Does Physics Answer Metaphysical

Questions?
1
JAMES LADYMAN
1. Introduction
According to logical positivism, so the story goes, metaphysical
questions are meaningless, since they do not admit of empirical
confirmation or refutation. However, the logical positivists did not
in fact reject as meaningless all questions about for example, the
structure of space and time. Rather, key figures such as
Reichenbach and Schlick believed that scientific theories often
presupposed a conceptual framework that was not itself empirically
testable, but which was required for the theory as a whole to be
empirically testable. For example, the theory of Special Relativity
relies upon the simultaneity convention introduced by Einstein that
assumes that the one-way speed of light is the same in all directions
of space. Hence, the logical positivists accepted an a priori
component to physical theories. However, they denied that this a
priori component is necessarily true. Whereas for Kant, metaphys-
ics is the a priori science of the necessary structure of rational
thought about reality (rather than about things in themselves), the
logical positivists were forced by the history of science to accept
that the a priori structure of theories could change. Hence, they
defended a notion of what Michael Friedman (1999) calls the
relativised or the constitutive a priori. Carnap and Reichenbach
held that such an a priori framework was conventional, whereas
Schlick seems to have been more of a realist and held that the
overall relative simplicity of different theories could count as
evidence for their truth, notwithstanding the fact that some parts of
them are not directly testable. All this is part of the story of how
the verification principle came to be abandoned, and how logical
positivism transmuted into logical empiricism.
1
This text is more or less that of a talk of the same title given to the
Royal Institute of Philosophy on 10th February 2006. The relationship
between science and metaphysics is discussed in more detail in Ladyman
and Ross (2007).
179
Yet there is no doubt that the crude view that the positivists
banished metaphysics altogether was considerably influential
despite the more nuanced history that is now being told. For
example, here is John Watkins in 1975:
The counter-revolution against the logical empiricist philosophy
of science seems to have triumphed: I have the impression that it
is now almost as widely agreed that metaphysical ideas are
important in science as it is that mathematics is. (91)
Watkins celebrates the work of historians of science who showed
the decisive role that certain metaphysical speculations have played
in the advancement of science (91). Since he wrote it has become
ever more widely acknowledged that metaphysical questions are
meaningful and important, and metaphysics is now enjoying
something of a renaissance. In philosophy of science, the
acceptance of metaphysics is due in part to the emergence and
dominance of scientific realism. For one definition of metaphysics
is that it is the study of the nature of reality beyond the
appearances of things, and scientific realism is the position that we
ought to believe in the unobservable entities, such as atomic and
subatomic particles, fields, and black holes, postulated by our best
scientific theories. Unobservable entities are now fundamental to a
great deal of science and especially physics, and the theories that
describe these unobservable entities clearly purport to describe
reality beyond the appearances. Therefore, that metaphysical
questions can be answered by physics seems to follow from
scientific realism.
2
This is surely a significant general consideration
(of which more below), but perhaps a more important part of the
explanation of the seriousness with which philosophers of science
now take metaphysics is the fact that a number of fundamental
metaphysical doctrines are widely perceived to be important parts
of the content of physical theories. For example, Special Relativity
is thought to imply that all times are real, and quantum mechanics
is thought to imply indeterminism. (More of these specific cases
below.)
However, other factors have led to metaphysics being regarded as
not merely a meaningful and legitimate part of philosophy, but also
as an autonomous and non-empirical science. One of these is that
philosophers of science began to turn to metaphysics in an attempt
2
Indeed a recent textbook (Couvalis, 1997) on philosophy of science
defines scientific realism as the view that metaphysical questions can be
answered by science.
James Ladyman
180
to explain the continuity of reference between successive scientific
theories that gave competing descriptions of unobservable entities.
The causal theory of reference, and ideas such as rigid designation,
were in part a response to the challenge to scientific realism raised
by the historicism of Kuhn and others. Hence, it began to look as if
scientific realism needed to be grounded in metaphysics. Further-
more, there was a general acceptance in analytic philosophy after
the heyday of Quineanism that modality could not be eliminated
either from philosophy or from the rational reconstruction of
science, and so the metaphysics of possible worlds, causation, and
laws became central areas of philosophical concern. The philoso-
phers pursuing these issues did not detect any prospect of science
being able to settle them, and so nowadays metaphysics is pursued
by many who regard it as sovereign.
2. Autonomous versus Naturalised Metaphysics
It is a commonplace among contemporary philosophers that the
systematic divide among the great early modern philosophers into
rationalists and empiricists is an artefact of the histories of
philosophy from the late nineteenth century that whiggishly
construed rationalism and empiricism as thesis and antithesis prior
to celebrating Kants cunning synthesis. That said, there is no
doubt that we can detect the binary opposition within contempo-
rary philosophy, and as ever no where is it more evident than within
metaphysics. Representing the resurgent voice of the analytic
metaphysicians here is Jonathan Lowe:
... metaphysics goes deeper than any merely empirical science,
even physics, because it provides the very framework within
which such sciences are conceived and related to one another
(2002, vi)
According to him the universally applicable concepts that
metaphysics studies include those of identity, necessity, causation,
space and time. Metaphysics must say what these concepts are and
then address fundamental questions involving them, such as
whether causes can have earlier effects. Metaphysics other main job
according to Lowe, is to systematise the relations among
fundamental metaphysical categories such as things, events,
properties, and so on. Leaving aside the latter, we might reasonably
Does Physics Answer Metaphysical Questions?
181
ask how metaphysical enquiry ought to proceed and Lowe adopts
the familiar methodology of reflecting on our concepts (conceptual
analysis).
Frank Jackson (1998) has also recently defended the idea of
metaphysics as conceptual analysis. The problem these philoso-
phers have is with explaining why we should think that the
products of their activity reveal anything about the deep structure
of reality rather than merely telling us about how we think about
and categorise reality. Even those fully committed to a conception
of metaphysics as the discovery of synthetic a priori truths shy
away from invoking a special faculty of rational intuition that
delivers such knowledge; rather they usually just get on with their
metaphysical projects and leave the matter of explaining the
epistemology of metaphysics for another occasion. Ted Sider
defends this strategy by pointing out that lack of an epistemological
foundation for science and mathematics does not prevent practi-
tioners from getting on with the business of advancing the state of
knowledge in these domains (2001, xv). The obvious rejoinder to
this, which he does not consider, is that we know a tree by its fruits
and mathematics and science have undoubtedly borne fruits of
great value; pure metaphysics has not achieved anything compara-
ble, if it has achieved anything at all.
Furthermore, naturalists have good positive reasons to be
sceptical about a priori metaphysics. There is no reason to believe
that human beings naturally endowed cognitive capacities extend to
the provision of reliable information about the fundamental
structure of reality.
3
So the only kind of metaphysical enquiry that
ought to be taken seriously, if any, is naturalised metaphysics. (Of
course the naturalist concedes that if a metaphysical system is
internally consistent then it can be rejected a priori.)
Lowe has two arguments against this:
(i) ... to the extent that a wholly naturalistic and evolutionary
conception of human beings seems to threaten the very
possibility of metaphysical knowledge, it equally threatens the
very possibility of scientific knowledge ... (6) Since natural
selection cannot explain how natural scientific knowledge is
possible so the fact that it cannot explain how metaphysical
knowledge is possible gives us no reason to suppose that such
knowledge is not possible.
(ii) Naturalism depends upon metaphysical assumptions.
3
For more on this see Ladyman and Ross (2007), 1.1 and 1.2.
James Ladyman
182
In response though note that even if it is granted that natural
selection cannot explain how natural scientific knowledge is
possible, we have plenty of good reasons for thinking that we do
have such knowledge. On the other hand, we have no good reasons
for thinking that metaphysical knowledge is possible. (Recall the
point against Sider above.) It seems that their predilection for a
priori argument leads metaphysicians to discuss the status of
metaphysics compared to science a priori as if we didnt know a
posteriori that the track record of science gives us reason to take it
epistemically seriously that we lack in the case of metaphysics.
With respect to Lowes second argument here, it is enough to point
out that even if naturalism depends on metaphysical assumptions,
the naturalist will argue that the metaphysical assumptions in
question are vindicated by the success of science, by contrast with
the metaphysical assumptions on which autonomous metaphysics is
based which are not vindicated by the success of metaphysics since
it can claim no such success.
According to Lowe, it is the job of metaphysics to tell us what is
possible, but it may be conceded that which of the possible
fundamental structures of reality exists can be answered only with
empirical evidence. The problem with this is that philosophers
have often regarded as impossible states of affairs that science has
come to entertain. For example, metaphysicians confidently
pronounced that non-Euclidean geometry is impossible, that it is
impossible that there not be deterministic causation, that non-
absolute time is impossible, and so on. Furthermore, there is no
agreement now among metaphysicians about what is metaphysically
possible. For example, some metaphysicians believe that infinitely
divisible matter is possible and others that it is not.
In any case, suppose that naturalised metaphysics is the only
option. It remains an open question whether it is possible to learn
metaphysical lessons from science. Since metaphysics does not give
rise to empirical predictions it cannot be directly confirmed by a
physical theory. However, metaphysical theories can be incompat-
ible with physical theories, and hence the former can be ruled out or
at least undermined by the confirmation and adoption of the latter.
For example, Cartesian metaphysics is incompatible with Newto-
nian physics because the former has it that space is filled with a
plenum and that all action is action by contact, whereas the latter
involves particles moving in a void subject to the force of gravity
that seems to act at a distance. For Watkins, metaphysics is part of
the hard core of a Lakatosian research programme. It cannot be
directly refuted but if the research programme persistently
Does Physics Answer Metaphysical Questions?
183
degenerates then the hard core must eventually be abandoned.
Arguably this happened to Aristotelian and Cartesian metaphysics,
and then later it happened to the metaphysical framework of
classical physics in the face of relativity and quantum physics. As
Sider says in a more naturalistic mood: ... in cases of science
versus metaphysics, historically the smart money has been on
science ... (2002, 42). (Sider seems to think that metaphysical
knowledge can be a priori but that metaphysics is also required to
be compatible with current science.)
Let us grant then that the failure of a metaphysical framework to
help in the production of good science can count as a reason against
it. Is it also plausible that the success of science based on a
particular metaphysical framework provides some evidence for the
latter? For example, the success of materialism as a research
programme in science in the seventeenth and eighteenth centuries
was arguably evidence for it. (Of course, the conception of matter
integral to materialism, namely that of extended substance, became
untenable in the nineteenth and twentieth centuries.)
Katherine Hawley (2006) helpful distinguishes a number of
possible positions one might take on the relationship between
metaphysics and science. Hawley assumes that it is sometimes at
least possible to provide reasons for and against metaphysical
claims and that science is not the only source of such reasons; she
therefore thinks that metaphysical enquiry can be reasonable even
where it outruns the scope of science (454). Purely a priori reasons
for metaphysical claims may involve appeals to intuition, explana-
tory power, simplicity, or perhaps the results of conceptual analysis.
As was mentioned above, naturalists will be suspicious of appeals to
intuitions not least because intuitions vary so much, and also
because even were we agreed in our intuitions, they doubt that
there is any reason to believe that our intuitions track the truth
about metaphysical matters any more than they do about motion,
gravity or the constitution of material objects. When it comes to
explanatory power, naturalists must agree that inference to the best
explanation is indispensable in science and so the question can only
be whether there is any difference between metaphysics and science
in respect of explanation. There is an important such difference,
namely that what metaphysical hypotheses are supposed to explain
are often metaphysicians intuitions or common-sense beliefs. In
the light of this, naturalists ought to be very sceptical about the
extension of the methodology of inference to the best explanation
James Ladyman
184
into metaphysics, and about the worth of pursuing metaphysical
enquiry where it goes beyond the scope of scientific enquiry.
4
However, we can at least agree with Hawley that an outright
sceptic about inference to the best explanation like Bas van Frasssen
must eschew metaphysics, just as he does scientific realism. Hence,
some form of realism about science is a necessary condition for
believing that in principle science can answer metaphysical
questions. According to Hawley, optimism is the view that the
involvement of a metaphysical claim in an empirically successful
scientific theory provides some reason to think that the claim is
true (456). This idea that in principle science can answer
metaphysical questions is the negation of Hawleys Radical
Pessimism. She thinks that scientific realism is sufficient for
optimism because there is no in-principle difference between
claims about unobservable scientific entities and metaphysical
claims (459). The idea is that since the theoretical claims of science
concerning unobservable entities go beyond the empirical content
of the relevant theories, the scientific realist is committed to the
idea that the empirical success of theories gives us reason to believe
their theoretical content. The onus is on the sceptic about
metaphysics to point out some relevant semantic or epistemic
difference between the theoretical content and the metaphysical
content of theories and there is good reason to think this cannot be
done. One of way of thinking about this is in terms of the Quinean
notion of the web of belief. Metaphysical hypotheses are as much
part of the web as theoretical scientific hypotheses, and since
neither can be confirmed or refuted in isolation, in principle they
have the same epistemic status.
Let us grant all this for the moment. It follows that the scientific
realist must deny radical pessimism and so they must chose
between the following two positions:
(1) Metaphysical claims have in fact been established by science.
(2) Although in principle metaphysical claims might be estab-
lished by science, in fact they have not been.
(Hawley talks not of science establishing metaphysical claims but of
it providing some reason to think they are true. However, this is too
weak: The sceptic about metaphysics may concede that science
gives us some reason to believe a particular metaphysical claim, but
deny that the reasons in question are anywhere near compelling.)
4
Again see Ladyman and Ross (2007), 1.1 and 1.2 and also van
Fraassen (2002), chapter 1.
Does Physics Answer Metaphysical Questions?
185
If it is supposed that scientific realism implies that metaphysical
hypotheses have been established by science, and it turns out that
metaphysical hypotheses have not in fact been established by
science, then by modus tollens, scientific realism must be false.
Hence, close attention to the role of metaphysics in physics and to
the fate of metaphysical hypotheses in the history of physics may
bolster scepticism, not just about metaphysical knowledge, but also
about scientific realism. On the other hand, if it is supposed that
scientific realism only requires that metaphysical hypotheses can in
principle be established by science, but they have not in fact been,
then it may be objected that the form of scientific realism that
results is too attenuated to really be worth the name. After all, as
Hawley says, there is no clear demarcation between a theoretical
claim and a metaphysical one, and indeed claims about unobserv-
able entities become empty if nothing is said about their
metaphysical status. For example, the claim that light is transmitted
as waves in the ether is empty without some account of what the
ether is supposed to be.
5
3. Metaphysical questions that might be thought to be
answerable by physics:
Here are some examples:
(i)Are all times real?
(ii)Does time flow?
(iii)Is there a global asymmetry in time in the universe?
(iv)Is time travel possible?
(v)Is the world deterministic
6
?
(vi)Can a cause succeed its effect in time?
(vii)Is there action at a distance or does all causation happen
locally in space and time?
(viii)Is the identity of indiscernibles true?
(ix)Do space (and/or time) exist independently of their material
contents?
(x)Are space and time discrete or continuous?
(xi)What is the geometry of space(time)?
5
For more on the status of the metaphysical content of theoretical
claims see Ladyman and Ross (2007), 2.2.
6
Determinism is here understood as a modal thesis: given the state of
the world at some time, there is only one way it could be at any later (or
earlier) time.
James Ladyman
186
(xii)How many dimensions are there to spacetime
(xiii)What is the topology of space(time)?
(xiv)Is the universe finite or infinite?
(xv)Are there parallel universes?
4. The Philosophy of Time
Questions (i) to (iv) all concern time. They are related but
nonetheless distinct although often conflated. In relation to (i), call
the view that all events are real eternalism, the view that only the
present is real presentism, and the view that all past and present
events are real cumulative presentism. (The latter is defended by
Michael Tooley (2000) although not under this name). (ii) is often
expressed in terms of whether there is temporal passage or
objective becoming. Those who believe in the passage of time or
objective becoming often also believe that the process of becoming
is that of events coming into existence and going out of existence,
but this need not be so; to suppose there is becoming, one need only
believe that there is some objective feature of the universe
associated with the passage of time. Objective becoming could be
like a light shining on events as they are briefly present, and is
therefore compatible with eternalism.
7
On the other hand, both
presentism and cumulative presentism entail a positive answer to
question (ii), since if events do come into existence, whether or not
they then stay existent or pass out of existence, this is enough to
constitute objective becoming. Since McTaggarts famous argu-
ment for the unreality of time, metaphysicians have often discussed
(i) and (ii) in the context of a question in the philosophy of
language namely: Is tensed language reducible to tenseless
language, or does tensed language have tenseless truth conditions?
(*) However, the philosophy of language cannot settle the
metaphysical issues as Michael Tooley (2000) argues. So even
though the standard opposition is between those who answer no to
(i), yes to (ii), and no to (*) on the one hand (the defenders of
McTaggarts A-series), and those who answer yes to (i), no to
(ii), and yes to (*) (the defenders of McTaggarts B-series), a
variety of more nuanced positions are possible. (Note that it is not
clear on reflection that the negation of presentism is a necessary
condition for the possibility of time travel.)
7
Tim Maudlin (2002) argues that the passage of time may be an
objective feature of the universe even if eternalism is true.
Does Physics Answer Metaphysical Questions?
187
In relation to (iv), clearly if (i) or (ii) are answered positively then
that is enough to privilege a particular direction in time. However,
eternalism and the denial of objective becoming are compatible
with time having a privileged direction, since there could be some
feature of the block universe that has a gradient that always points
in some particular temporal direction. For example, the entropy of
isolated subsystems of the universe, or the universe itself, might
always increase in some direction of time. Another well known
possible source of temporal direction was proposed by Reichenbach
(1956) who argued that temporal asymmetry is grounded in causal
asymmetry: in general, correlations between the joint effects of a
common cause are screened off by the latter but the joint causes of
a common effect are uncorrelated. Although some have claimed
that Reichenbachs Principle of the Common Cause is violated, not
least by the behaviour of entangled states in quantum mechanics
(see for example van Fraassen 1991), such considerations are
sufficient to show that conceptually the question of the direction of
time must be separated from questions about eternalism. However,
it may be that no physical meaning can be attached to the idea of
the direction of time in the whole universe, because no global time
co-ordinate for the whole universe can be defined. This seems to be
implied by special relativity.
The status of time in special relativity (SR) differs from its status
in Newtonian mechanics in that there is no objective global
distinction between the dimensions of space and that of time.
Spacetime can be split into space and time, but any such foliation is
only valid relative to a particular inertial frame, which is associated
with the Euclidean space and absolute time of the co-ordinate
system of an observer. This seems to imply eternalism, since if
there is no privileged foliation of spacetime, then there is no global
present, and so the claim that future events are not real does not
refer to a unique set of events.
8
Furthermore, many have argued
that, since SR implies the relativity of simultaneity, whether or not
two events are simultaneous is a frame-dependent fact, and
therefore there is no such thing as becoming.
9
However, this is too
quick. It is possible to advocate a form of becoming that is relative
8
Of course, one could argue that the very notion of reality must be
relativised to observers, but this is to give up on the kind of metaphysics at
issue.
9
The literature on these topics is voluminous but among the most
influential papers are Gdel (1949), Putnam (1967), and Stein (1968) and
(1989).
James Ladyman
188
to observers or events, so strictly speaking it is only absolute
becoming that is ruled out by the lack of absolute time in special
relativity. Since the light cone structure of Minkowski spacetime is
Lorentz invariant, it can be regarded as absolute. It is easy then to
define a notion of the open future of an event E, since any event E
in the forward lightcone of E will have events in its backwards
lightcone that are not in the backwards lightcone of E, meaning
that there is a sense in which it can be claimed that E is not
determinate at E. This notion of becoming is objective in the sense
that all observers will agree about which events are in the open
future of a given observer at a particular point in his or her history,
because all observers agree about the light cone structure of
spacetime.
10
In any case there is a fundamental problem with drawing
metaphysical conclusions about the nature of time from special
relativity, namely that it is a partial physical theory that cannot
describe the whole universe.
11
For that we must turn, in the first
instance, to general relativity (GR), and the implications of that
theory for time are not clear. This is because GR gives us field
equations that are compatible with a variety of models having
different global topological features, and different topological
structures may have very different implications for the metaphysics
of time. For example, if the topology of the universe is globally
hyperbolic then it is possible to define a single global foliation of
spacetime for it; otherwise it may not be. Clearly we must then turn
to cosmological models of the actual universe, of which there are
many compatible with the observational data. As yet there is no
agreement about which of these is the true one. Highly
controversial issues about the cosmological constant and so-called
10
Another possibility is to argue that while Special Relativity is
empirically adequate, the empirical evidence is nonetheless compatible
with the existence of a privileged foliation. This would be to advocate
along the lines of that originally proposed by Lorentz and Fitzgerald.
This is the strategy adopted by Tooley (2000) for example. Naturalism
forbids the revision of scientific theories on purely philosophical grounds,
so the proposal of a privileged foliation contra to SR requires a scientific
motivation. One possible scientific motivation is the adoption of a solution
to the measurement problem that posits a preferred frame of reference.
Another is the identification of foundational problems with the account of
relativistic phenomena in the Minkowski spacetime framework (see Bell
1987 and Brown 2005).
11
Syder (2001) is a recent example of a metaphysician who confines
discussion to the implications of Special Relativity.
Does Physics Answer Metaphysical Questions?
189
dark energy, dark matter, and the nature of singularities, as well as
the various approaches to the search for a theory of quantum
gravity, all bear on the question of whether spacetime will turn out
to admit of a global foliation, and hence on whether absolute time is
physically definable. Even if it does turn out to be definable, there
remains the question of whether such a definition ought to be
attributed any metaphysical importance. For example, it is possible
to define something called cosmic time which is based on the
average properties of the universes global matter distribution
under the expansion of the universe. Some have argued that we can
regard Cosmic Time as giving us a privileged foliation (Lucas and
Hodgson, 1990). However, others argue that the fact that such a
foliation can be defined gives us no reason to regard it as having an
objective significance (Berry, 1989), not least because it is based on
averages that have nothing to do with the phenomenological
experience of everyday simultaneity for objects whose states of
motion are not the same as the state of motion of the galaxies in our
region of the universe with respect to which cosmic time is defined
(see Bourne, 2004).
Non-relativistic many-particle quantum mechanics does not
directly bear on the philosophy of time since the status of time in
the formalism is not novel in the same way as in relativity. However,
it has often been argued that quantum physics is relevant to
questions about the openness of the future, becoming, and the
direction of time, because of the alleged process of collapse of the
wavefunction. Since Heisenberg (1962) it has been popular to claim
that the modulus squared of the quantum mechanical amplitudes
that are attached to different eigenstates in a superposition
represent the probabilities of genuinely chancy outcomes, and that
when a measurement is made there is an irreversible transition from
potentiality to actuality in which the information about the weights
of the unactualised possible outcomes is lost forever. Hence,
measurement can be seen as constituting irreversible processes of
becoming that induce temporal asymmetry. However, quantum
measurements need not be so understood not least because some
deny that collapse is a genuine physical process and also because
realism about the wavefunction is highly contentious. Similarly, if
there is no collapse, as in the Everett interpretation, then again
there is no temporal asymmetry in quantum mechanics.
The upshot seems to be that the status of the arrow of time in
quantum mechanics is open.
There is also a vast literature about whether or not the second
law of thermodynamics represents a deep temporal asymmetry in
James Ladyman
190
nature. The entropy of an isolated system always increases in time,
and so this seems to be an example of the arrow of time being
introduced into physics. If the whole universe is regarded as an
isolated object, and if it obeys the second law, then it would seem
that there is an objective arrow of time in cosmology. However, it is
not clear what the status of the second law is with respect to
fundamental physics. One possibility is that the second law holds
only locally, and that there are other regions of spacetime where
entropy is almost always at or very near its maximum. (Boltzman
himself thought this was the case in his later years.) Even if
thermodynamics seems to support the arrow of time, it is deeply
puzzling how this can be compatible with an underlying physics
that is time asymmetric. Conservative solutions to this problem
ground the asymmetry of the second law in boundary conditions
rather than in any revision of the fundamental dynamics. The most
popular response is to claim that the law does indeed hold globally
but that its so doing is a consequence of underlying time-reversal
invariant laws acting on an initial state of the universe that has very
low entropy. This is called the Past Hypothesis by Albert (2000).
A much more radical possibility is that the second law is a
consequence of the fact there is a fundamental asymmetry in time
built into the dynamical laws of fundamental physics. Given the
outstanding measurement problem in quantum mechanics those
who propose radical answers to problems in thermodynamics and
cosmology often speculate about links between them and the right
way of understanding collapse of the wavefunction. Both the
question of eternalism, and the arrow of time, lead inexorably to
cosmology and thence to the realm of quantum gravity.
5. Quantum Gravity
12
A theory of quantum gravity must do all of the following: say what
happens in nature at the Planck length (10
-33
cm); recover GR as a
low-energy limit; and provide a background spacetime, at least
phenomenologically, for conventional quantum theories. What else
it should do is a matter of great contention. Some, such as
advocates of string theory and M-theory like Brian Greene (2004),
think it must also unify the four fundamental forces of nature;
12
Introductions to the philosophy of quantum gravity include
Callender and Huggett (2001) especially the introduction, and Rovelli
(2004), chapter 1. Wallace (2000) is also very helpful.
Does Physics Answer Metaphysical Questions?
191
others, such as Smolin (forthcoming), argue that, in the first
instance at least, it need only amount to a quantised version of GR.
A further question is whether quantum gravity must also be a
cosmological theory of one (unique and actual) universe, rather
than allowing for models representing a variety of universes.
Quantum gravity must reconcile a number of profound tensions
between GR and quantum theories. Most obviously, quantum
physics is the physics that best describes the phenomena when we
look at very short length scales, and GR is physics that was
specifically designed with a distinction between local and global
properties of spacetime in mind, and sought to describe deviations
from the topological, geometrical and metrical properties of
Minkowski spacetime that only show up in large scale structure
this is the scale tension. Secondly, GR depends on the identification
of inertial and gravitational mass, and the equivalence between
accelerating and gravitational frames, whereas quantum theory was
originally developed to account for the interaction between
electro-magnetic radiation and matter. Initially, it was only the
energy states of matter that were quantised, but subsequently it has
proved possible, with differing degrees of success, to quantise all
the fundamental forces, with the exception of gravitythis is the
force tension. Thirdly, relativistic theories obey the condition that
there are invariant, and hence objective causal pasts for events,
whereas in QM there are nonlocal correlations that some regard as
evidence of action at a distancethis is the causal tension. Finally,
there is the radically different status of time in quantum theory
versus GR. In the former, time is a parameter external to all
physical systems; in the latter it is a co-ordinate with no particular
physical significance. More specifically there is something called
The Problem of Time the upshot of which is that theories of
quantum gravity are in danger of saying there can be no change in
the universe over time (I return to this below).
Depending on who one listens to, string theory has either already
led us a considerable distance down the road to a complete theory
of quantum gravity, or it has achieved absolutely nothing that
counts as physics rather than mathematics. String theorists have
followed the methodology that was used in the construction of
quantum field theories, namely the search for fundamental
symmetries. If the string theory vision is correct then the ultimate
fundamental physics will describe the universal symmetries of the
universe. (Another important commonality between string theory
James Ladyman
192
and classical and quantum physics is that they posit a continuous
space and time, which is departed from by some rival
programmessee below.)
Lee Smolin is highly critical of string theory and argues that it is
not falsifiable in the sense that it makes no falsifiable predictions
for doable experiments (2006, 197). He claims there are no
fundamental global symmetries, on grounds that those theories that
posit them are not fully empirically adequate. His view is that the
two big ideas that drive string theory, namely unification and
symmetry, have run their course, and that there have been no
substantially new results in particle physics since 1975. He points
out that the standard model has so many adjustable parameters that
any likely experimental data from particle accelerators can be
accommodated by it.
13
He also emphasises that there are at least
10
100
possible string theories, and argues that all the ones that have
been studied disagree with the data. Super-symmetric string theory
has 105 free parameters. Hence, Smolin claims, partisans will be
able to maintain that whatever comes out of the next generation of
particle accelerators confirms super-symmetry.
He also criticises string theory for being background dependent
in the sense that it relies on a background spacetime structure.
Smolin predicts that the correct theory of quantum gravity will be
relational in the sense that it wont posit any background structure
which does not change with time but which is necessary for the
definition of kinematical quantities and dynamical laws (ibid.).
Newtonian mechanics, SR, quantum theories including quantum
field theories, and string theory are all background-dependent and
rely on various structures that are outside the scope of the
dynamics of the theory. For example, in ordinary quantum
mechanics the spacetime and the algebraic structure of Hilbert
space are part of the background structure. On the other hand, GR,
understood as a cosmological theory of the whole universe, is a
relational theory in the sense that the physically important
structural features of the theory are dynamical. The only
background structure in GR consists of the dimensionality, the
differential structure, and the topology.
String theorists now seem to have accepted that background
independence is a desideratum. Brian Greene speculates about a
background free version of string theory, and the search for
so-called M-theory is partly motivated by the need for a way of
13
There are at least 19 free parameters in the standard model plus
almost as many from cosmology.
Does Physics Answer Metaphysical Questions?
193
thinking about strings that does not treat them as vibrations in a
background spacetime. However, no such theory yet exists.
Meanwhile, Smolin has inspired a significant minority of research-
ers to seek background independence in other approaches to
quantum gravity, and he shows how this notion plays out in the
context of a variety of these theories. He suggests that the history
of physics testifies to the success of the pursuit of background
independence. It is true that progress has sometimes been made in
physics by eliminating background structure. SR eliminated the
background structure of absolute space and time, and then GR
eliminated the background structure of Minkowski spacetime. On
the other hand, there have been many background dependent
theories that have been highly successful, including quantum
theories. Smolin himself concedes that GR is background
dependent in certain respects. Furthermore, consider the success of
the pursuit of symmetry and unification of forces that motivates
string theory, in generating the standard model and the unified field
theories based on the knitting together of the symmetry groups
previously discovered to be governing the separate forces. Smolin
and Greenes dispute can be construed as concerning which of the
following two desiderata for fundamental physics holds trumps:
symmetry or background independence. The empirical evidence is
equivocal, to say the least.
Among background independent approaches the most well-
known is canonical quantum gravity. This approach seeks a
quantum theory of gravity, but not necessarily a unification of all
the fundamental forces. This gives rise to the famous Wheeler-
DeWitt equation, and the infamous Problem of Time: the physical
states of the universe must be time-independent, and so nothing
changes. The latest version of the canonical programme is loop
quantum gravity. The pioneers of this approach include Abhay
Ashtekar, Carlo Rovelli, John Baez, and Lee Smolin. Other
approaches include:
Causal set theory: This is a background independent approach
motivated by the assumption that at the Planck scale spacetime
geometry will be discrete, and by the fact that a discrete causal
structure of events is almost sufficient to define a classical General
Relativistic spacetime (see Malament 2006). The formalism models
spacetime as a partially ordered set of primitive elements with a
stochastic causal structure representing the probabilities for future
elements to be added to a given element. The volume of
spacetime is then recovered from the number of elements. The
dynamical structure is compatible with eternalism because the
James Ladyman
194
whole of spacetime can be considered as a single mathematical
structure, and temporal relations regarded as just the order of
elements. The probabilistic structure is required to be local. It can
be shown that a classical spacetime can always be approximated by
a causal set. However, the converse does not hold and this is a major
problem for this approach (see Smolin 2006). Note that a causal set
is a partially ordered set where the intersection of the past and
future of any pair of events is finite: [T]he fundamental events
have no properties except their mutual causal relations. (210).
Causal triangulation models: These models use a combinatorial
structure of a large number of 4-simplexes (the 4-d version of a
tetrahedron), from which a classical spacetime will emerge as a low
energy limit, and from which quantum theory can be recovered if
background assumptions are made. Interestingly, the latest
simulations of spacetime emerging dynamically from these models
generate it as four-dimensional on large scales, but two-dimensional
at short distances (and it is known that a quantum theory of gravity
is renormalisable in two dimensions).
Topological quantum field theories, twister theory, and non-
commutative geometry: these approaches are highly abstract and
speculative at present.
Another approach worth mentioning has the unique feature that
the temporal dimension is abandoned altogether, namely Julian
Barbours relationism: Time is supervenient on change, but change
is just differences between distinct instantaneous three dimensional
spaces.
All of these research programmes use new and highly abstract
mathematical structures to describe the universe, and theorists
hope to get the familiar behaviour of spacetime and quantum
particles to emerge as limiting behaviour. It seems clear that we
cannot yet say what the metaphysical implications of quantum
gravity are, but the possibilities range from eleven dimensions to
two, from a continuous fundamental structure to a discrete one, and
from a world with universal symmetries to one with none.
With respect to the metaphysics of time, it seems that it is an
open question whether there is a objective global asymmetry in
time, and whether such dynamical structure as there is in the
universe reflects a fundamentally tensed reality or whether
eternalism is true. (Mathematically, perhaps the real issue is
between threeor moreplus one dimensions and four.) If
M-theory is the correct theory of quantum gravity then there will
be universal symmetries that are not time dependent, and they may
define a background independent structure. If there is no
Does Physics Answer Metaphysical Questions?
195
background independent structure and asymmetry in time is part of
fundamental physics, then it may be that there is dynamics all the
way down and reality is fundamentally tensed.
One possible motivation for the dynamical view is a principle that
van Fraassen sometimes seems to endorse, namely that there is
nothing thats both perfectly general about all of reality and also
true. This coheres with the idea that there is dynamics all the way
down in the universe, since any fundamental properties that hold
generally would necessarily be time-independent and hence amount
to background structure. Consider Smolin: The universe is made
of processes, not things (2001, 49). Smolin insists that a lesson of
both relativity theory and quantum theory is that processes are
prior to states. Classical physics seemed to imply the opposite
because spacetime could be uniquely broken up into slices of space
at a time (states). Relativity theory disrupts this account of
spacetime and in QM nothing is ever really still it seems, since
particles are always subject to a minimum amount of spreading in
space and everything is flux in quantum field theory, within which
even the vacuum is the scene of constant fluctuations.
However, one of the lessons of quantum gravity is that some
philosophersespecially standard scientific realistshave jumped
to overly strong metaphysical conclusions on the basis of not taking
account of all the possibilities still held open by physics. There are
two leading examples where what some philosophers treated as
decisive rulings from physics are now questioned. One of these is
the case we have just been discussing, namely, whether or not we
live in a block universe. The other is the alleged discovery by
quantum theory that the world is not deterministic. In Bohm
theory and the Everett approach, the world comes out deterministic
after all. Clearly, theories that seem to wear their metaphysical
implications on their sleeves often turn out to admit of physical
reconstruction in different terms. Many physicists have attempted
to resolve tensions between QM and GR by seeking what can be
regarded as the key metaphysical truth that lies behind each
theorys empirical success. For example, Barbour and Smolin think
that relationism is the basis for the success of relativity theory.
Often it is argued that the truth of quantisation that lies behind the
empirical success of QM means that theorists should pursue
discrete structures of space and time as in the programmes of
causal set theory. If each instant is ontologically discrete then why
should the timeline be continuous? On the other hand, the key
insight of QM might also be regarded as the superposition
principle, and the consequent problem of entangled states.
James Ladyman
196
So it is not clear which aspects of the metaphysical foundations
of contemporary physicsfor example, continuous space and time,
or four-dimensionalismwill be preserved in quantum gravity.
Given the diversity of philosophical and foundational presupposi-
tions and implications currently abroadfor example, ranging
from a return to absolute time to nihilism about time, and from ten
dimensional continuous spaces to discrete graphsthere is little
positive by way of implications for metaphysics that we can adduce
from cutting edge physics. However, when it comes to negative
lessons, there is more to be said. None of the existing contenders
for a theory of quantum gravity is consistent with the idea of the
world as a spatio-temporal manifold with classical particles
interacting locally. This is not surprising to anyone who has
thought about the implications of Bells theorem. It is important
that this is not a theorem about quantum mechanics, but rather tells
us something about any possible empirically adequate successor to
quantum mechanics, namely that it cannot be both local and posit
possessed values for all measurable observables. We are justified in
treating as unmotivated the idea that any theory of quantum
gravity will be a local realist theory, and we should restrict
consideration in metaphysics to theories that are compatible with
the violation of Humean supervenience implied by entanglement.
6. Scientific Realism and the Metaphysical Content of
Physical Theories.
Here is Larry Sklar:
... it is a great mistake to read off a metaphysics superficially
from the theorys overt appearance, and an even greater mistake
to neglect the fact that metaphysical presuppositions have gone
into the formulation of the theory in the first place. (Sklar 1981,
131)
Steven French is another philosopher of physics who has expressed
scepticism about the attempt to learn metaphysical lessons from
physics. French argues that it is not possible to read the answers to
metaphysical questions off our best physics because there is an
underdetermination of metaphysics by physics (1998).
It turns out that when we examine any of the instances of alleged
metaphysical knowledge being delivered by scientific theories, there
are always a number of extra assumptions needed to derive the
conclusion which we can contest. So, for example, Einstein
Does Physics Answer Metaphysical Questions?
197
contested indeterminism in the face of quantum mechanics by
denying that the latter is a complete description of reality. In
general, there is the problem of scientific theories underdetermin-
ing metaphysics. In the case of quantum mechanics, for example, it
is widely known following the work of John Bell that no local
deterministic hidden variable theory can replace quantum mechan-
ics and still be empirically adequate. However, a nonlocal hidden
variables theory can do so, as is the case with Bohm theory. So we
see that a metaphysical package may be ruled out or in, but not
metaphysical assumptions taken singly.
It is worth remembering that scientific theories have issued
metaphysical claims before, only for subsequent developments to
have shown the theories in question to be erroneous in certain
fundamental respects. So, for example, Fresnels ether theory is
now regarded as featuring a central theoretical term, ether, which
does not refer to anything in the world. Given this the best we can
say about our best theories is that they are approximately true. This
may give cause for doubt that we ought to believe their
metaphysical implications.
Recent realist responses to the arguments against scientific
realism from theory change have separated the parts of theories
that essentially contributed to their novel predictive success from
the idle parts of theories. The problem for our project of learning
metaphysics from physics is that the idle parts often turn out to be
the metaphysical parts. For example, Psillos (1996) argues that in
the case of caloric theory all the important predictive successes of
the theory were independent of the assumption that heat is a
material substance, and in the case of the ether, the predictive
success of Fresnels wave theory of light were independent of the
assumption that the ether is a material solid. The other problem
with the historical record is that the pessimistic meta-induction
becomes much stronger if we confine ourselves to considering the
fate of metaphysical posits in the history of science rather than
theoretical posits in general. Claims about the existence of
unobservable entities may not often be renounced but claims about
the metaphysical nature often are.
Other problems in reading off our metaphysics from our best
science include the following:
(a) There is no fundamental unity in physics since quantum
mechanics and general relativity have not so far been
conjoined consistently. This might suggest that we should
James Ladyman
198
suspend judgement about the metaphysical implications of
both theories until we see which are carried over to their
successor.
(b) The measurement problem in quantum mechanics means that
we lack a consistent interpretation of the theory that is agreed
upon by most physicists.
(c) Scientific theories often come in multiple formulations which
may suggest different metaphysical pictures, and theories are
only applied to the world via models. So we might doubt
whether the empirical success of science ought to make us
believe that a particular fundamental theory is true.
(d) Metaphysical component of scientific theories makes the
problem of defining approximate truth come to the fore.
The topic of this paper was addressed by Sir James Jeans in 1942 in
his book Physics and Philosophy. Towards the end comes this
passage:
There is a temptation to try to round off our discussion by
summarizing the conclusions we have reached. But the plain fact
is that there are no conclusions. If we must state a conclusion, it
would be that many of the former conclusions of nineteenth-
century science on philosophical questions are once again in the
melting pot. (216)
He continues:
... physics and philosophy are at most a few thousand years old,
but probably have lives of thousands of millions of years
stretching away in front of them. They are only just beginning to
get under way, and we are still, Newtons words, like children
playing with pebbles on the sea-shore, while the great ocean of
truth rolls, unexplored, beyond our reach. (217)
I can offer no better conclusion than his:
... to travel hopefully is better than to arrive. (217)
References
Albert, D. (2000), Time and Chance, Cambridge, MA: Harvard University
Press.
Berry, M. (1989), Principles of Cosmology and Gravitation, Bristol: Adam
Hilger.
Does Physics Answer Metaphysical Questions?
199
Bourne, C. (2004), Becoming inflated, The British Journal for the
Philosophy of Science 55, 107119.
Callender, C. and Huggett, N. eds. (2001), Physics Meets Philosophy at the
Planck Scale: Contemporary Theories in Quantum Gravity, Cambridge:
Cambridge University Press.
Couvalis, S. (1997), The Philosophy of Science: Science and Objectivity.
London: Sage.
Dummett, M. (2000), Is Time a Continuum of Instants?, Philosophy 75,
497515.
French, S. (1998), On the withering away of physical objects, in E.
Castellani, ed., Interpreting Bodies: Classical and Quantum Objects in
Modern Physics, 93113. Princeton: Princeton University Press.
Friedman, M. (1999), Reconsidering Logical Positivism, Cambridge:
Cambridge University Press.
Gdel, K. (1949), A remark about the relationship between Relativity
Theory and idealistic philosophy, in A. Schilpp, ed., Albert Einstein,
Philosopher-Scientist, 557562, La Salle: Open Court.
Greene, B. (2004), The Fabric of the Cosmos: Space, Time and the Texture
of Reality, London: Allen Lane.
Hawley, K. (2006), Science as a Guide to Metaphysics, Synthese, 149,
451470.
Heisenberg, W.(1962), Physics and Philosophy, New York: Harper and
Row.
Jackson, F. (1998), From Metaphysics to Ethics: A Defence of Conceptual
Analysis, Oxford: Oxford University Press.
Jeans, J. (1943), Physics and Philosophy, Republished by New York:
Dover 1981.
Ladyman, J. and Ross, D. (2007), Every Thing Must Go: Metaphysics
Naturalised, Oxford: Oxford University Press.
Lowe, E.J. (2002), A Survey of Metaphysics, Oxford: Oxford University
Press.
Malament, D. (2006), Classical General Relativity, in J. Butterfield and J.
Earman, eds., Handbook of the Philosophy of Physics, Elsevier.
Lucas, J. and Hodgson, P. (1990), Spacetime and Electromagnetism,
Oxford: Oxford University Press.
Maudlin, T. (2002b). Remarks on the passing of time. Proceedings of the
Aristotelian Society 102: 237252.
Putnam, H. (1967), Time and physical geometry, Journal of Philosophy
64, 240247.
Psillos, S. (1996), Scientific Realism: How Science Tracks Truth, London:
Routledge.
Reichenbach, H. (1956), The Direction of Time. Berkeley: University of
California Press.
Rovelli, C. (2004), Quantum Gravity. Cambridge: Cambridge University
Press.
James Ladyman
200
Sider, T. (2001), Four-Dimensionalism: An Ontology of Persistence and
Time, Oxford: Oxford University Press.
Sklar, L. (1981), Time, Reality and Relativity, in R. Healey (ed.),
Reduction, Time and Reality, Cambridge: Cambridge University Press.
Smolin, L. (2006), The relational idea in physics and cosmology, in D.
Rickles, S. French and J. Saatsi, Structural Foundations of Quantum
Gravity, Oxford: Oxford University Press.
Stein, H. (1968), On Einstein-Minkowski space-time, Journal of
Philosophy 65, 523.
Stein, H. (1991), On relativity and openness of the future.
Tooley, M. (2000), Time, Tense and Causation. Oxford: Oxford University
Press.
Van Fraassen, B.C. (1991), Quantum Mechanics: An Empiricist View.
Oxford: Oxford University Press.
Van Fraassen, B.C. (2002), The Empirical Stance. New Haven: Yale
University Press.
Wallace, D. (2000), The quantization of gravityAn introduction
unpublished manuscript. available at arXiv:gr-qc/0004005 v1.
Watkins, J. (1975), Metaphysics and the Advancement of Science, The
British Journal for the Philosophy of Science, 26, 91121.
Does Physics Answer Metaphysical Questions?
201