubts; at fty,
I understood Heaven’s Mandate; at sixty, my ear was attuned; at
seventy, I could follow my heart’s desires without overstepping the
bounds of propriety’. (Slingerland 2003: 9)

EISEGETICAL EXEGESIS

Do we really know what the ctional diviner Que thought when he

was divining or who he was consulting when he made his charge to
the spirits upon the command of King Wu Ding? If we follow
Keightly’s wonderful recreation of the ritual, the turtle, that most
silent of creatures, nally does indeed speak. While we can never
know if Keightly’s imaginative reconstruction is anywhere near
accurate, heated turtle plastrons do make this kind of puk-like
sound. Who does the turtle speak for? Is it the high spirit of the
ancestors, shangdi ? Is it tian supernal heaven? Or some other
powerful spirit? What is the nature of Chinese religion in its
formative phase?

Wilfred Cantwell Smith wrote with oracular insight when

considering the speci c case of Confucianism, ‘For the moment, we
may simply observe once again that the question, “Is Confucianism
a religion” is one that the west has never been able to answer, and
China never able to ask’ (Smith 1991: 69). The question of the
nature of Chinese religion in general and Confucianism in particular
lingers whenever there is an ecumenical discussion of divine,
supernal matters. All religious traditions are unique yet the Chinese
case seems to occupy a di erent niche. It is obvious that the Shang
kings (c. 1750–1040 BCE) believed they could contact the spirits,
divine beings who were the cause of good or woe in the Shang
world.

Moving forward a millennium, the rst two quotes from Master

Kong (Confucius) are often cited in order to demonstrate that he
shared a set of beliefs about and even veneration for a set of early
Zhou dynasty (c. 1040–249 BCE) religious ideals such as tian
supernal heaven and various minor and major spirits gui and gods
shen . With a typical Confucian air Master Kong observes that if
tian has given him a mission in life and helped to nourish his ethical
virtue, what could such a mere odious ru an as Huan Tui do to
him? Master Kong was convinced that only a person who has
o ended tian had no one to pray to and no protection from the
vicissitudes of an imperfect and dangerous world. Scholars see in
such passages in the Lunyu connections to the archaic Shang and
Zhou religious world-view reconstructed by Keightly.

The account we have of Master Kong’s religious sensibility is

tantalizingly vague and ambiguous about the precise nature of his
religious beliefs and practices. He cherished and promoted the
inherited rituals of the Zhou and yet prefers not to speak about
prodigies or strange events. He is most famous as the ‘ rst teacher’
and reveres tian as intimately connected to the virtue of his own life
and conduct. As his students lamented, we simply have not heard
about the religious vision of the master in any fulsome fashion save
for a clear piety towards tian and a commitment to the moral life.
We also know that he makes only one claim to fame when pressed
by his students to tell them in what he excels: others are wiser,
braver, faster, stronger, more intelligent, but no one loves learning
xue more than Master Kong. This love of learning is one key
element of Confucian ‘theology’ or re ection on the supernal path
tiandao . Any great Jewish rabbi would understand the religious
aura of pious and persevering learning. To revere tian is to set one’s
mind-heart xin on learning.

Contrariwise, the third quote, Lunyu 11. 12, is cited to prove that
Master Kong did not believe in the spirits and gods. However, such
an inference is not clear in the text. The most that can be drawn
suggestively from the quote is an agnostic view of how the spirits
and gods interact with human beings. Remembering that tian not
only gives us life but also nourishes de virtue, I interpret the
passage to mean that Kongzi is reminding his audience that
whatever role the spirits and gods play in the world, the proper
charge or supernal mandate tianming for a person was to
cultivate humane virtue ren . If a person persistently cultivates
virtue, then what is there to worry about in relationship to the
spirits, gods and even tian? As Herbert Fingarette (1972) so
elegantly observed, for Master Kong the secular is the sacred and
the axiology of ethical self-cultivation trumps any kind of bhakti
devotion to the spirits, gods, and tian. This does not mean that we
lack a natural or cultivated piety towards the spirits and gods, but
that our primary task is to learn how to serve our fellow human
beings in the way we would ourselves like to be treated in order to
create a ourishing, harmonious, and shining human society.

The famous outline of a cultivated life given in Lunyu 2. 4 is

paradigmatic for what a person must do in order to create and live
within a ourishing human society. The path of self-cultivation
begins with xue and at the end of life embodies the ability to
follow the desires of the mind-heart and never to overstep the
bounds of propriety. The attuned mind-heart also hears something
else, what in the western tradition we would call the music of the
spheres. The cultivated person has found how to walk the supernal
path of being fully human by means of personal self-cultivation of
virtue and through the ordering of society and an appreciation of
the supernal heaven’s mandate tianming a developed attention
to the mind-heart’s desires and a mature sense of propriety. We will
argue that this pithy statement encapsulates the salient features of
being religious following the Confucian Way—from learning,
cultivating a moral path, conforming to supernal heaven’s mandate,
and the consummate balance of desire, propriety, and the goal of
humane harmony in society.

What is truly fascinating about the Confucian tradition is that,

unlike so many other Eurasian religions, later Confucians did not
make the obvious—at least obvious to the Children of Abraham—
extension from the veneration of tian to veneration of an omnipotent
and omniscient God. Although known and remembered in the
classics, the gambit of worshiping or consulting a monotheistic high
god embraced so strongly by the western Asian religions is declined
by the Confucians, even though being religious has been and
remains integral to the Confucian Way. The Confucians may not
have a ‘theology’ if by this we demanded an a rmation of a
monotheistic creator, sustainer, and redeemer of the cosmos.
Confucians do have what modern Chinese scholars have called a
‘Daology’, a way to follow the Dao with reverence, creativity, and
propriety by means of elaborating the notion of tiandao the
supernal path or way of heaven.
 The early classical canon has been called the Confucian classics
but at the time these early texts were composed they were simply
the cultural records in many genres for the emerging intellectual
elite of the Shang, Zhou, and Han dynasties (c. 1750 BCE to 220 CE).
Other considerations beyond a focused reverence for tian as a high
god won out in the development of the Confucian Way rudao
when compared to west and south Asian religions and even
extensive segments of the Daoist world as well. While there is a
wonderful exfoliation of religious themes and traits over the next
three millennia of Confucian history, what is striking is just how
di erent Confucianism is from Daoism, Buddhism, and the other
religious theologies and cumulative traditions of Eurasia. In his epic
history of Chinese science Joseph Needham (1954–) once observed
that in Confucianism we have a creation without a creator and we
could add to this, a religion without supernatural theism. As we
shall see, the later Neo-Confucians (Song and post-Song
developments, i.e. 960 to the twentieth century) and the modern
New Confucians Mou Zongsan (1909–95) and Tu Wei-ming (1940–)
are explicit in defending the religious dimension of the Confucian
Way.

CONFUCIAN RELIGIOSITY

The literature about the complexity of discerning what ‘religion’

means in the Chinese setting is vast and growing daily as more and
more students of Chinese thought both in China and the Chinese
Diaspora ponder this question they are often asked by western
scholars of religion.1

What could ‘theology’ mean for Confucians? Theology is

certainly a Christian enterprise, though many other cultural
traditions have clear analogs to such Christian endeavors; for
instance Indian sages such as Shankara and Ramanuja wrote
treatises that any western theologian who learned something of the
Hindu world-view would instantly recognize as theological
re ection. We cannot remain con ned by the comfortable western
religious lexicography when we review the Confucian canon2 and
how it was interpreted and practiced for thousands of years as a
form of being ‘religious’. While it is true that we can discover
pervasive and persuasive analogs to western norms and forms of
theological discourse in Daoism and Buddhism, the truly hard nut to
crack remains the Confucian tradition. That is why it is so important
to ponder the nature of Confucian religious discourse.

Whitehead (1996: 67–8) wrote, ‘To-day there is but one religious

dogma in debate: What do you mean by God? And in this respect to-
day is like all its yesterdays.’ Whitehead informs us that when we
talk about religion, we are talking about our notion of God or what
Tillich labeled our object of ultimate concern or what people hold to
be the most supernal, divine reality. By subtly changing the focus
from God to an exploration of the Chinese supernal matrix we can
then readily include, as worthy markers of supernal reality, such
concepts as tian, the Dao, or the taiji the Supreme Polarity
[Ultimate] beloved of Song (960–1279) Confucian masters.

We will now focus our attention on the Confucian tradition as

the test case for explicating theology as the science or study of
supernal things or events in the Chinese cultural context.
Confucianism is a western term invented by the early Jesuit scholar-
missionaries in order to catalog the various forms of religious theory
and practice they found in late Ming China (1560–1644). Of course,
earlier Chinese intellectual historians asked the same sorts of
questions about how to classify the various teachings of the golden
age of Chinese thought in the Zhou dynasty. Scholars such as Sima
Tan and Ban Gu created bibliographical categories of intellectual
teachings and lineages, for instance, such as the designations of ru
and dao designating what we now call Confucian and Daoist
discourse. However, ru originally probably meant a scholar and dao
simply someone who followed the dao as exempli ed in texts such
as the Daodejing or the Zhuangzi. In contemporary Chinese we nd
terms such as rudao , rujao rujia and ruxue all of
which can be roughly translated as the Confucian Way, Confucian
teaching, the Confucian school, and Confucian learning. When
modern Chinese scholars discuss the cumulative Confucian tradition
they use rujia Confucian school for the totality of the tradition and
rujiao if they want to specify or recognize the religious dimensions
of the tradition. The most common sophisticated response about the
religious traits of the tradition now is to recognize that while
Confucianism is not a religion per se if the model of religion is the
template of the monotheisms of Judaism, Christianity, or Islam,
nonetheless there are certainly religious elements or dimensions of
the cumulative Confucian tradition that are religious. As Tu Wei-
ming (1989) has argued, rather than trying to de ne Confucianism
as a religion it is better to ask, what does being religious mean for a
Confucian? ‘Being religious, in the Confucian perspective, informed
by sacred texts such as the Zhongyong, means being engaged in the
process of learning to be fully human. We can de ne the Confucian
way of being religious as ultimate self-transformation as a communal
act and as a faithful dialogical response to the transcendent’ (ibid. 94).
Tu’s statement is considered a classic modern de nition of
Confucian spirituality. The task is to travel from the ‘puk’ voice of
the sacred turtle in the Shang to the a rmation of the Confucian
way of being religious a rmed by a modern New Confucian scholar
such as Tu Wei-ming.

THE CONFUCIAN WAY OF BEING RELIGIOUS

A fascinating historical question to ponder is, how did the classical

Chinese ru become transformed from the Shang court diviners into
the social system of modern New Confucians in China and the East
Asian Diaspora? We must remember that Confucianism also played
a vital role in the cultural, social, political, and religious life of
Korea, Japan, and Vietnam as well as China. As a religious question,
the Confucian Way is a long and complicated journey of the human
xin (mind-heart) seeking to embody the supernal reality of the
cosmos.

In order to simplify the task of discerning the ru way of being

religious I am pausing at three critical locations in the history of the
development of the Confucian Way. The rst epoch, Classical
Confucianism, is the foundation or classical period represented by
Master Kong, Master Meng (Mencius), Master Xun , and the
Zhongyong . These Confucian masters and texts are considered
essential elements of the golden age of classical Chinese philosophy
(c.551–221 BCE). The second era, Neo-Confucianism, is the great
revival of Confucianism thought and praxis during and after the
Song dynasty (960–). Song and post-Song thought is celebrated as
the second golden age of Confucian religious speculation and is
indebted to the interaction of generations of Confucian scholars
with their learned Daoist and Buddhist cousins. It is impossible to
think of Neo-Confucianism (Song and post-Song Confucian
discourse) without the impact and in uence of Daoism and
Buddhism. Neo-Confucianism is a bridge between the classical and
modern periods. The third temporal portal we will enter is the
modern period (1911 to the present) dominated by the encounter
with the west. In the world of contemporary Confucian thought the
third epoch is distinguished from its classical and Neo-Confucian
ancestors by being labeled New Confucianism—that is, a
Confucianism revived, transformed, and renewed by its encounter
with western culture, politics, religion, and philosophy. We will
select the work of Mou Zongsan and Tu Wei-ming as paradigms for
the contemporary Confucian art of being religious.

CLASSICAL CONFUCIANISM: MENGZI, XUNZI, AND THE ZHONGYONG

The second great classical Confucian master after Kongzi was

Mengzi (c.371–289 BCE). Although Master Meng protested that
he disliked the philosophically contentious nature of his age, it was
his duty to refute all the false schools and to place the true ru
learning of self-cultivation, propriety, and ritual back in their proper
place by following the insights found in Kongzi. Ideas that were
merely hinted at by Master Kong are elaborated and expanded by
Mengzi to give the rst relatively comprehensive philosophical
exposition of the Confucian Way.

Against adversaries as diverse as Mozi and his utilitarian

rejection of Confucian ritual conduct, Yang Zhu and his sophistic
defense of egoism, together with wandering Daoists who rejected
any value whatsoever in the moral deportment of a Confucian
worthy, Mengzi’s defense of the Confucian Way highlighted a
number of concepts that were to remain at the core of Confucian
religious and philosophical discourse. Two critical concepts Mengzi
de ned and defended from a Confucian perspective were xing as
human nature, dispositions, or tendencies, and xin as the human
mind-heart. Mengzi argued against all-comers that the supernal xing
was the matrix of the four seeds of the virtues of humanity,
righteousness or justice, ritual propriety, and wisdom or
discernment. Using botanical metaphors, Mengzi argued that if a
person successfully cultivated these four cardinal virtues then there
would be progress along the path of true virtue and a person could
realize the authentic Confucian way.

However, the task of such cultivation had to take place

somewhere, and the locus of the cultivation of the four seeds of
virtue was in the mind-heart. One of the most fascinating things
about Mengzi’s vision is his various strategies for helping us to
nurture the mind-heart, to reform or temper less than perfect desires
and inclinations in order to embrace the cultivation of the four
virtues as the proper path of becoming human. If Master Kong
professed the necessity to follow the path of the Dao, his disciple
Mengzi showed what innate seeds of morality human beings needed
to embody in order to allow them to become virtuous and provided
guidance by means of the careful praxis of self-cultivation contained
in our xing human nature.

In a tantalizing passage, Master Meng, in responding to a

question from a student about his own forms of excellence, is
recorded to have made the following response. By naming the truly
protean concept of qi 3 as an object of ethical cultivation, he
evokes an almost mystical sense of the power of Confucian self-
cultivation of the mind-heart, itself constituted by the qi of the
cosmos.

May I ask what your strong points are?

 I have an insight into words. I am good at cultivating my ‘ ood-
like qi’ [haoran chi qi ]
May I ask what this ‘ ood-like qi’ is?
It is di cult to explain. This is a qi which is, in the highest
degree, vast and unyielding. Nourish it with integrity and place no
obstacle in its path and it will ll the space between Heaven and
Earth. It is a qi which united rightness and the Way. (Lau (trans.),
Mencius, 1984: i. 57–8)

Master Meng makes two points. The rst is that he has insight
into words, and this means that he understood the role of
philosophical disputation about the true nature of the Dao. But
secondly, more than merely having insight into words, he had the
ability to cultivate his ood-like qi. His disciple is as perplexed by
what Mengzi means by this qi as we and countless Confucian
scholars have been over the centuries. Although to pro er some
‘obvious’ exegesis is the height of folly, it does appear that if one
can achieve this feat of the cultivation of the ood-like qi then the
person will be in tune with heaven and earth in manifest yi
rightness or righteous justice within the matrix of the Dao. In short,
such a person can achieve oneness with the very ow of the cosmos
in the highest degree. This truly is a teaching about self-cultivation
that becomes part of the Confucian religious quest. In fact, based on
Mengzi’s teachings, many of the great mystics of the Confucian
tradition have counted themselves as followers of Master Meng in
seeking to cultivate the ood-like qi.
 The third of the great classical Confucians was Xunzi (c.310–238
BCE). Because Master Xun had the audacity to contradict Master
Meng concerning xing by declaring that human nature or
dispositions were odious and perverse, he was banished from the
inner circle of the ‘orthodox’ founders of the Confucian way.
However, it is wise to remember that this banishment was only
formalized in the Song dynasty; prior to the rise of Neo-
Confucianism Master Xun was considered one of the greatest of the
classical Confucians. Given the range of his works, from poetry,
ritual theory, ethics, military theory, and epistemology, he was by
far and away the most systematic and coherent of any of the early
Confucian masters.

It was also Xunzi who resolutely rejected the supernatural

theological path. What was at least ambiguous in the religious
speculation of Kongzi and Mengzi takes a decidedly naturalist turn
in Xunzi. We are no longer in the realm of spirits and gods; we are
in the world of human contrivance. Even the human good comes,
according to Master Xun, not from innate seeds of virtue implanted
in our mind-hearts by heaven but rather is the outcome of conscious
human e ort or wei , contrived action as conformed and shaped by
ritual action li and yi justice/righteousness.

John Knoblock, the meticulous translator and commentator on

Xunzi, translates the chapter tianlun as ‘Discourse on Nature’.
While Xunzi can use tian in the older sense of heaven, for the most
part, he means nature as the orderly structure of the natural world.
17. 1 The course of Nature is constant; it does not survive because
of the actions of a Yao: it does not perish because of the actions of a
Jie. (Yao is a paragon of virtue whereas Jie is an exemplar of
perversity—my explanation of these two important gures and not
in Knoblock’s translation.)

17. 2b The constellations follow their revolutions; the sun and moon
alternately shine; the four seasons present themselves in succession;
the Yin and Yang enlarge and transform; and the wind and rain
spread out everywhere.…We do not perceive the process, but we
perceive the result—this indeed is why we call it ‘divine’.

17. 5 Heaven does not suspend the winter because men dislike cold
weather. Earth does not reduce its broad expanse because men
dislike long distances. The gentleman does not interrupt his pattern
of conduct because petty men rant and rail. Heaven possessed a
constant Way; Earth has an invariable size; the gentleman has
constancy of deportment.

(Knoblock 1994: iii. 14–17)

Notwithstanding his commitment to a naturalist cosmos, Xunzi

was more than a narrow rationalist or strict materialist—there is
still an element of the supernal in his cosmos. Moreover, his form of
naturalism is expansive and holistic in that it continues to
commingle the grand triad of heaven, earth, and humanity as
essential for the ourishing of the cosmos. What he explains is a
creation without a supernatural creator: creation is the coordinated
work of heaven, earth, and humanity, each playing its own
primordial role in the cosmic drama. As with so many later
Confucians, Master Xun can be religious without invoking the
transcendent powers of a creator God.

Xunzi had a vision of the world as grand ritual inspired by

powerful poetry and music. He was also moved by the virtue or
process of cheng , true integrity, a term crucial to his theory of
self-cultivation. In his chapter on ‘Nothing Unseemly’ he included
this hymn to self-realization that demonstrates how true a Confucian
he really is, and wherein his religious sensibility really lies.

As for the gentleman’s cultivation of the mind-heart, there is

nothing better than cheng (true sincerity or self-realization), for he
who perfects true integrity/sincerity need do nothing else than
allow humanity to be maintained and justice acted upon. With the
realization of the mind-heart and the maintenance of humanity,
they become manifest, and being manifest they are spirit-like, and
being spirit-like they are capable of transforming; with the
realization of the mind-heart and the practice of justice, there is
order, and when there is order, clarity, and with clarity there is
change. The transformations and changes act together and this is
what is called the Virtue of Heaven …Heaven and Earth are great,
but without true integrity they cannot transform the ten thousand
things; although the sage has knowledge, yet lacking true integrity,
he is out of touch. The exalted ruler is indeed eminent, and yet,
without true integrity, he is lowly. Therefore true integrity is what
the gentleman seeks to approach, and is the root of the cultivation
of a airs. Only when he rests in the perfection of true integrity will
he assemble the like minded to himself…After having succeeded [in
perfecting true integrity] the character of the person is completed
and has progressed such a long way that does not revert to its
origins, and this is transforming.

(A Concordance to the Xunzi 1996: 1; for an alternative translation,

see Knoblock 1988: i. 177–8)
It would not be an exaggeration to call this Master Xun’s hymn to
cheng. If Kongzi and Mengzi show us an emerging form of Confucian
spirituality, Xunzi provides us a way to cultivate our mind-heart so
that we can put ourselves in tune with the constancies and
transformation of the Dao.

THE ZHONGYONG

In the Zhongyong4 we clearly perceived the creative holistic and

expansive religiosity of the classical Confucian period. If Xunzi
provides us with a hymn to cheng, the Zhongyong is an extended
religious meditation on the role of cheng in acknowledging and
discerning the secular as the sacred. ‘What tian commands (ming
) is called natural tendencies (xing ); drawing out these natural
tendencies is called the proper way (dao ); improving on this way
is called education (jiao )’ (Ames and Hall 2001: 89). The how of
this method of instruction for improving the way in order to follow
the Dao and obey the natural tendencies of tian delineates the
function of cheng . No other classical text better exempli es the
religious dimension of the notion of sincerity or self-actualization
that is the process of cheng.

Ames and Hall translate cheng by calling it ‘creativity’5 in order

to stress the processive and generative nature of the religious vision
of the text. ‘Creativity (cheng Image) is the way of tian (
Image); creating is the proper way of becoming human (
Image Image). Creativity is achieving equilibrium and
focus (zhong Image) without coercion; it is succeeding without
re ecting’; ‘and only if one can assist in the transforming and
nourishing activities of heaven and earth can human beings take
their place as members of the triad’ (ibid. 104 and 105). As a form
of classical Confucian spirituality, this is pretty exalted company
indeed and bespeaks of the holistic, relational, and creative supernal
discernment of a holistic and expansive Confucian humanism in its
most robust religious sensibility.

‘Thus, the utmost creativity (zhicheng Image) is ceaseless. …

This process of utmost creativity is in full display without
manifesting itself, changes without moving, and realizes without
doing anything. …The way of heaven and earth can be captured in
one phrase: Since events are never duplicated, their production is
unfathomable’ (ibid. 107).
 It is this unfathomable nature of the Dao that provides what the
New Confucian scholar Mou Zongsan calls the supernal dimension
of the Confucian Way. Mou (1994) sees this as a form of immanent
transcendent, yet it is so productive, unceasing, and unfathomable
that it is surely as ‘divine’ as Xunzi asserted in his description of
cheng in the cosmos.

The journey from Kongzi to the Zhongyong illustrates the outline

of classical Confucian spirituality; a theology of the secular as the
sacred; of a holistic and expansive humanism that seeks to nd a
creative place for humanity in the ceaseless transformations of the
Dao. It is a Trinitarian theology to be sure, yet it is a natural, non-
theistic spirituality that locates the supernal Dao through the
interaction of tian, earth, and humanity. Furthermore, along with
locating the supernal in this primordial triad, it suggests that the
truly supernal or divine is the proper coordination of the cosmos
founded and funded by this triad of heaven, earth, and humanity.

We now must move forward more than a millennium to take up

the thread of the story again in the Song dynasty, the second golden
epoch of the Confucian Way.

THE SONG REVIVAL

The second great epoch of the Confucian Way began in the late
Tang and gained momentum to in uence in the Northern Song
(960–1126 CE). It is important to remember that the contemporary
New Confucians of the third epoch are the direct descendants of the
great masters of the Tang, Song, Ming, and Qing dynasties. Many
western students of Confucianism have a marked preference for the
classical period even though the Song–Ming masters argued that
they were faithfully reviving and restoring the grand classical
Confucian tradition that had been partially eclipsed by the arrival of
the Buddhist dharma in China in all its resplendent Mahāyāna glory
and the ourishing of various indigenous Daoist sects.

Later in the Qing dynasty (1644–1911) many great Confucian

scholars revolted against what they considered the metaphysical
excesses of the Neo-Confucian masters. Great thinkers such as Wang
Fuzhi (1619–92), Dai Zhen (1724–77), and the equally impressive
Japanese scholar Ogyū Sorai (1666–1728), protested that the Song–
Ming masters had won a pyrrhic victory over their Buddhist and
Daoist cousins by incorporating too much of the metaphysics of the
Buddhists and Daoists in Song and Ming Confucian philosophy. For
instance, the whole of Zhu Xi’s complicated theory of the abstract
nature of li, coherent principle, was taken to be a perfect example of
borrowing a Buddhist idea and applying it inappropriately to
Confucian philosophy. This is the same charge that many western
scholars also level against the Neo-Confucians: the Song
philosophers muddied the waters of the perfectly good classical
Confucianism with a hodgepodge of Buddhist and Daoist
metaphysics that ought to be left behind by modern Confucians. The
problem is that even the great Qing and Japanese scholars had to
unlearn Song–Ming Neo-Confucianism before they could begin to try
to purify the Buddhist-tainted works of scholars such as Zhang Zai
(1020–77), Zhu Xi (1130–1200), and Wang Yangming (1472–1529).
The irony is that many contemporary New Confucians see absolutely
no reason to abandon any of the religious, historical, aesthetic, and
philosophical achievements of the Song–Ming masters or the critical
Qing and Japanese scholars.

We will brie y review two examples of how the Northern Song

and Southern Song masters appropriated, expanded, and renewed
the teachings of their respected Zhou dynasty teachers concerning
the supernal way. The two writings reviewed are the justly famous
‘Western Inscription’ of Zhang Zai and ‘The Treatise on the
Completion of the Mind-heart’ by Zhu Xi.

Zhang Zai was one of the greatest of the Northern Song masters,
and with his strong a rmation of the role of qi vital energy as the
key concept of his philosophy, he was deeply appreciated by the
later Qing scholars because of this commitment to the dynamic,
concrete nature of the cosmos. Zhang is lauded for writing the
famous short essay the ‘Western Inscription’.6 Not only does the
‘Western Inscription’ capture the cosmic scope of Song cosmology,
axiology, and ethics, it also has an unmistakable religious sensibility
that has inspired the entire Confucian tradition.

Although the ‘Western Inscription’ is not a long document, I will

select the passages that are the most important for the future
‘theological’ development of the Song tradition.
The Western Inscription

Heaven is my father and Earth is my mother, and even such a small

creature as I nds an intimate place in their midst.

Therefore that which lls the universe I regard as my body and

that which directs the universe I consider as my nature.
All people are my brothers and sisters, and all things are my
companions.
The great ruler (the emperor) is the eldest son of my parents
(Heaven and Earth), and the great ministers are his stewards.
Respect the aged—this is the way to treat them as elders should be
treated. Show deep love to the orphaned and the weak—this is the
way to treat them as the young should be treated. The sage
identi es his character with that of Heaven and Earth, and worthy is
the most outstanding man. Even those who are tired, in rm,
crippled, or sick; those who have no brothers or children, wives or
husbands, are all my brothers who are in distress and have no one
to turn to.
When the time comes, to keep himself from harm—this is the
care of a son. To rejoice in Heaven and to have no anxiety—this is
lial piety at its purest.
He who disobeys [the Principle of Nature] violates nature. He
who destroys humanity is a robber…
 One who knows the principles of transformation will skillfully
carry forward the undertakings [of Heaven and Earth], and one who
penetrates spirit to the highest degree will skillfully carry out their
will…
Wealth, honor, blessing, and bene ts are meant for the
enrichment of my life, while poverty, humble station, and sorrow
are meant to help me to ful llment.
In life I follow and serve [Heaven and Earth]. In death I will be
at peace.
(Chan 1963: 497–8)

The classical Confucian virtues are all recon rmed, but within a
spiritual vision of the cosmos that expands its sense of concern
consciousness from the immediate family and clan to the whole of
humanity and beyond to all the things or events of the world. Both
blessings and di culties are the methods by which the Dao
strengthens the sage and the worthy. If one is successful in this
spiritual, moral, and intellectual quest, then as Zhang says with such
eloquence, even ‘In death I will be at peace.’ If this is not an
evocation of a supernal religious sensibility, a meditation on how a
person ought to deport him or herself in the presence of the Dao,
the reader is deaf to the spiritual yearnings of the Northern Song
mind-heart.

Later in the Southern Song, the greatest of the Neo-Confucian

masters was Zhu Xi. It was his interpretation of the Song legacy that
became the foundation of the Chinese civil service from 1313 to
1905. Moreover, through his extensive writings on ritual praxis, he
in uenced not only the philosophical development of Song-Ming
thought but also had a profound impact on the ritual and spiritual
life of the Chinese, Korean, Japanese, and Vietnamese people.
Although he was known as a champion of li Image, coherent
principle, pattern and order, he was also profoundly implicated in
the study and cultivation of xin Image, the mind-heart of
humanity. The spiritual aspect of Master Zhu’s teachings emerges in
works such as:

Treatise on the Completion of the Mind-heart [Jinxin shuo

Image]

‘The completion of the mind-heart [xin Image] is knowing nature.

Knowing nature is knowing heaven’ [from Mengzi] and hence we
a rm that a person can know the mind-heart when the person
knows nature and this capacity to know the nature is to know
heaven. Now, heaven is the spontaneity of coherent principle and
the source from which human beings are born. Nature is the essence
of coherent principle which the person receives to become a person.
The mind-heart is the way a human being controls the person and
fully sets forth this coherent principle. ‘Heaven is great and
boundless’ [quoting Zhang Zai] and nature receives [this boundless
nature] completely. Therefore the essence of a person’s fundamental
mind-heart is itself expansive and without limitation. Only when it
is fettered by the sel shness of concrete things, hemmed in by
seeing and hearing pettiness, does it become concealed and
incomplete. A person can in each event and in every thing
exhaustively examine their coherent principles until one day the
person will penetratingly comprehend them all without anything
being left out: then a person can complete the broad essence of the
fundamental mind-heart. The reason for a person’s nature being
what it is and the reason for heaven being supernal heaven is not
beyond this and is connected in its unity.

(Zhu Xi 2002: xxiii. 3273)

The Treatise7 encapsulates not only the outlines of Zhu’s mature

philosophical interpretation of the role of xin, but it shows the
structure of Confucian spirituality, the knowledge, comprehension,
and unity of supernal things, as it were. The three major elements
are learning, action, and discernment through the cultivation of the
mind-heart. Tian is the source of coherent principle and its
manifestation is human nature xing. When we understand our
nature, we understand our coherent principle li, and when we
understand our coherent principle we will penetratingly
comprehend the holistic interconnections of heaven, earth, and
humanity. The method for the encompassing envisagement of the
coherent principle of supernal heaven depends upon the proper
ethical cultivation of the mind-heart. The mind-heart is potentially
expansive and limitless in its powers of knowledge and holistic
discernment if it can overcome sel shness and the pettiness of
unexamined things and events. Even Xunzi would have found little
trouble with this formulation: he would simply add that the proper
way to overcome sel shness and pettiness of spirit is always and
everywhere by ritual civility and righteousness. The unity of
learning, action, and discernment of the proper axiological values of
self-cultivation remain as the core of the Confucian search for tian,
the marker of supernal things or events in a holistic world-view.

THE NEW CONFUCIAN MOU ZONGSAN’S Image RELIGIOUS THOUGHT

The traditional Confucian world, a culture that stretches back to the

primordial sages of antiquity, Kongzi, the baroque philosophical
edi ces of the Song-Ming masters, the harsh critiques of the Qing
Evidential Research scholars, all came crashing down under the
hammer blows of the colonial west. By 1905 the civil service
examinations, based on Zhu Xi’s interpretation of the Confucian
Way, were abolished and in 1911 the empire itself was
extinguished.

In 1949 when Chairman Mao, standing in Tiananmen Square,

proclaimed that the Chinese people had stood up in order to liberate
themselves from their semi-colonial status to the western powers
and Japan, he did not make his announcement in the name of a
revived Confucianism. Mao proclaimed that modern China would
follow the Marxist path and that Confucianism was t only, and
barely that, for the museum of intellectual history as a disgraced
example of the worst of feudal ideology. Many non-Marxist scholars
would have agreed with Mao’s assessment of the negative future for
Confucianism even if they did not embrace his passionate messianic
version of Marxism.

Nonetheless, the reports of the demise of Confucianism were

premature to say the least. By the early 1920s a small group of
scholars, while recognizing that a great deal of Confucian culture
was in need of drastic reform, began the process of reviving and
reforming the Confucian Way as had Kongzi, Mengzi, Xunzi, and the
Song-Ming masters before them. This group of scholars, now into its
fourth generation, has become a vigorous force in China and the
Chinese Diaspora. In order to distinguish it from its classical and
Neo-Confucian ancestors, it is now called the New Confucian
movement.

While there are many distinguished reformers within New

Confucianism, most agree that Mou Zongsan (1909–95) was the
most comprehensive and creative philosopher of the second
generation. Mou was more than a philosopher per se; he also
defended the religious nature of the Confucian Way and tried to
show that it was both a spiritual and philosophical tradition worthy
of respect as it underwent a drastic reformation in the twentieth
century. Mou was a passionate defender of Confucianism; he was
also a harsh critic and pointed out that a renewed Confucianism
must and should include a place for democratic theory and political
practice; the development of ecumenical and modern scienti c
learning; and a complete rejection of its patriarchal mistreatment of
women. Only a thoroughly modern and reformed Confucianism
could stand a chance of nding a place again in the mind-hearts of
the Chinese people. One crucial element of Mou’s reform project
was to examine the religious and spiritual dimension of the
Confucian Way rudao Image

In the twelfth lecture of The Special Features of Chinese

Philosophy, Mou (1994: 125) provides a concise discussion of his
understanding of the religious dimensions of the Confucian Way:

According to our way of seeing things, a culture cannot but have a

foundational inner spirituality. This is the power of the creative
spirit of culture, and it is where the special nature of the culture
resides . . . this motive power is religion and it does not matter in
what form. Following this we can say that the motive power of the
foundation of the life-values of a culture are necessarily found in
religion.

Mou hypothesized that if Confucianism is a religion or way of being

religious, it must satisfy the broad functional features of any
religion. There are two such traits: (1) Any religion ‘should perfect
the duties of the proper way of everyday life’ (ibid.
126). Mou strongly a rms that this is the case for the Confucian
Way with its great emphasis on the moral cultivation of the ve
cardinal virtues, social rituals, and pedagogies for ‘the proper
conduct of daily life’ (ibid. 127) it has devised over the centuries to
sustain human ourishing. (2) A religious way of being human must
be a path of living spirituality, something more than a secular path
of conduct, though such a practical, even mundane path is critical to
the holistic vision of the spiritual quest. ‘But the reason why a
religion is a religion is not merely on this level, for it also has yet
another necessary function to ful ll …A religion must stimulate the
incipient seeds of a person’s spirituality towards transcendence, to
point out a path of spiritual life’ (ibid. 128). In short, a religion has
both daily praxis and a coherent and supernal principle that govern
its teachings and structures.

According to Mou the two foundational traits of Kongzi’s

spirituality are the traits of ren Image, supported by nature xing
Image and the supernal way tiandao Image. Ren has both
the meaning of the moral code of the ve cardinal virtues, namely
ren Image humaneness, yi Image righteous justice, li
Image ritual action or civility, zhi Image discernment or
wisdom, and xin Image faithfulness, and an even deeper,
spiritual meaning as well. Ren is ‘the fundamental root of creativity’
(ibid. 130) or, as he earlier in the text glosses his own formulation
in English, ‘creativity itself’ Image Image(ibid. 32).
Therefore, ‘Ren is the essence of the myriad things of the cosmos
and in the primary sense it is not a substance but rather spiritual’
(ibid. 131).

Mou adds that for something to be spiritual, it must have two

traits: intuition and perseverance. ‘It [intuition] has the meaning of
morality. From the perception of yielding, intuition is the sense that
life does not head into a deadlock, is not constricted, and is the
opposite of the meaning that something is dead and without feeling’
(ibid. 131). In his earlier discussion of creativity itself, Mou de nes
it as unceasing generativity, shengsheng buxi Image (ibid. 31).
The mind-heart yields to the dictates of the true intuition of
generativity and moral connection and this is what Mou calls
perseverance without ceasing in the cultivation of the way of
creativity itself. Moreover, this is the unceasing movement of tian, or
tianming Image Image, as the ‘profound and unceasing’
urgency of the decree, ming Image, of the creativity of tian
(ibid. 131). If we follow tianming Mou believed that we will
cultivate the essence of creativity itself as ren.

Human beings are given xing Image as a set of basic

dispositions and if these are cultivated correctly, Mou a rmed that
a person can conform the mind-heart to tiandao itself (ibid. 131 .).
Mou goes on to point out that this form of Confucian spirituality
does not focus on a creator god, though such an idea was clearly
present in the early classical literature. Confucian religious thought
simply took a di erent direction. Because of their emotional nature,
human beings often take their feelings of despair or joyous piety and
turn them into a prayer to God. Again, while Mou understands and
approves of this kind of religious insight, it simply and historically is
not what the Confucians did with their spiritual sense of supernal or
divine things (ibid. 133 .).

If Confucians decline the life of prayer to a supreme Creator God,

then wherein does the main thrust of Confucian spirituality direct
itself? ‘I answer that it lies in the question of how the person goes
about embodying the Way of Heaven’ (ibid. 135). Mou links the
primordial ‘how’ of embodying tiandao to his notion of inter-
subjectivity zhuguanxing Image (ibid.). ‘If we open up the idea
of inter-subjectivity, then what is above and what is below become
uni ed. The inter-subjective nature and the objective nature are
correlated and the virtue of the Dao becomes the praxis of the mind-
heart’ (ibid.). The ultimate goal of the cultivation of inter-
subjectivity is to become a sage, the person who assists in the
unceasing creative activity of the cosmos and becomes one with
heaven and earth. In one sense Confucianism is humanistic in its
focus on the embodying of the Dao. But this is a holistic humanism
that is open to tiandao and tianming. It has both an immanent and
transcendent focus and is hence a profound form of spirituality.
Confucianism de nitely has incredibly strong religious dimensions
in that the truly cultivated person seeks to become a sage, to be
conjoined as one with the Way of Heaven analogous to a Buddhist
bodhisattva or a Daoist Immortal. While it is often asserted that
Confucianism is an overly optimistic faith, Mou counters that the
task of self-cultivation is extremely di cult, so much so that only a
sage or worthy can achieve such a feat. Most human beings have
limits to their ability to perfect the mind-heart; the path is long and
the burden is heavy. But the true Confucian never desists from
asking, in any situation, how he or she can advance towards the
Way of Heaven (ibid. 135 .). Mou (and other New Confucians of
his generation) provide the context for the current discussions
underway in China and the Chinese Diaspora about Confucian
spirituality.

Tu Wei-ming has an oft-quoted de nition of being religious that

is inspired by Mou’s defense of the religious dimension and
spirituality of the Confucian Way.

Being religious, in the Confucian perspective, informed by sacred

texts such as Chung-yung [ Zhongyong], means being engaged in the
process of learning to be fully human. We can de ne the Confucian
way of being religious as ultimate self-transformation as a communal
act and as a faithful dialogical response to the transcendent [italics in
the original]. This is also the Confucian prescription for learning to
be fully human …we can say that Confucian religiosity is expressed
through the in nite potential and the inexhaustible strength of each
human being for self-transcendence. (Tu [Du] 1989: 94)

Tu’s formulation of ‘being religious’ as a Confucian adds the trait

‘communal act’ as constitutive of the tradition, namely the need to
act within what Tu calls a duciary community. Tu’s notion of
‘being religious’ meshes perfectly with Mou’s argument that the
religious dimension of the Confucian way is in the method of how a
person embodies tiandao. Of course Mou recognized this need for
communal religious solidarity, and for him the prime exemplar of
this trait is his notion of inter-subjectivity, the fusions of subject and
object relations between the myriad things. For Mou and Tu the task
of being religious is unthinkable without a person’s interaction with
other people, society, and the natural world. The classic formulation
is that a Confucian must be a sage within, a king without neisheng
waiwang Image Image.

THE FUTURE OF THE CONFUCIAN TASK FOR BEING RELIGIOUS

For a modern Confucian religious thinker, the issue is, as Mou

clearly understood, ‘the question of “how” the person goes about
embodying the Way of Heaven’ (Mou 1994:135). Tu Wei-ming
glosses this task of embodying the tiandao as the quest for learning
how to be fully human. Any contemporary Confucian spiritual
re ection must cohere with this primordial Confucian insight into
how the human person is conjoined through self-cultivation with the
most spiritual, transcendent, and supernal traits of the Dao.

Lunyu 15. 29 says, ‘The Master said, “Human beings can broaden
the Way—it is not the Way that broadens human beings”’
(Slingerland 2003: 185). It is this kind of insight that has always
separated the Confucian religious tradition from the great theistic
teachings of west and south Asia. Nonetheless, a sampling of
modern New Confucians such as Mou Zongsan and Tu Wei-ming
illustrates why Lunyu 15. 29 remains relevant to the current New
Confucian revival project. One can imagine many changes or
transformations of the Confucian Way in the future, but one cannot
imagine that Confucians will turn the saying around to a rm that it
is the Way that broadens human beings, even if Confucians are
concerned to discern how they should respond to the Way of
Heaven. It is precisely this note of immanent transcendence that
many western thinkers nd appealing as they struggle to overcome
the feeling that the western religious path has overemphasized the
nature of God as the wholly other, the completely transcendent One
of worship. If the Jesuits and early Protestants found the
Zhongyong’s thesis that heaven, earth, and humanity form a living
trinity to be heretical, then modern men and women nd this an
increasingly plausible notion about how to be spiritual in the world
that modern science reveals to us.

We now wander into speculation about the future of a tradition.

This is a hard business and I profess no abilities at prophecy.
Moreover, the Confucian Way has been dramatically challenged
over the last 150 years, perhaps more so than in any other period of
its long journey. Confucians have experienced other great
transformations before with marked success. The rst triumph was
in its classical period with the teachings of the late Zhou masters
Kongzi, Mengzi, and Xunzi. The second grand transformation
occurred when the late Tang and Northern Song Neo-Confucians
responded to the religious and philosophical challenges of Daoists
and Buddhists, creating a form of religious, intellectual, and social
life that lasted for more than eight hundred years as the dominant
ideology of the Chinese empire. The o cial examination system
ensured the continuity and homogeneity of the legal state and social
mores throughout the empire. The radical impact of the west since
1842 has provoked a third great transformation, one that is really
only its rst phases now as the New Confucian movement and
reformation. It is truly a reformation because all the New
Confucians realize and a rm that they must nd a method for a
positive engagement with modern democratic theory and practice,
ecumenical science, the changing roles of women in the modern
world and an appropriate response to globalization and the
ecological crisis globalization announced via its rapid
industrialization and pollution of the earth.

Notwithstanding all the pressures for change, some things will

remain constants in New Confucianism: (1) learning as an act of
religious piety; (2) the praxis of civility through ritual actions and
patterns of social behavior; (3) discernment of the ultimate moral
good, the transcendent and supernal elemental moral seeds that are
as immanent in the human mind-heart as they are transcendently
supernal; (4) self-cultivation of the mind-heart in order to cultivate
the Dao for oneself in service to others; and (5) the ongoing search
for the harmony he Image of human ourishing. The path is
long and New Confucians, as much as their ancestors, will be called
upon to discern what constitutes the merely mundane from the
morally necessary and to seek a way to move from what is to what
ought to be.

Confucianism has returned with a vengeance to the cultural

agenda of China and the Chinese Diaspora. But then, so many things
are ourishing in China, including monumental architectural and
building projects such as the Three Gorges Dam; the fastest train the
world in Shanghai; a newly discovered love of the automobile;
amazing industrialization linked to an equally amazing pollution of
the environment; successful though closely supervised revivals of
Daoism, Buddhism, and Christianity; and a renewed pondering of
the meaning of the Confucian Way. In fact, the Chinese government,
on the model of France’s Alliance française and Germany’s Goethe
Institutes, is now in the process of establishing hundreds of
Confucius Institutes in the western world. One wonders what
Chairman Mao would make of all of this. Bookstores are crammed
with large sections on Confucian thought and the practical
applications of this ancient wisdom. Everyone is asking, what can
Confucianism contribute to the ourishing of modern China? No one
has a complete or coherent answer.

One common criticism of New Confucianism when it is

compared to the revivals of Daoism, Buddhism, and Christianity is
that there has been precious little renewal or reformation of the
public side of the Confucian Way. At one level this is not hard to
understand. The Chinese government, though now relatively
friendly to the revival of certain aspects of the Confucian Way, is
vigilant about any large religious or cultural movement gaining too
much public support and in uence beyond the careful monitoring
and control of the government. The Chinese civil service has
de nitely not revived the study of the Four Books as the basis of
government service,8 though even the highest governmental o cials
are pleased to quote from the Confucian classics when it suits their
purposes. The cynical among Confucian scholars note that the
government loves to cite those aspects of the Confucian Way that
encourage a hard-working, educated population dedicated to social
harmony and the continued material development of a uni ed
country. Criticism of state policy is not high on the government’s list
of things needed in the Confucian revival.

When the government decided that it was timely to allow the

teaching of literary Chinese again in the private schools,9 public
schools, and universities, the Four Books and the other classics again
resumed their centuries-long role as common textbooks to instruct
modern Chinese students in the crafts of reading classical and early
modern literature, poetry, history, and philosophy. The criticism is
often made that there has been a commendable renewal of neisheng,
the sage within, in terms of the modern critical study of the
Confucian tradition but previous little of the waiwang, or king
without, in the sense of public ritual or policy. One of the tasks of a
Confucian scholar has always been to speak truth to the powerful,
even at the risk of great personal harm.

Even amidst all these problems of social orientation, policy

formation, and reformed ritual praxis, most Chinese intellectuals
acknowledge that Confucianism is de nitely a major part of the
Chinese cultural DNA. For instance, there is legislation in China that
mandates that children must care for their aged parents: xiao
Image, lial piety, it appears, is even a good Chinese
Communist virtue. A specter is haunting the Chinese intellectuals:
what about the spiritual resources of the Confucian tradition? Some
choices are obvious. Zhang Zai’s ‘Western Inscription’ has become
the favorite manifesto for the ecological movement and is seen as
precisely the sort of indigenous spiritual resource the Chinese
people will need to remember while they grapple with the
incredible ecological crisis provoked by the amazingly rapid
industrialization of China.

The other side of the coin is the recognition of the revival of

Confucian scholarship. As we have seen, one of the critical traits of
spirituality is learning. While merely arcane research is never
commended (even though it is admired by other scholars), learning
as a way to embody the tiandao is still foundational to the survival
and revival of Confucianism in all its aspects, including the
religious; nor is such a form of spirituality unknown in the west. For
thousands of years the rabbis have taught that the study of Torah is
a sacred act, a way of discerning God’s will in the midst of the daily
life of the people. As the Chinese people, and Koreans, Japanese,
Vietnamese, and even assorted westerners, ponder the spiritual
dimensions of the Confucian Way in the midst of its third epoch,
just as with Kongzi and Zhu Xi, new forms of Confucian ‘theology’
or re ection on supernal, divine matters will ourish—though what
this renewed appropriation of tiandao Image will become is
something left to present and future generations to determine.

With its emphasis on immanent transcendence, New

Confucianism also speaks to the nervous culture of the west. Many
people search for a spirituality that allows for the full development
of humanity and a sense of connection to the environment nurtured
by a commitment that makes a religious place for them within the
holistic and ontologically uni ed cosmos illumined by modern
ecumenical science. People want a unity of heaven, earth, and
humanity in order to nd a balance and harmony for their lives.
With this kind of yearning, there will always be a place for being
religious in a Confucian mode.

NOTES

Save for the names of authors and book titles, I have modi ed
all texts and translations to use the contemporary pinyin
Romanization system for consistency.

1. Actually this is always more of a problem with the elite

segments of the Chinese world; Chinese popular and local
religions look very religious with all kinds of rites, revealed
texts, gods and goddesses, spirits, ghosts, demons, tightly
organized religious hierarchies and societies, priests, art,
temples—in short, all the functional elements we would expect
from a ‘religious’ tradition. In particular it is the emerging
Confucian elite and literati culture that diverges from the
commonplaces of Eurasian religious history.

For two excellent large collections of material on the

question of the religious nature of Confucianism, see Yao 2003
and Tu (Du) and Tucker 2003–4. For general accounts of the
 historical and religious issues involved see Berthrong 1994,
1998, 2000.

2. The Confucian Canon began with the recognition of ve great

classical texts, but by the Song dynasty (960–1279) eight more
early texts had been added for a total of thirteen texts. From
the Song on, when people refer to the canon, it would be these
thirteen early texts and would be known in the west as the
Thirteen Confucian Classics. All the texts were composed,
according to critical scholarship, at least by the Han period
(221 BCE–220 CE).

3. Qi de es any easy translation: it can accurately be translated as

vapor, force, action, material force, vital force, vital energy,
matter-energy—and many other reasonable suggestions have
been made over the centuries. If there is a truly metaphysical
or cosmological term in the Chinese philosophical
lexicography, qi is it. It is simply assumed by most indigenous
Chinese philosophical and religious traditions to be one of the
foundational traits of the cosmos, the very vital stu or vital
force or vital material energy out of which all the ten thousand
things, the favorite term used to designate everything that is,
arises, and perishes. It becomes a particular element of
analysis, as we shall see, with the Song Neo-Confucians.

4. It is impossible to say just when the current text of the

Zhongyong was composed. Tradition holds that it is after
Kongzi and before Mengzi and Xunzi. However, it could be as
 late as the early Han dynasty for all that. I have chosen to
place it after Xunzi not because I have a better insight into its
textual history but because I want to follow up on the
theological/spiritual point about cheng raised by Master Xun.

5. If translating cheng as creativity seems a stretch, Andrew Plaks

(Ta Hsüeh and Chung Yung trans., 2003) suggests ‘integral
wholeness’ in his 2003 translation of the Zhongyong (Chung
Yung), which neatly captures the expansive humaneness of the
process of cheng. Wing-tsit Chan 1963 stays much closer to the
common meaning of cheng and translates cheng as being
sincere, with the recognition that only through sincere
commitment to the highest ethical values can we learn to
become fully human.

6. It was called the ‘Western Inscription’ because it was inscribed

on the western wall of Zhang’s study.

7. Mou 1968–9: iii. 439–47 has an extended commentary on the

essay. He contends that it clearly shows how far Zhu Xi has
departed from the correct teachings of Cheng Hao and Hu
Hong. Levey 1991, in his superb dissertation, is much less sure
that Zhu got it wrong.

8. It is important to remember that the Four Books were the

bedrock, the core of the texts used by all students seeking to
join the imperial civil service from 1313 to 1905.
 9. For instance see a BBC Internet article at
<http://news.bbc.co.uk/1/hi/world/asia-
paci c/7169814.stm>, accessed 10 June 2008, about the
teaching of Confucianism in private schools.

REFERENCES

AMES, ROGER T., and Rosemont, Henry, Jr. (1998). The Analects of
Confucius: A Philosophic Translation. New York: Ballantine.

——— and HALL, DAVID L. (2001). Focusing on the Familiar: A

Translation and Philosophical Interpretation of the Zhongyong.
Honolulu: University of Hawaii Press.

BERTHRONG, JOHN H. (1994). All Under Heaven: Transforming

Paradigms in Confucian-Christian Dialogue. Albany: State
University of New York Press.

——— (1998). Transformations of the Confucian Way. Boulder:

Westview.

——— and BERTHRONG, EVELYN NAGAI (2000). Confucianism: A Short

Introduction. Oxford: Oneworld.

CHAN, WING-TSIT (1963). A Source Book in Chinese Philosophy.

Princeton: Princeton University Press.

A Concordance to the Xunzi (1996). A Concordance to the Xunzi. The

Chinese University of Hong Kong Institute of Chinese Studies.
Hong Kong: Commercial.
FINGARETTE, HERBERT (1972). Confucius—the Secular as Sacred. New
York: Harper& Row.

KEIGHTLEY, DAVID N. (ed.) (1983). The Origins of Chinese Civilization.

Berkeley: University of California Press.

KNOBLOCK, JOHN (1988–94). Xunzi: A Translation and Study of the

Complete Works. 3 vols. Stanford: Stanford University Press.

LAU, D. C. (trans.) (1984). Mencius, trans. D. C. Lau. 2 Vols. Hong

Kong: The Chinese University Press.

LEVEY, MATTHEW A. (1991). ‘Chu Hsi as a “Neo-Confucian”: Chu Hsi’s

Critique of Heterodoxy, Heresy, and the “Confucian” Tradition’,
Ph.D. Dissertation. Chicago: University of Chicago.

MOU, ZONGSAN (1968–9). Xinti yu Xingdi

Image [Mind and Nature]. 3
vols. Taipei: Zhengjing Shuju.

——— (1994). Zhongguo Zhexue de tezhi

Image [Special
Traits/Features of Chinese Philosophy]. Taipei: The Student
Book Company.

——— (2003). Spéci cités de la philosophie chinoise, trans. d’Ivan P.

Kamenarovic and Jean-Claude Pastor. Paris: Éditions du Cerf.
[French trans. of Mou 1994.]

NEEDHAM, JOSEPH (1954–). Science and Civilisation in China. 8 vols.

Cambridge: Cambridge University Press.

PLAKS, ANDREW (trans.) (2003). Ta Hsüeh and Chung Yung (The Highest
Order of Cultivation and On the Practice of the Mean), trans. with
 an introduction, Andrew Plaks. London: Penguin.

SLINGERLAND, EDWARD (trans.) (2003). Confucius Analects with

Selections from Traditional Commentaries. Indianapolis: Hackett.

SMITH, WILFRED CANTWELL (1991). The Meaning and End of Religion.

Fortress edn. Minneapolis: Fortress.

TU WEI-MING [Du Weiming] (1989). Centrality and Commonality: An

Essay on Confucian Religiousness. Albany: State University of New
York Press.

——— and TUCKER, MARY EVELYN (eds.) (2003–4). Confucian

Spirituality. 2 vols. New York: Crossroad.

WHITEHEAD, ALFRED NORTH (1996). Religion in the Making. New York:

Fordham University Press.

YAO, XINZHONG (ed.) (2003). Encyclopedia of Confucianism. 2 vols.

London: Routledge-Curzon.

ZHU XImage
I (2002). Zhuzi quan shu
Image [The Collected Works of
Master Zhu], ed. Zhu Jieren, Yan Zuozhi, Liu Yongxiang. 27
vols. Shanghai: Shanghai gu ji chu ban she & Anhui jiao yu chu
ban she.
 INDEX

Abelard, Peter 434–5

Abrabanel, Isaac 546
Abraham, William: Crossing the Threshold ofBerulle, Pierre de Divine
Revelation 22
Achtemeier, Paul 43
Acts of the Apostles, The 12, 31, 48, 204, 298, 484
Adam 571
and original sin 437–40
temptation of 268
Adams, Marilyn McCord 456, 499–500
Adams, Robert M. 317–18, 329 n43
afterlife see Heaven, doctrine of; Hell; resurrection, doctrine of;
Islam, afterlife
agent causation 118–20, 179
agnosticism: and Incarnation 454
al-Farabi 557–8, 565–6
Alfonso X, King 369
al-Ghazali 557, 558–9, 570–1
al-Juwayni 562
al-Ma’mun, caliph 556
al-Sira 557
al-Suhrawardi 561
Alston, William
and belief 24–5, 27, 133
and ‘Christian mystical practice’ 24–5, 27
‘Divine Mystery Thesis’ 82–3
and human understanding 132, 233, 338, 380
and interventionism 242
‘Two Cheers for Mystery’ 80
Ames, Roger T. 584
Amos (prophet) 33
animal su ering 369–70
Anselm, St
Cur Deus Homo (Why The God-Man) 432–3
and divine aseity 107
and divine omnipotence 169
and divine omnipresence 200–1, 203
and divine omniscience 129–30
and divine simplicity 209
and God: concept of 212–3
and Hell: account of 494
 Monologion 109, 200–1
ontological argument 212
Proslogion 84, 108, 169, 200, 201
Apollinarius 470
apologetics 348
apophatic tradition 96 n7
Apostles’ Creed 482
Aquinas, St Thomas
and atonement 434
and creation 156, 231, 233, 234–5
and divine eternity 146–7
and divine knowledge 92, 129, 131
on divine omnipotence 169–70, 182
on divine omnipresence 170, 201–3
on divine punishment 361
on Eucharist 529, 533–5
and ve ways 107
and foreknowledge 139
on Hell 495–6
and human reason 88–9
and the Trinity 88–9, 409–10
on liberty 267
 and local presence 519, 520
and moral perfection 220, 230–1, 235
Summa contra Gentiles 83, 107, 201
Summa Theologiae 107, 129, 131, 132, 146–7, 156–7, 169–70,
201–2, 515, 520, 533, 534, 535
see also Thomism
on the will 234
Arabic language 568
Arian controversy 18–19
Aristotle
and approach to philosophy 545–7
and ethical theory 229, 547–8
and rst mover 107
and form and matter 418–19
and Islam see Islam, Peripateticism
and natural law theory 314–16
and practical reason 232
and the soul 470
and substance 418–19, 533–4
and views, expression of 566
aseity see God, aseity of
Armstrong, David 111–12 598
Ash‘arites 562
Athanasian Creed 403, 415
atheism
arguments for 340 see also evil, problem of
and evil 355
and human su ering 341
and naturalism 61–2
and theism 340–3
atonement, doctrine of 11, 362, 430–47
commercial theory 433
forensic ction 436
and Hell 494
and moral exemplar theory 434–5
and penal substitution theory 435–47
and ransom view 431–2
and satisfaction theory 432–5
Augustine, St
and Biblical interpretation 14–16
and compulsion 503
and Creation 17, 247
and divine foreknowledge 138
De Genesi ad Litteram 17, 28 n7
 De Trinitate 108, 409, 421
and divine omnipotence 168–9
and divine revelation 14–15
and divine simplicity 107
and Heaven 502
and Hell 496
and human freedom 223
and moral perfection 223, 230
and original sin 437–40
Aulén, Gustaf 431
Auschwitz: visit by Pope Benedict XVI 352
Averroes see ibn Rushd Avicenna see ibn Sina axiological goodness
225–6
axiology 237 n16
Ayer, A. J. 2

Baconian inductivism 61
Baillie, John: The Idea of Revelation in Recent Thought 51 n5
Ban Gu 579
Barbour, Ian 74 n5
Barnes, Michael 420
Barr, James 49, 145–6
Barth, Karl: Church Dogmatics 441, 442–3
Barton, John: Reading the Old Testament 27 n5
Basil, St, of Caesarea 27 n3
Epistle 52 423
On the Holy Spirit 27, 67, 27 n3
Basinger, David 294
Bayes’ theorem 339–49
Bayle, Pierre 86–7
Bayne, Tim 467
beati c vision 91–3
Beaudoin, John 395 n9
belief
and evidence 503–4
and evil 332–3
Islam 563, 567
justi cation for 19–27
Maimonides and 546
systems of 469–70
and trust 26
see also logic; science: and religion
Benedict XVI, Pope 352
Bernstein, Mark 385
Berulle, Pierre de 80
Bible
Apocrypha: Acts of John 81
authority of 11–13, 21–2, 25, 45–6, 50
authorship 41–2, 47–9
and Christian tradition 12–19
and Creation 249–50, 252
and divine-human relationship 500
and divine revelation 11–13, 15, 18, 20–1, 35–6, 37–9, 43
Hebrew: book of Job 354
and history 35, 36–7
inconsistencies in 13–14
and inspiration 41–50
metaphor in 16
New Testament 12, 13, 14, 18, 39
Old Testament 12, 13–14, 16
and science 15
see also Scripture
bilocation, Eucharist 526–33
Bivalence, Principle of 164 n7, 219
Blount, Charles 85
Blount, Douglas K. 455–6
Boethius, Anicius Manlius Severinus 138–9, 147
Brower, Je 425 n22
Brümmer, Vincent 448 n19
Buddha 319, 320
Buddhism 75 n18, 585
Madhyamikas 87
Burrell, David 454
Byrne, Peter and Houlden, Leslie: Companion Encyclopedia of
Theology 79

Cain, James 496

Calvin, John 20–1
Cappadocian Fathers 412, 413, 415, 416, 421–2, 423
Carnap, Rudolph 2
Categorical Imperative 223
Chalcedon, Council of 452–3, 455, 461, 472 n3, 515
Chalcedonians 172
Chang Tsai 258–9
Charlesworth, M. J. 84
Chesterton, G. K. 6 n3
Chinese philosophical theology see Confucianism
Choi, Isaac 295, 297–8
Christ see Jesus Christ
Christianity: and divine omnipresence 211–12
Christian doctrine
justi cation of 19–27
source of 11–19
Christian tradition 12–13, 17–9
Christology 452–71
coherence of 453
and divine nature 172–3, 460–1
Hick, John on 435 ‘Kenotic’ 465
Molinist 276
and parts and properties 458–63
relative identity solutions 457–8
restriction solutions 463–6
and ‘two minds’ solutions 467, 468
see also Nestorianism
Chrysostom, St John 79, 92
church: and identity 19
Clark, Kelly James 76 n42
classical foundationalism 3
Clayton, Philip 67
Clouser, Roy 69
Coburn, Robert 153–4
cognition 71
Cohen, Richard 352
Collins, Robin 258
Colossians, Letter to the 250
commandments: Judaism 542, 543, 546, 547–51, 549, 550, 551
communion 513
communion of saints 486
Companion to the Philosophy of Religion, A 79
compatibilism 121–3, 141, 236 n8, 266–7
Conee, Earl 226
Confucianism 574–94
Classical 580, 580–3
and Daoism see Daoism future of 592–4
Neo-Confucianism 580, 582, 585
New Confucianism 580, 585, 588–94
religosity of 576–80, 589, 591
Song Revival 585–8
Zhongyong 583–4
Confucius 580, 581, 592
see also Lunyu (Analects)
consequences, natural: theodicy of 361–2
consequentialism 221–2
Constantinople, rst Council of 470
Constantinople, Third Council of 453–4, 471
consubstantiality 406, 422
consubstantiation 514
contradictions 87
conversions 503
Corinthians
First letter to the 33, 39, 41, 91, 250, 441, 484, 518
Second letter to the 35, 444
Council of Chalcedon see Chalcedon, Council of
counterfactuals: freedom of action 141–2, 174–7, 180–1, 268, 270–
1, 274, 276–82, 394 n3, 395 n7
Craig, William Lane 42–3, 414, 415–16, 423
created rational beings 325
Creation: explanation for 243–6, 578, 583 see also God: and
creation creationism 61
Credulity, Principle of 386
Crescas, Hasdai 543
crime, and punishment 436 see also punishment, theodicy of; Hell
Cross, Richard 409, 420
Cyril of Alexandria 462
Cyril of Jerusalem, St 168

Dai Zhen 585

Damasio, Antonio 71
Daniel, book of 484
Daoism 579, 580–1, 584, 585, 590, 591–2
Darwin, Charles 62–3, 73
Davis, Stephen 130, 135
Day of Resurrection 478, 479, 482, 484, 485
dei cation 173
Deism 30–1, 85–6
Dennett, Daniel 62
Descartes, René 370
determinism 260, 267, 496, 567
Devil, the 431–2
Dionysius the Areopagite: Celestial Hierarchy 80
disasters, natural 357
divine action 68, 155–9
and evolution 241–59
divine attributes 107–17, 147
divine authority 306–26
divine essence 83–5, 87, 89, 91–5, 97 n27, 100 n68, 132, 569
divine eternity 145–64
divine freedom 219–20, 225–7, 230–3, 264–5, 291–2, 301 n15, 301
n18, 318, 570–1
divine goodness 217–35, 292–3, 319–20, 324
divine hiddenness 246
divine immutability 146–7
divine inspiration 2–3, 38, 41–9
divine knowledge see God: and knowledge
divine life 252–3
divine love 230, 233, 235, 384–5
divine omnipotence 167–91, 192 n36, 324, 519
accounts of 168–70, 171–83, 185–6, 190–1
and divine authority 324
and impossibility 188–90
range of 183–5, 186–7, 190
divine omnipresence 199–212
and Christianity 211–12
and co-location 211
and containment 210
and controversy 202–3
historical view of 200–2
and incorporeality 201–2, 210–11
 as knowledge 201–3
and multiple location 209–10
and non-occupation relations 204–5
occupation accounts 205–6, 209–11
occupation relations 206–9
as power 201–3
and simplicity 209–10, 213 n17
and timelessness 210
and the Trinity 212
divine omniscience 129–42, 148–9, 154, 159–63, 266, 324, 454,
465
de ned 130, 160–1
and divine authority 324
and Islam 559
and philosophical objections 133–42
divine perfection 84–6, 323–4, 326
divine personhood see God: personhood of
divine providence 43, 49, 90, 243–53, 262–82, 290, 298, 301 n14,
361
divine reason 88
divine retribution 494–5
divine revelation 11–13, 15, 18, 20–2, 27, 30–40, 83, 87
 Deist interpretation of 31
general 36
modes of 34–9
progressive nature of 39–40
propositional and non-propositional 34–6
purpose of 32–4
and skeptical theism 389
special 36–8
verbal and non-verbal 34–5
see also divine inspiration
divine simplicity 105–23, 146–7, 209
and coherence 110–17
and contingency 117–23
doctrine of 105–10
interpretation of 108–10
divine substance see Trinity, the
divine temporality 153–63
divine timelessness 146–52, 284 n4
divine volitions 118–20
divine wisdom 550
Dostoevsky, Fyodor: The Brothers Karamazov 336
Draper, Paul 339, 343, 383–6
Dulles, Avery: Models of Revelation 35
Duns Scotus, John 170, 173, 175
duty: and moral perfection 223–7

Eastern Orthodox: and Creation 250

Ecumenical Council 18
Edwards, Jonathan
on Hell 494
on the Trinity 79, 89–91
‘Miscellaneous Observations’ 91, 92, 99 n50
on mysteries 81–3, 89–92
Einstein, Albert 58, 71, 149–51, 157, 530
election: doctrine of 441–2
Elijah: and the prophets of Baal 293, 294
emanation 558, 560
Empiricus Sextus 553
emotions: human 71
Ephesians, Letter to the 249–50
Ephesus, Council of 461
Epicurus 553
epistemic evaluation 71
Erigena, John Scotus 80, 83–4, 175
eschatology 249–53
eternity 145–64
ethics: Islam 562
ET-simultaneity 157–8
Eucharist 510–36
Evans, Stephen: The Historical Christ and the Jesus of Faith 25
evil
Bayesian arguments 339–49
conditions for 355–60
existence of 257 see also theism, skeptical; theodicy
and freedom of choice 245
good and 81, 376–81
horri c 386, 388–90
and misery 500–1
moral 335, 362–5
MT (modus tollens) arguments 333–9
natural 335, 362, 365–7
power of 431–2
problem of 222, 223, 243–5, 257–8, 374–94
and the afterlife 332–3
logical 334–9
practical 332
evolution 23, 62–3, 241–59
existence: eternal 145
Exodus, book of 35
externalism 23

facts
moral 309, 310–20
tensed 159–63
faith
Christianity 25–6, 35
Judaism 543
Fallen Humanity 438–9, 444–5
Falsafa see Peripateticism families 254, 256
Faraday, Michael 65
Fiddler on the Roof 277
Fingarette, Herbert 577
Flew, Anthony see MacIntyre, Alasdair and Flew, Anthony
Flint, Thomas 130, 142, 174–83, 234, 293
foreknowledge 133, 138–41, 148–9, 263–7, 270–3, 275, 277, 357–
8
forgiveness 435
Francis of Marchea 65
Frankfurt, Harry G. 178
Freddoso, Alfred 133, 140, 174–83, 494
freedom
of action 132, 222, 223, 233, 245, 263, 289–90, 567
in Islam 571
challenges to 499–502
of choice 121–2, 295–6, 356–8, 363, 365, 366–70
and divine foreknowledge 138–42, 218–19, 262–72, 274–6, 356–
8
and God see divine freedom and Hell 496, 499–502
libertarian 43, 141, 154, 180–1, 187, 248, 258, 263–70, 274–6,
496, 497, 501, 504
and non-human creatures 370
Free Process Defense 244
free will 68, 244–6, 362–5
Free Will Defense 222–3, 335
Free will theodicy see theodicy, free will
Freud, Sigmund 62
Friedman, Alexander 151

Gamaliel, Rabbi 548

Garcia, Jorge 229–30
Gavrilyuk, Paul 81
Geach, Peter 139, 218–19, 417, 500
Gellman, Jerome 179
Genesis, book of 13–14, 15, 16–17
Adam 268, 437
Gnosticism 81
god: de nition of 407
God
Anselmian concept of 212
aseity of 107–8, 120–2, 313–14
and Atonement 362
as author of Scriptures 42, 47–8
and belief 26–7
and church doctrine 19
and control see divine providence and creation 25, 155–6, 220,
231, 233, 241–59, 262, 313, 315, 325, 368, 558
in Confucianism 583
in Islam 569
de nition of 306
Deist interpretation of 30, 31
and divine interpretation 108–9
and divine revelation 34, 36–9, 83
in Eastern Orthodox theology 250
embodiment of 201–2, 210–11
essence of 83–5, 87, 89, 91–5, 97 n27, 100 n68, 132, 569
eternal nature of 145–6, 152, 163
and evil see evil, problem of
and evolution see divine action: and evolution
existence of 4, 212, 243, 309, 342, 374–5, 382–3, 384–5 see also
belief
and evil 333–5, 337, 338, 339, 342 see also theodicy and evil,
problem of
in Islam 569
as rst cause 156, 267
freedom of see divine freedom
and gratitude 325
and inspiration 2–3, 38, 41–9
and knowledge 91–5, 118, 121–2, 129–42, 159–63, 262–3, 274–
5, 559
foreknowledge 133, 138–41, 148–9, 263–7, 270–3, 275, 277,
357–8
free knowledge 43, 141, 274–5
middle knowledge 42–3, 140–2, 149, 275–7, 301 n14
natural knowledge 42, 141–2, 274–5 602
probabilities 272–3
 scientia visionis 148
self-knowledge 83–4, 87–8, 92–5, 101 n68
and love 230, 233, 235, 384–5
and moral perfection 217–35, 324
nature of 21, 108–10, 117, 145–6, 152, 163, 174, 189–90, 323–4
and obedience 310–2, 321, 323
omnipotence of see divine omnipotence
omnipresence of 199–212
omniscience of see divine omniscience
Islam 559
perfection of 84–6, 323–4, 326
personhood of 107–8, 153–5, 319
power of 187–90
and prayer 287–99
purpose of 245–53
Islam 570–1
redemption 32–3
and risk 272–3, 276, 358
sovereignty of 262, 268, 274, 313–4
see also divine providence and divine authority
temporality of 153–63
timelessness of 146–52, 284 n4
 as truthmaker 118, 120
will of 117–20, 223–4, 231, 264–5, 267–9, 313–4
Word of 80
see also Trinity, the
Gödel, Kurt 88
Godhead, and the Trinity see Trinity, the
good
and evil 81, 376–81
maximizing 221–3
goodness, moral: and divine authority 324
grace, optimal 498
Gracia, Jorge: How Can We Know What God Means? 17
Gregory of Nyssa 14, 93, 421, 423
Gregory the Great 13
Grim, Patrick 136, 192 n25

haecceity 136
Hájek, Alan 396 n26
Hall, David L. 584
happiness, eternal 234, 506–8
Hare, John 318
Hartshorne, Charles 202–3
Hasker, Willam 139–40, 279, 280
Haught, John 243, 244–6, 259 n3
Heaven: doctrine of 491–2, 506–8
Hebblethwaite, Brian 472 n3, 473 n13
Hebrews, Letter to the 33, 129, 202, 441, 464, 513
Hell 491–506
accounts of 499–502
defenses of 493–9
eternal 495, 502–6
and freedom 496, 499–502, 506–7
proportionality objection 495–9
separation conceptions of 498
Helm, Paul 37, 163, 220
hermeneutics 13–14
Hesse, Mary 70
Hick, John 368, 435
Evil and the God of Love 499
Hilbert, David 91
Himma, Kenneth 497–8
Hobbes, Thomas 314–16, 328 n24
Hodge, Charles 448 n17
Ho man, J. and Rosenkrantz, G. 139, 175
Holocaust 352
Holy Spirit 24, 26–7, 27 n3, 36–8, 40, 41–2, 45–6, 48–9
Holy Trinity see Trinity, the homoousion clause 404, 416
Hookway, Christopher 71
Houlden, Leslie see Byrne, Peter and Houlden, Leslie
Howard-Snyder, Daniel 368, 376
Hudson, Hud 487 n6, 536 n5
human body 255
human identity 478
human intellect 83
humanity
attributes of 463–4
fallen see Fallen Humanity
human nature 581–2, 587, 590
and morality 314, 315, 318
human perception 23, 71
human reason 87–8
humans
emotions 71
interconnections 254–7
human su ering 233, 337–8, 339, 340–2, 344, 346, 354, 393, 495
Hume, David 78, 343
Hunt, David 131–2
Hyperactive Agent Detection Device (HADD) 62

Ibn al ‘Arabi 560

Ibn Bishr Matta 557
Ibn Rushd (Averroes) 557, 558, 565
Ibn Sab‘in 560
Ibn Sina 557, 559–60
Ibn Tibbon, Judah 545
Illuminationism: Islam 560, 561–2, 563
Incarnation, the 237 n25, 452–71
indi erence, hypothesis of (HI) 343, 344, 355, 383, 385–6
inspiration see divine inspiration intellect: human 83
Intelligent Design (ID) 61
interconnections
and Creation 248–58
ancestral 249
emergent 248–9
redemptive 249–53
interventionism 241–3
intuition theory of morality 223
Irenaeus, St 14, 17, 368
Isaiah (prophet) 145, 444
Ishraqi thought see Illuminationism
Islam 556–72
and afterlife 564, 569
and bodily resurrection 476, 478, 482, 488 n11
ethics 562
Illuminationism 560, 561–2, 563
logic 564–8
mysticism (Su sm) 560–1
and Neoplatonism 557–60
Peripateticism (Falsafa) 557–60, 563–4, 565–6, 569
politics 563
and Qur’an see Qur’an
rituals 570
soul, the 559, 563–4
tradition 568–70
truth 565–6
Israelites 39–40

James, Letter of 169, 246

Jeremiah (prophet) 167, 353
Jesus Christ and absence, real 533–5
 and atonement 430–47
Cruci xion of 48, 81, 431, 442, 487 n3
divine attributes of 172–3, 464–6
human attributes of 456, 457–8, 464–5, 466
Incarnation of 237 n25, 452–71
on Mount of Olives 81
mystery of 80, 81
Passion of 354
presence, real 514, 515–33, 516–18
as prophet 22
resurrection of 482, 517–8
as revelation 41
sacri ce of 433, 513
and the Trinity 452–3, 461, 462–3
union with 442, 444
Jews 541–3
Job, book of 90, 129, 167, 354
John Chrysostom, St see Chrysostom, St John
John, Gospel according to 18, 33, 484, 535
John, St 18, 81
John the Scot see Erigena, John Scotus
Joshua, book of 39
Judaism 507
and bodily resurrection 476, 478, 488 n11
and philosophical theology 541–53
and practice 541–4
and reform 541–2
traditional 541

Kant, Immanuel 223, 224, 228, 308, 326, 329 n40, 547
Kaufman, Gordon 492, 493
Keightly, David N. 575
Kellenberger, James 90
Kellner, Menachem 546
Kennedy, John Fitzgerald 543
kenosis 465
Kepler, Johannes, Mysterium Cosmographicum 57
Kings, First book of 293
Kitcher, Philip 369
Knight, Gordon 510 n27
Knoblock, John 582
knowledge 132–3
accounts of 302 n26
de dicto 134
 de praesenti 135–7, 162–3
de re 134, 162
de se 136, 162
divine 42–3, 91–5, 117–18, 121–2, 129–42, 159–63, 262–3,
274–5, 559
rst-person 135–7, 162
foreknowledge 133, 138–41, 148–9, 263–7, 270–3, 275, 277,
357–8
free 43, 141, 274–5
middle 42–3, 140–2, 149, 275–7, 301 n14
moral 318
natural 42, 141–2, 274–5
Kongzi see Confucius
Koons, Robert C. 526, 528
Kretzmann, Norman 135–6, 139, 151, 157–8
Kuhn, Thomas Samuel 65
Kvanvig, Jonathan 136, 160, 497, 498 604
Lamont, John: Divine Faith 25
Lang, Bernhard 492
language: and Islam 556–7, 558, 568
law
of nature 242, 246, 251–2
 natural 314–17, 365–7
Judaism 548, 549, 550
Leftow, Brian 151, 158–61, 410–12, 461–2, 489 n16
legalism 31
Leibnitz, Gottfried Wilhelm 87–8, 99 n44, 353, 361–2, 516–7
Lewis, C. S. 6 n3, 235, 297, 496, 498, 503
Lewis, David 113, 360, 436–7, 528
libertarianism 43, 141, 154, 180–1, 187, 248, 258, 263–70, 274–6,
496, 497, 501, 504
life, eternal 507–8
liturgy 512
Locke, John 89, 543
logic: Islam 564–5, 566–8
logical empiricism 2–3
Lombard, Peter 169
Luke, Gospel according to 14, 167
Lunyu (Analects) 574, 576, 577

McCabe, Herbert 454

McCann, Hugh 266–7
McDannell, Colleen 492
McGinn, Colin 226, 237 n19
Mach, Ernst 150
MacIntyre, Alasdair and Flew, Anthony: New Essays in Philosophical
Theology 3
Mackie, J. L. 335
Madhyamikas 87
Maimonides, Moses 488 n11, 541–2, 546–53
The Guide of the Perplexed 545, 548, 549, 551, 552–3
Shemonah Peraqim (Eight Chapters) 548, 550
Malachi (prophet) 446, 513
Malebranche, Nicholas 370
Mann, William: The Blackwell Guide to the Philosophy of Religion 79
Mao Zedong 588
Marcel, Gabriel 82
Marcion 14
Mark, Gospel according to 432
Martin, C. B. 105, 112
Marty, Martin 492–3
Marx, Karl 255
Mary, Mother of Jesus 461–2
Masek, Lawrence 288
materialism 70
mathematics 71
Matthew, Gospel according to 40, 50, 167, 432, 464, 484
Mavrodes, George 51 n4, 88, 89
Mawson, T. J. 227
maximization problem 222–4, 226, 228, 234, 292
Maximus the Confessor 93, 173
Maxwell, James Clerk 65
Mencius, see Mengzi
Mendelssohn, Moses 541, 542–4, 551–2
Mengzi 580–2, 592
Merricks, Trenton 410, 462, 463
meta-epistemology 3
metamorphosis: and redemption 250–1
metaphysics, monadic 516–17
Metzger, Bruce: The Canon of the New Testament 12
Midgley, Mary 76 n50
middle knowledge 42–3, 140–2, 149, 275–7, 301 n14
miracles 293–4, 367, 525–6
Miskawayh 570
modalism 407, 409–10, 413
modality: and God’s power 187–90
Mohammad see Muhammad
Molina, Luis de 42–3, 122, 141–2, 263, 282 n4
Molinism 42–3, 140–2, 175, 195 n57, 269, 274–82, 394 n3, 395 n7
grounding objection to 278–81
monotheism 414–15
Islam 560
morality
and absolutes 230
de ned 307
and divine authority 306–26
skepticism about 391–4
Theistic Explanation of Morality (ThEM) 307–8, 310–11, 312,
313, 317, 320–1
moral perfection see God: and moral perfection
Moreland, J. P. 414, 415–16
Morris, Thomas
and Christ, attributes of 464, 465–6
and duty theory 225–6, 227, 230–2
and God, Anselmian concept of 213
reduplicative analysis 455
and the Trinity 410
and two ‘minds’ solutions 466–7
Mou Zongsan 584, 588–91
Muhammad 556, 566
Murphy, George 245, 256
Murphy, Nancey 67, 68, 256
Murray, Michael 244, 246–7, 293–5, 297, 358, 497
Muslims 568 see also Islam
Mu‘tazilites 562
mysteries 78–95
mysticism
Christian 80, 83
Islam (Su sm) 560–1
Myth of God Incarnate, The 453

Nasr, Seyyed Hossein 561

naturalism 58–9, 61–2, 65, 346–9, 582–3
natures: de ned 420
Nazianzen, Gregory 422, 423
Needham, Joseph 578
negative theology see apophatic traditon
Neo-Confucianism 258–9, 580
Neoplatonism, and Islam 557, 558, 559, 560
Nestorianism 452, 456–7, 467, 471
Nestorius 461–2
Neurath, Otto 2
New Confucianism 580, 585, 588–91
Newman, John Henry 18, 223, 264
New Testament see Bible: New Testament
Newton, Sir Isaac 57, 65, 150–1
Nicaea, second Council of 28 n3
Nicene Creed 17, 19, 404–5, 415–16, 420, 482
Nicene Council, First 404–5
Nietzsche, Friedrich 507
nihilism 31

obedience 321
obligations, moral 317–18 ‘occasionalists’ 314
Ockham, William of 138, 140, 170
Old Testament see Bible: Old Testament
omnipotence see divine omnipotence
omnipresence see divine omnipresence
omniscience see divine omniscience
‘Open Theism’ movement (Openists) 76 n42, 139–40, 263, 264,
269–74, 281, 356–8, 510 n27
Oppy, Graham 176, 177–9, 195 n47
Origen of Alexandria 14, 15–16, 28 n11, 168, 361, 478
original sin 6 n3, 257, 437–40, 443–7
 and imputation 437–8
traducianism 438, 440
Orr, James 43
orthodoxy 413
Eastern Orthodox 250
Otte, Richard 342
Otto, Rudolph 79, 82, 94
Oxford Handbook of the Philosophy of Religion 4–5, 79

pain see su ering

panentheism 69, 75 n34, 75 n35, 341
Pannenberg, Wolfhart 68
pantheism 62, 204, 341
paradox of the stone 184
Parmenides 219–20
Parsons, Josh 206
Pascal, Blaise 503–4
Patristic statutes 423 n2
Paul, St 13, 14
Chrysostom, St John of 79
and Corinthians 43
and creation 235, 252
 on divine revelation 39
on eucharistic liturgies 518
and evil 258
on God, existence of 246
on knowledge 91
on scripture 48
teachings of 41
Paul VI, Pope: ‘Credo of the People of God’ 514
Peacocke, Arthur 67, 73
penance 434
Penelhum, Terence 86
Pereboom, Derek 393
perfection 84–5
moral 212–35
models of 221–32
Peripateticism (Falsafa) 557–60, 563–4, 569
Persia: philosophy 558, 562
persons, inseparability of 422
Peter, St 293
First letter of 48
Second letter of 41
Phillips, D.Z. 336
philosophical theology see theology, philosophical
philosophy
analytic tradition 1–4
Continental tradition 1–4
of religion 3–4
physics: and life 251–2
Pike, Nelson 138, 335
Plantinga, Alvin
and divine foreknowledge 140
and freedom of action 278
and Free Will Defense 222–3, 335
and God, existence of 212, 348–9
God, Freedom and Evil 353
and Incarnation 489 n16
and moral good 222–3
and omniscience 130
Warranted Christian Belief 22
and warrant theory 23–4, 26–7, 132
Plantinga, Cornelius, Jr. 414
Plato 564
Plotinus 97 n31
politics: Islam 563
Polkinghorne, John 244, 245, 259 n3
polytheism 407, 413, 415
Popes 18
Benedict XVI 352
Paul VI 514
prayer
petitionary 286–99
answered 287–91
predication 109–17, 124 n1, 125 n3
Prior, Arthur 134
Process Theists 282 n3 see also free process defense
prophecy 41, 272, 277
prophets 46–7
propositional revelation 12–27, 34–6
providence see divine providence
Pruss, Alexander 125 n23, 310
Psalms 12, 15, 32, 35, 95, 167, 169, 212 n1
Pseudo-Dionysius 78, 93
Celestial Hierarchy 80
Mystical Theology 78
punishment
crime and 436
 theodicy of 360–1
theory of 435
see also Hell

Qing dynasty (China) 585

quantum mechanics 68, 251
Quine, W. V. 492
Quinn, Philip 313–17, 323, 328 n25, 435, 508
Quinn, Philip and Taliaferro, Charles, Companion to the Philosophy
of Religion, A 79
Qur’an 559, 564
and logic 566–8
and reason 568–9, 571–2

rabbis 548–9
Rachels, James 311
Rahner, Karl 85, 92–3, 95
Ramanuja 578
rationality, religious 62–3, 67
realism
Augustinian 437–40
and regions 199
Rea, Michael 302 n19, 358
real absence: doctrine of 533–5
real presence: doctrine of 512–36
reason, human 87–8
Redeemed Humanity 441, 442, 443, 444–5
redemption 249–53, 256, 434
reductionism 69
reduplicative analysis 455
Reformed theological tradition 220
Régnon, Theodore de 424 n15
Reid, John: The Authority of the Scripture 52 n33
Reitan, Eric 506
relativism 31
Relativity Theory 251, 359 see also Einstein, Albert
reliabilism 23, 24
religion 591
erosion of 63–4
point of 570
and reason 565–6, 568
and science 54–73
religious rationality 62–3, 67
religious studies: schools 492–3
representationalism 443
Resurrection Day of see Day of Resurrection
resurrection, doctrine of 476–86
Revelation, book of 485
revelation see divine revelation
risk see God: and risk
ritualism 31
Romans, Letter to the 129, 246, 249, 255, 258, 444, 446, 447
Rosenkrantz, G. see Ho man, J. and Rosenkrantz, G.
Roughgarden, Joan 249
Rowe, William 81, 219–20, 222, 226, 341, 376, 381–2, 390
Russell, Robert 242

Saadya Gaon 552

Book of Doctrines and Beliefs 544–5, 546
sacri ce 513
Sadra, Mulla 562
saints 485–6
see also Anselm, St; Aquinas, St Thomas; Augustine, St;
Chrysostom, St John; Cyril of Jerusalem, St; Irenaeus, St;
John, St; Paul, St
salvation 32, 36, 45–6, 434, 441
Satan 431–2
Schellenberg, John 382–5
Schleiermacher, Freidrich 505
Schlick, Moritz 2
science
metaphysical principles 70–1
natural 70
and religion 54–73
con ict (warfare) views 60–7, 72
dialogue 67–70, 72
independence 57–60, 72
scientism 67–8
Scripture 39, 48
and atonement 45, 100 n50
and authority 11–13, 21–2, 25, 45, 100 n50
authors of 42, 48, 488 n14
God, revelation in see divine revelation
interpretation of 13–19
see also Bible
Segal, Alan 507
Sellars, Wilfred 2
Senor, Thomas 455, 456–7
Seymour, Charles 497, 498
Shang kings (China) 576
Shankara 578
Sheldrake, Rupert 250
Shi‘as, Islam 567, 569
Sima Tan 579
sin
and free choice 269, 357
original 6 n3, 257, 437–40, 443–7
and revelation 33
and truth 33
see also atonement and Hell
skeptical theism see theism, skeptical
skepticism 86–7
about external world 390–1
moral 391–4
and principles of reason 86–7
Smalley, Beryl 16
Smart, John D.: The Interpretation of Scripture 51 n6
Smith, Quentin 105
Smith, Wilfred Cantwell 576
Socinianism 435
Socinus, Faustus 435, 448 n13
Socrates 320, 566
Song dynasty (China) 580
Song Revival: Confucianism 586, 594
Sorai, Ogyū 585
soul-making 254, 367–9
soul, the 248–9, 255
in Islam 559, 563–4
and resurrection 479–80, 485
Spinoza, Benedict de 545, 553
spirituality: New Confucianism 589
Stenmark, Mikael 67
Stump, Eleonore 139, 151, 157–8, 296, 460, 509 n20
subordinationism 407
su ering
animals 243, 247, 253, 369–70
horri c 388–90
humans 337–8, 339, 340–2, 344, 346, 354, 393, 495
see also evil, problem of
Su sm see mysticism, Islam
Sunnis, Islam 567, 569
supererogation 224, 226, 232, 433–4
Swinburne, Richard
and atonement 433–4
 and the Bible 50, 51 n16
and Christ 463–6, 468–71, 473 n12
nature of 459
omniscience of 469–70
on damnation 496–7
and dei cation 173
and divine omnipotence 168
and divine omnipresence 201–2, 203
and divine omniscience 139, 228
The Existence of God 21
and freedom 234
Revelation 21–2
and human attributes 463–4
and kenosis 465
and petitionary prayer 295–6
The Resurrection of God Incarnate 6 n3
and skeptical theism 386–8
and theodicy 258
and the Trinity 414–15
and two ‘minds’ solutions 468, 470

Talbott, Thomas 500–6

Taliaferro, Charles 203 see also Quinn, Philip and Taliaferro,
Charles
Tanner, Kathryn 454
Taoism see Daoism
Ten Commandments 224
Tertullian 17
Thagard, Paul 71
theism
and atheism 339–42, 346–9
canonical 22
and divine aseity 123
and divine attributes 116–17
and divine personhood 106–7
and evil 334–5, 342–3, 352–5, 367–8
and free will 254
and human su ering 341, 346
hypotheses of 347
and morality 307–8, 310–20
and moral perfection 218–20
Open Theists (Openists) 76 n42, 139–40, 263, 264, 269–74, 281,
356–8, 510 n27 608
and ‘perfect-being theology’ 212
 and petitionary prayer 286–99
Process Theists 282 n3
and science 67
skeptical 344, 354–5, 374–94
Theistic Explanation of Morality (ThEM) 307–8, 310–11, 312, 313,
317, 320–1
theodicy 352–70
animal su ering 369–70
connection-building (CBT) 257–8
free will 244, 254, 362–5
natural consequence 361–2
natural law 365–7
and Open Theism 356–8
punishment 360–1
recent deterrents to 353–5
soul-making 244–5, 254, 367–9
standards for success 355–60
virtuous response 257–8
theology
Eastern Orthodox 250
Hindu 578
philosophical 1
 Chinese (Confucian) 574–94
Islamic 556–72
Jewish 541–53
thermodynamics, second law of 246, 251–2
Thessalonians, First letter to the 484
Thomas Aquinas, St see Aquinas, St Thomas
Thomism 263, 266–9, 270–1, 274–5, 278
Tillich, Paul J. 578
time
cosmic 151
nature of 148–51, 163–4
Timothy
First letter to 169
Second letter to 13, 48
Toland, John 78, 85–6, 87, 96 n3, 98 n36, 98 n41
Tooley, Michael 390
tradition
analytic 1–3
Christian 12–13, 17–9
Islam 568–70
traducianism 438, 446
transubstantiation 514–15, 516, 529
Trembath, Kern Robert, Evangelical Theories of Biblical Inspiration: A
Review and Proposal 44–6, 52 n34
Trent, Council of 515, 529
Trinitarianism 408–16
Constitutional 417–23
Greek (eastern) 412–16
Latin (western) 409–12
part–whole 414–16
and relative identity 417–8
social 412, 413–4, 416, 423
Trinity, the 403–23
adoration of 95
and Christ’s identity 452
and consubstantiality 405–7, 417
doctrine of 86–7, 88–91, 406, 424 n4
logical problem of 405–6
and matter 420
mystery of 79
and omnipresence 212
Relative Identity Strategy 417–18
truth
counterfactuals and 277–82
 double truth issue: Islam 565–6
God and 118
truthmakers 110–17, 125 n17, 190, 197 n75, 278
Tuggy, Dale 426 n36
Tu Wei-ming 579, 591

universalism 499, 500

U.S. News and World Report 493
utilitarianism 221–3

van Huyssteen, J. Wentzel 59, 74 n10

van Inwagen, Peter 131, 362, 380, 457–8, 487 n6, 488 n9
veri cationism 150
virtue theory 228–35, 319–20
volition see divine volitions

Wainwright, William, The Handbook of Philosophy of Religion 79

Walls, Jerry L. 498
Wang Fuzhi 585
Wang Yangming 585
Ward, Keith 242, 250, 256
warrant theory 23–4
Westminster Confession 17, 264
Westminster Shorter Catechism 17
Whitehead, Alfred North 63, 179–80, 195 n54, 578
Wielenberg, Erik 185–6
Wierenga, Edward 136–7, 160, 161, 171–4, 182, 201, 203
Wilkinson, Sir Denys Haig 71
William of Auxerre 169, 180
William of Ockham see Ockham, William of
Williams, Bernard 507–8
Wilsom, David Sloan 62
Wolterstor , Nicholas 3, 4, 135
Divine Discourse 46–7
world-views: dialogue 72–3
Wright, J. Edward 507
Wykstra, Stephen 375

Xunzi 580, 582–3, 587, 592

Zagzebski, Linda 130, 318–19

Zaleski, Carol and Philip 491
Zhang Zai 585–6, 594
Zhongyong (Confucianism) 583–4
Zhou dynasty (China) 576
Zhu Xi 585, 586–7
 Treatise on the Completion of the Mind-heart 587–8
Zimmerman, Dean 488 n10
Zionism, Political 542