Moore, Michael - Justifying Retributivism PDF
Moore, Michael - Justifying Retributivism PDF
Moore, Michael - Justifying Retributivism PDF
15 1993
Content downloaded/printed from
HeinOnline (http://heinonline.org)
Thu Jul 17 14:55:14 2014
-- Your use of this HeinOnline PDF indicates your acceptance
of HeinOnline's Terms and Conditions of the license
agreement available at http://heinonline.org/HOL/License
-- The search text of this PDF is generated from
uncorrected OCR text.
-- To obtain permission to use this article beyond the scope
of your HeinOnline license, please use:
https://www.copyright.com/ccc/basicSearch.do?
&operation=go&searchType=0
&lastSearch=simple&all=on&titleOrStdNo=0021-2237
JUSTIFYING RETRIBUTIVISM
Michael S. Moore*
I shall address two concerns in this paper: first, what retributivism
is, and second, how one justifies retributivism as the only proper theory
of punishment. Since this paper is necessarily short, treatment of these
topics is likewise abbreviated, although hopefully not so abbreviated but
that it whets the appetite for those who wish to pursue them in greater
depth.1
I. What is Retributivism?
Retributivism is the view that we ought to punish offenders because
and only because they deserve to be punished. Punishment is justified,
for a retributivist, solely by the fact that those receiving it deserve it.
Punishment of deserving offenders may produce beneficial consequences
other than giving offenders their just deserts. Punishment may deter
future crime, incapacitate dangerous persons, educate citizens in the
behavior required for a civilized society, reinforce social cohesion, pre-
vent vigilante behavior, make victims of crime feel better, or satisfy the
vengeful desires of citizens who are not themselves crime victims. Yet
for a retributivist these are a happy surplus that punishment produces
and form no part of what makes punishment just; for a retributivist,
* Leon Meltzer Professor of Law and Philosophy, University of Pennsylvania. An
earlier version of this paper was given to the Fulbright Colloquium on Penal Theory,
held at the Philosophy Department, Stirling University, Stirling, Scotland in Sep-
tember, 1992. My thanks go to the participants at the Colloquium for their helpful
comments, and particularly in that regard to my commentator, Dudley Knowles of
the Glasgow University Philosophy Department.
1 My own views on these topics have been spelled out in greater detail in: 'Closet
Retributivism" (Spring-Summer 1982) USC Cites 9-16; Law and Psychiatry: Re.
thinking the Relationship (Cambridge U. P., 1984) 233-243; "The Moral Worth of
Retribution", in F. Shoeman, ed., Character, Responsibility, and the Emotions (Cam-
bridge U. P., 1987) 179-219, reprinted in Joel Feinberg and Hyman Gross, eds.,
Philosophy of Law (Wadsworth, 4th ed., 1991) 685-707; Placing Blame: A Theory of
the Criminal Law (Oxford U. P., forthcoming), chaps. 2 and 3.
ISRAEL LAW REVIEW
deserving offenders should be punished even if the punishment pro-
duces none of these other, surplus good effects.
A number of ambiguities are present in retributivism that should be
clarified before we move on to its justification. One of these has to do
with the connection of punishment to the desert of the offender.
Retributivists assert that punishment is justified only if those who
receive it deserve it,
2
but that is not the connection between desert and
punishment that is distinctive of retributivism. Rather, the more dis-
tinctive assertion of the retributivist is that punishment is justified if
it is given to those who deserve it. Desert, in other words, is a sufficient
condition of a just punishment, not only a necessary condition.
The second ambiguity has to do with the moral demands that just
punishment makes on state officials and citizens. The desert of offend-
ers certainly gives such officials permission to punish offenders, and this
permission to give just punishment is no mere Hohfeldian privilege but
a (claim) right to punish. But retributivism goes further. As a theory
of a kind of justice, it obligates us to seek retribution through the
punishment of the guilty. This means that officials have a duty to
punish deserving offenders and that citizens have a duty to set up and
support institutions that achieve such punishments.
3
The third ambiguity has to do with punishment itself. Does
retributivism purport to justify why we have institutions of punishment
generally, or only why we punish particular people on particular occa-
2 The late John Mackie called this view "negative retributivism*: J. L. Mackie,
"Morality and the Retributive Emotions', (Winter/Spring, 1982) Criminal Justice
Ethics 3-10; Mackie, "Retributivism: A Test Case for Ethical Objectivity,, in Joel
Feinberg and Hyman Gross, eds., Philosophy of Law (Wadsworth, 4th ed., 1991) 677-
684. Note that the negative retributivist, if he is truly a retributivist at all, must
also be a positive retributivist. For the negative retributivist asserts that the
justification for not punishing those who do not deserve it lies in the fact that there
is no morally acceptable reason to punish in the absence of desert; this implicit appeal
to the deserving of retributive justice is distinct from justifying the non-punishment
of the innocent on grounds of fairness, liberty, or utility. (On the latter point, see
Moore, Act and Crime (Clarendon Law Series, Oxford U. P., 1993) chap. 3.
3 See Moore, 'The Moral Worth of Retribution", supra n. 1, at 182. David Dolinko has
come to label this a "bold" retributivism, to be contrasted with a "weak" retributivism
according to which it is only permissible to punish those who deserve it. See Dolinko,
"Some Thoughts About Retributivism", (1991) 101 Ethics 537-559, at 542. The late
John Mackie called this pair "positive" versus "permissive" retributivism, respec-
tively: Mackie, 'Retributivism: A Test Case for Ethical Objectivity', supra n. 2, at
678; Mackie, "Morality and the Retributive Emotions", supra n. 2, at 4.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
sions? Some theorists have wished to distinguish the general justifying
aim of the criminal law from the justifications of particular decisions
within an ongoing criminal law system.
4
While such a distinction
certainly has its uses, it is of no use in wedding retributivist to, say,
utilitarian concerns with justifying punishment. Even if one were to
concede the existence of "practice rules" - rules which require moral
justification but whose application to particular cases is to be made
without resort to that moral justification
5
- such concession goes no
distance towards introducing any alien concerns into either a retribu-
tive or a utilitarian punishment scheme. That is, if one is a utilitarian
in his justification of the practice of punishment, one will remain purely
a utilitarian even if one bars utility calculations in applying criminal
law rules; and mutatis mutandis for the retributivist. Retributivism is
thus the view both that punishment institutions in general are justified
by the giving of just deserts and that the punishment of each offender
is justified by the fact that he or she deserves it.
The fourth ambiguity has to do with the "consequentialist" versus
"deontological" nature of retributivism. A consequentialist about moral-
ity believes that the rightness of an action is exclusively a function of
the goodness of the consequences that that action produces; a deontologist
about morality believes that the rightness of an action is (sometimes at
least) a function of the action's conformity with "agent-relative" norms,
norms that are addressed to each person individually at a particular
time and that are not concerned with maximizing conformity to such
norms by oneself or others on other occasions.'
The distinction becomes clearer if we apply it to retributivism. Take
Kant's old example of the last murderer in an island society that is about
to disband and leave its island.' A retributivist believes that the
murderer should be punished because he deserves it, even though no
other good will thereby be achieved. The guilty, in other words, must
4 Most notably, H. L. A. Hart, Punishment and Responsibility (Oxford U. P., 1968), and
John Rawls, "Two Concepts of Rules", (1956) 64 Philosophical Review 3-32.
5 This Rawls/Hart distinction has been carried into more recent discussions by Joseph
Raz. See his Practical Reason and Norms (Oxford U. P., 1975). I argue against Raz's
version of practice rules in Moore, "Law, Authority, and Razian Reasons', (1989) 62
S. Cal. L. R. 827-896, and Moore, "Three Concepts of Rules', (1991) 14 Harv. J. L.
and Public Policy 771-795.
6 This distinction is explored in much greater detail in Moore, 'Torture and the
Balance of Evils", (1989) 23 Is. L. R. 280-344.
7 Kant, The Metaphysical Elements of Justice (J. Ladd, trans., 1965) 102.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
be punished. If one construes this retributivist norm as an agent-
relative norm, then the members of Kant's island society are obligated
to punish the last murderer even if by doing so other guilty murderers
elsewhere will go unpunished. (We might imagine, for example, that if
we punish this murderer, he will not testify against other murderers
who will thereby escape conviction and punishment.) Alternatively, one
may look to the overall consequences of applying the norm; if one
construes this retributivist norm as a norm describing a state of affairs
that is to be maximized, then the members of Kant's island society
should not punish the last murderer but should rather maximize the
punishment of the guilty by their actions. The "deontological" or "agent-
relative" retributivist regards the act of punishing the guilty as categori-
cally demanded on each occasion, considered separately; the
"consequentialist"
or "agent-neutral"
retributivist regards the state of
the guilty receiving punishment as a good state to be maximized even
when this means that some guilty persons are intentionally allowed to
escape punishment.
It is usually assumed that retributivism must be of the agent-relative
kind.
8
Often this assumption is made about retributivism, the better to
criticize it. Thus, it might be argued that retributivism cannot justify
any actual institution of punishment because any actual institution of
punishment will make some mistakes in its findings of guilt, namely,
some actually innocent persons will be found guilty.
9
How, then, could
an agent-relative retributivist justify setting up retributivist institu-
tions of punishment? Not by weighing the evil of a guilty many not being
punished (if there were no institutions of punishment) against the evil
of a few innocents being punished (as will be the case if there are
institutions of punishment) - for this mode of justification would be
consequentialist, a mode not allowed to the (supposedly necessarily)
agent-relative retributivist.
Alternatively, it might be thought that we rightly refuse to punish
some guilty persons in order to be able to punish other, more seriously
guilty persons - as when we give immunity, in order to extract testi-
mony needed to convict the latter. 10 How can the retributivist accom-
8 See, e.g., Dolinko, "Three Mistakes of Retributivism", (1992) 39 UCLA L. R. 1623-
57.
9 See Dolinko, ibid; Schedler, "Can Retributivists Support Legal Punishment?* (1980)
63 Monist 185-198. Compare Larry Alexander, "Retributivism and the Inadvertent
Punishment of the Innocent' (1983) 2 Law and Philosophy 233-246.
10 I owe this point to Dudley Knowles.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
modate these practices, given that the retributivist regards the punish-
ment of the guilty to be categorically imperative whenever the opportu-
nity to give such punishment presents itself?
Yet if a retributivist may be a consequentialist, these are obviously
non-problems for his theory of punishment. With regard to the first, a
consequentialist-retributivist need not maximize the punishment of the
guilty to the exclusion of maximizing other, equally valuable states of
affairs, such as the non-punishment of the innocent. Such a retributivist
may-easily maximize both that those deserving punishment receive it
and that those not deserving of punishment not receive it. Where in the
actual design of punishment institutions the consequentialist-
retributivist comes out on this balance need not here detain us;
1
the
important point is that there is no incoherence, even prima facie, in such
a retributivist setting up an institution that will punish some innocents
in order to punish the guilty and forego punishing some guilty in order
not to punish too many innocents.
Likewise with regard to the second problem, the consequentialist-
retributivist will intentionally refuse to punish guilty persons whenever
more guilty persons (or greater guilt) will be punished thereby. For the
consequentialist-retributivist, no matter how intrinsically good it is that
the guilty receive their deserts, more of that good is to be preferred to
less of it.
Can the retributivist be a consequentialist? One can, of course,
define retributivism so that a retributivist must be a deontologist. Yet
I suspect that the temptation to so define retributivism stems from a
simple confusion. The confusion is between the intrinsic goodness of
retribution being exacted, on the one hand, and the categorical duty to
punish the guilty on each occasion where that can be done, on the other.
As I shall argue shortly in the next section, what is distinctively
retributivist is the view that the guilty receiving their just deserts is an
intrinsic good. It is, in other words, not an instrumental good - good
because such punishment causes other states of affairs to exist that are
11 The retributivist might adopt a principle of symmetry here - the guilty going
unpunished is exactly the same magnitude of evil as the innocent being punished -
and design his institutions accordingly. Or the retributivist might share the common
view (that the second is a greater evil than the first) and design punishment
institutions so that "ten guilty persons go unpunished in order that one innocent not
be punished". See Jeffrey Reiman and Ernest van den Haag, 'On the Common
Saying that it is Better that Ten Guilty Persons Escape than that One Innocent
Suffer: Pro and Con' (1990) 7 Social Philosophy and Policy 226-248.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
good. Even if punishing the guilty were without any further effects, it
would be a good state to seek to bring about, on this intrinsic goodness
view of punishing the guilty. This distinction - between intrinsic
versus instrumental goodness - is commonly confused with the distinc-
tion between deontological and consequentialist views of right action.
The result is that something that is not distinctive of retributivism (that
we each have an agent-relative obligation to see that the guilty are
punished) is taken to be essential to retributivism.
Although I thus conclude that a retributivist may be a consequentialist,
neither of the two problems above raised for deontological versions of
retributivism should incline the retributivist towards the consequentialist
version. The probable punishment of the innocent by any real-world
punishment scheme is not much of a worry even for deontological
versions of retributivism. We rightly set up many social institutions
where we know that some percentage of individuals affected by them
will be hurt or even killed, e.g., coal-mining, high-rise construction, and
speed limits on freeways. That we know that some percentage of
individuals will likely suffer these harms if we arrange these institu-
tions as we do arrange them is not to be equated with either our
intending that they be so harmed, or knowing that some identified
individual will suffer that harm. Agent-relative moral norms bind us
absolutely only with respect to evils we either intend or (on some
versions) knowingly visit on specified individuals. One can thus arrange
safeguards in coal mines, high rise construction, automobile travel, and
punishment in ways that predictably will hurt some who do not deserve
to be hurt, without for a moment ceasing to be an agent-relative theorist
about morality. 12
The same cannot be said about the intentional foregoing of any
opportunity to punish a guilty offender in order to obtain the conviction
and punishment of an even more guilty offender, which is why this
common prosecutorial practice is more of a problem for the deontological
version of retributivism. Yet the deontological retributivist might sim-
ply deny the propriety of the practice. More plausibly, if he is a
"threshold deontologist", as am I, he might more qualifiedly disavow the
12 I explore both the doctrine of double effect, and the analogous doctrine distinguishing
knowing visitation of harms on specified people versus risking harm to some uniden-
tified part of a class of persons, in Moore, supra n. 6. As I there show In detail, any
deontological view of morality must incorporate some such limits as these in order
to specify the domain over which morality's absolute prohibitions range.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIMSM
practice except when it is needed to punish some very deserving
criminal(s).
13
In any case, nothing in the present project requires that we decide
between deontological versus consequentialist retributivism. Both are
recognizable versions of retributivism. The most important point here
is not to confuse the distinction between these two versions of
retributivism with the quite different distinction between valuing the
giving of just deserts as only instrumentally good, as opposed to valuing
it as intrinsically good.
II. Justifying Retributivism
A. Morally Justifying The Retributive Principle by Particular
Judgments
How is retributivism, so defined, to be justified? Very generally,
there are three sorts of things which might be said to justify the
retributivist principle of punishment. One is a conceptual claim: only
when harsh treatment is imposed on offenders in order to give them
their just deserts does such harsh treatment constitute punishment.
1 4
The conceptual claim is made about our concept of punishment, and the
claim presupposes that that concept is captured by the purpose for
which punishment is imposed as much as by any structural feature that
it may possess.
A second sort of justification of retributivism is a legal claim: within
our legal system with its established doctrines and procedures, the
implicit goal is the retributive one of seeking to give just deserts to
offenders. The claim is an "interpretive" (or as I prefer to say, a
"functional") one, in that it accepts established legal institutions at face
13 A "threshold" deontologist refuses to violate a categorical norm of morality until not
doing so produces sufficient bad consequences as to pass some threshold - then, he
will override such categorical norms. See Moore, supra n. 6.
14 For this older form of argument see, e.g., A. M. Quinton, 'On Punishmente (1954) 14
Analysis 1933-42. For a related argument, seeking to tease retributivism out of our
concept of law, see Herbert Fingarette, "Punishment and Suffering' (1977) 50 Pro-
ceedings of the American Philosophical Association 499-525.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
value and asks, "what goals make the most sense of these"?
5
The claim
is that the retributive goal fits our established doctrines well, whereas
competing utilitarian goals like crime-prevention fit so poorly that they
seem like so many of F. H. Bradley's bad reasons for what we believe
on instinct anyway.
A third sort of justification of retributivism is a moral claim: our
obligation to punish offenders so as to give them their just deserts is
justified in this sense if the practice meets whatever epistemic stan-
dards we impose to justify any of our moral beliefs. If we are
consequentialists about morality, and if there are other goods to be
maximized besides giving the guilty their due, then we may seek to
justify this retributive value by its contribution to the maximizing of
some other good(s) that we accept more readily than we accept the good
of retributive punishment. If we are not consequentialists about moral-
ity, or if our consequentialism admits of no goods that are more basic
than the good of achieving retributive justice, then our mode of moral
justification cannot be in terms of some other value served by retributive
punishment. Rather, we must justify retributivism in whatever way we
justify actions and practices as being intrinsically right. However we
justify the intrinsic rightness of not punishing the innocent, for ex-
ample, is how we should justify the intrinsic rightness of punishing the
guilty, on this general view of morality.
The most important sort of justification is the third one, the moral
justification of retributive punishment, and so I shall focus on it alone.
In order to understand what we can rightfully ask for by way of moral
justification we need to keep three distinctions in mind. One of these
is the consequentialism/deontology distinction of the last section. The
second of these was also adverted to earlier, that between the intrinsic
versus the merely instrumental goodness of a state of affairs or of an act
or institution. A state of affairs, act or practice is intrinsically good if
its goodness does not depend on some further effect that the state of
affairs, etc., produces. One might think that giving a novelist some
reward for producing a novel is intrinsically good, it being no more than
she deserves. Alternatively, if the goodness of such a practice of reward-
15 Ronald Dworkin regards such claims as interpretive, although I prefer the older if
less fashionable label, "functional". For the difference, see Moore, "Law as a Func-
tional Kind", in Robert George, ed., Natural Law Theories (Oxford U. P., 1992) 201-
203, 220-221. For the functionalist claim that Anglo-American criminal law is a
functional kind whose implicit goal is the seeldng of retributive justice, see Moore,
"A Theory of Criminal Law Theories", (1990) 9 T. A. U. Studies in Law 115-185.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
ing novelists with copyright monopolies inheres in the effects of the
practice - by giving rewards one gives incentives to other novelists to
produce novels, and novels are an intrinsic good - then the practice is
only instrumentally good.
Notice that the notion of instrumental goodness is not to be identified
as utilitarian. Utilitarianism is a kind of consequentialist moral theory
with a particular idea of what is intrinsically good - pleasure, happi-
ness, preference-satisfaction, or some other welfarist notion. For a
utilitarian, all other goods are merely instrumental. There is nothing
in the intrinsic/instrumental distinction that requires this kind of mo-
nism about what is intrinsically good. One might well be a pluralist in
this regard, admitting that the welfare of persons is an intrinsic good
while also thinking that not punishing the innocent, giving those who
labor the fruits of that labor, not demanding the impossible from people,
etc., are equally intrinsic and not merely instrumental goods.
As I mentioned in the last section, this intrinsic/instrumental distinc-
tion is often mistaken for the deontological/consequentialist distinction.
Yet these are quite distinct oppositions. One might believe that giving
those who labor the fruits of their labor is intrinsically good. This does
not determine whether one is a consequentialist, so that the right
institution (of copyright, say) is one that maximizes this intrinsic good
by sacrificing its attainment in the short run for even more of it in the
long run; or whether one is a deontologist, so that the practice of giving
the novelist the fruits of her labors is the intrinsically right thing to do
on any occasion on which one can do it.
The intrinsic/instrumental distinction is also often confused with the
third distinction pertinent here, the distinction between first principles
in ethics, and all other principles (let us call them "secondary prin-
ciples"). A first principle, as the name suggests, is the most general
description of what it is intrinsically right to do or what states of affairs
are intrinsically good to cause. Being first, such a principle cannot be
deduced from some more general principle. This feature of first prin-
ciples may suggest that such principles cannot be argued for at all and
that they must therefore be defended as "self-evident" or revealed by
religious faith.
My own view of justification in ethics rejects these last connotations
of the idea of first principles. On what is often termed a
"nonfoundationalist"
or "coherence" view of ethical justification, to say
that a principle is the most general that can be formulated - and
therefore that it cannot be argued for by showing it to be an instance
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
of a yet more general principle - is not to say that the principle cannot
be argued for at all. Rather, first principles are to be justified as
abductive inferences from more particular principles and judgments. 16
When John Stuart Mill argued for his first principle (of utility), for
example, he saw that that principle could not be deduced from some yet
more general principle - else the principle of utility would then not be
a first principle.1
7
Yet (on one interpretation of Mill) he did argue for
the principle of utility on the grounds that it best described what was
common to all more particular judgments about what states of affairs
are desirable, namely, that they are desired. Similarly, when John
Rawls argued for his two first principles of distributive justice, he did
not seek to deduce them from some more general principle of justice as
such. Rather, he sought to show how his two principles best described
a range of particular judgments, both substantive (no one deserves his
own talents) and procedural (hypothetical agreement in a fairly con-
structed setting for choice). is
It is a kind of category mistake to equate the first principle/second
principle distinction with the distinction between intrinsic goodness/
instrumental goodness. The latter is a metaphysical distinction be-
tween two sorts of value. The former is an epistemological distinction
between two sorts of principles through the grasping of which values are
known by us. Moreover, there is no tight linkage between the two
distinctions. While a first principle presumably describes what is
intrinsically good or intrinsically right, secondary principles may do so
as well. I may, for example, believe that giving this novelist this sort
of copyright protection is intrinsically right, believe that giving novelists
copyright protection is intrinsically right, believe that giving property
rights to anyone who creates a novel thing is intrinsically right, and
believe that giving anyone what they deserve is intrinsically right, even
though only the most general of these beliefs should be construed to
have as its content a first principle.
This minor detour into some fundamental distinctions in ethics is
essential if we are to keep proper perspective on various criticisms of my
mode ofjustifying retributivism. For what if one were to hold that: (1)
the retributive principle is a first principle; (2) the suffering that pun-
16 See Moore, "Moral Reality", [19821 Wisc. L. R. 1061-1156; Moore, "Moral Reality
Revisited" (1992) 90 Mich. L. R. 2424-2533.
17 John Stuart Mill, Utilitarianism (Liberal Arts Press, 1957) 44.
18 John Rawls, A Theory of Justice (Harvard U. P., 1971).
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
ishment entails is an intrinsic good when inflicted on those who deserve
it; and (3) we each have an agent-relative obligation to punish the guilty
(even if other guilty persons would escape punishment) because the
intrinsic goodness of such punishment is not to be maximized by our
actions? If retributivism is taken to be a first principle describing an
intrinsic value that we each categorically are enjoined to realize in our
actions, then we must tailor our demands for its moral justification
accordingly.
My own mode of justifying retributivism has tried to do this. 1
9
I take
seriously the sorts of particular moral judgments that Kant-like thought
experiments call forth in me and in most people I know: imagine an
offender who does a serious wrong in a very culpable way - e.g.,
Dostoevsky's Russian nobleman in The Brothers Karamazov, who turns
loose his dogs to tear apart a young boy before the eyes of the boy's
mother; imagine further that circumstances are such (e.g., Kant's island
society about to disband) that no non-retributive purpose would be
served by punishing this offender. Now imagine two variations: (1) you
are that offender; (2) someone else is that offender. Question: should
you or the other offender be punished, even though no other social good
will thereby be achieved? The retributivist's "yes" runs deep for most
people.
The main reason that thoughtful people often give for disavowing
their own retributivist intuitions about such cases is that they think
such impulses to be unworthy of them. More specifically, they under-
stand that such judgments are the products of emotional responses by
them to the brutal facts of such cases, and they wish to disavow those
emotions. That is why I have found it important to separate the first
and third person variations of the thought experiment. Such variations
call forth quite different emotions, even if they lead to a common
judgment of deserved punishment.
The third person variation of the thought experiment is admittedly
quite troublesome in terms of the sort of emotions it calls forth. Nietzsche
observed that such thought experiments called forth that witches' brew
of emotions he lumped under the French term, ressentiment: emotions
of resentment, fear, revenge, sadism, and cruelty. Because such emo-
tions are unworthy of us, Nietzsche urged, we should disavow the
retributive judgments to which they lead.2
19 Moore, 'The Moral Worth of Retribution, supra n. 1.
20 For discussion and citations, see Moore, 'The Moral Worth of Retribution', ibid
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
More recently a number of legal philosophers have sought to rescue
these kinds of retributive judgments from Nietzsche's sort of disavowal.
Jeffrie Murphy, Jean Hampton, and Samuel Pillsbury have each sought
to distinguish a virtuous emotional response from responses of
ressentiment to third person thought experiments.
1
In different ways
they have each sought to distinguish a "moral hatred" or a "moral
outrage" that is the only proper response of a moral being to the gross
violations of morality by others that violent crime represents. To my
mind, they have made their case: to be indifferent to culpable wrongdo-
ing by another evinces both a lack of empathetic identification with
others who are victims, and a lack of attachment to morality. The
virtuous person may turn the other cheek when it is his cheek that has
been slapped (although I doubt even this), but there is no virtue in
turning someone else's cheek when they've been slapped. Violations of
others' moral rights should make us angry at those who flout that
morality, and that anger need not be tainted by cruel, sadistic, fearful,
or resentful emotional accompaniments.
My own rescue of retributivism from Nietzsche's charge has been
different. I have focused on the first person variation of the Kantian
thought experiment, where we are to imagine that it is we who have
culpably done some great wrong. Here there is a much lessened danger
that our intuitions about desert will be tainted by the emotions of
ressentiment. Rather, one emotion here predominates, and that is the
emotion of guilt. A virtuous person would feel great guilt at violating
another's rights by killing, raping, assaulting, etc. And when that
emotion of guilt produces the judgment that one deserves to suffer
because one has culpably done wrong, that judgment is not suspect
because of its emotional origins in the way that the corresponding third
person judgment might be.
It may be tempting to concede that in one's own case: (1) one would
indeed feel very guilty if one culpably did such a wrong; (2) that such
feelings of guilt would be virtuous to feel - indeed, the only tolerable
response of a moral person; (3) that therefore the judgment which such
emotion produces - that one is guilty and deserving of punishment -
is true. Yet, one might refuse: (4) to generalize this judgment about
one's own deserved punishment to others who do like wrongs. One
21 Jean Hampton and Jeffrie Murphy, Forgiveness and Mercy (Cambridge U. P., 1988);
Samuel Pillsbury, "Emotional Justice: Moralizing the Passions of Criminal Punish-
ment" (1989) 74 Cornell L. R. 655-710.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
might think that the standard that we apply to ourselves should not be
applied to others.
This sounds like a generous way to think. It is important to see that
it is not. To grant that you would be guilty and deserving of punishment,
but that others who do the exact same wrong with the exact same
culpability would not be, is to arrogate to yourself a god-like position.
Only you have those attributes of moral agency making you alone a
creature capable of being morally guilty; others are simply lesser beings,
to be pitied perhaps, but not to be blamed as you would blame yourself.
This is not moral generosity. Rather, it is an elitist arrogance that
denies one's common humanity with those who do wrong. Such elitist
condescension is no virtue, and provides no basis for refusing to endorse
the fourth step in the first person version of the Kantian thought
experiment, a step generalizing one's own potential for guilt and de-
served punishment to other persons.
B. Critical Reconsideration of this Mode of Moral Justification
A variety of criticisms have been levelled at my justification of
retributive punishment via particular thought experiments and the
kind of particular judgments that they call forth. The leading criticisms
seem to be five in number, which I shall deal with seriatim.
1. First is the charge of circularity. The charge is that the general
judgment - that punishment should be imposed just because culpable
wrongdoers deserve it - is being justified by nothing but itself. For
what are the judgments in particular thought experiments but judg-
ments that punishment should be imposed (e.g., on Kant's last murderer
in the island society, on Dostoevsky's cruel nobleman, etc.) just because
each of these culpable wrongdoers deserve to be punished? The worry
here is one of question-begging: since the particular judgments with
which we start are themselves retributive judgments, how can they
justify the general retributive judgment without begging the question?
There are three ways of construing this question-begging (or circu-
larity) worry. One is to construe it strictly: literally, the very same
principle is being used to justify itself. So construed, the worry is
groundless, for the particularjudgments called forth by Kant-like thought
experiments are not identical in content with the retributive principle
itself. The retributive principle is the universalized judgment that all
culpable wrong-doing ought to be punished; Kantian thought experi-
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
ments give rise to particular judgments that particular persons in
particular situations ought to be punished. The former judgment stands
to the latter judgments as universally generalized to singular state-
ments: aspects of the situations involved in the latter judgments are
abstracted as relevant, and these are universalized into the general
judgment that whenever these aspects are present punishment is obliga-
tory. Strictly speaking, thus, the retributivist is not justifying the
retributive norm by reference solely to itself. Rather, he is inferring a
universally generalized statement from a set of singular statements.
A second, less-literal construal of the question-begging worry does
not claim identity between the judgment to be justified, retributivism,
and the particular judgments doing the justifying. Recognizing the
distinction between universally generalized versus singular judgments,
this form of the worry nonetheless insists that the particular judgments
are themselves retributive. The worry is that the particular judgments
are of the form "the nobleman should be punished because and only
because he deserves it". The worry is that merely universalizing this
singular judgment into the general judgment - that all those who
deserve punishment should be given the punishment that they deserve
- is trivial.
22
22 David Dolinko advances this form of the question-begging objection against my
coherentist methodology with an example: "Consider, as an analogy, a similar
coherence justification for the view, say, that women are inferior to men, offered circa
1800 by one man to another. Someone has raised the question of whether it is really
justifiable to treat women as inferior and is met with the response that, 'Well, of
course, we all have the intuition that in this situation we should discriminate against
women, and likewise in this and this; now, the best way of accounting for these
particular judgments is to suppose women to be morally and intellectually inferior
to men'. Wouldnt one feel that the crucial question is being begged?* Dolinko, supra
n. 3, at 557-58. Notice that Dolinko trivializes the inference in his analogy in the
manner attacked in the text: the particular judgments are not that this person
should be treated in some inferior way in this context; rather, the particular judg-
ments already include most of the terms of the principle to be abstracted from them
- that the inferior treatment is discrimination; that the only feature to be mentioned
is that the persons discriminated against are women.
By his choice of analogy, Dolinko also of course hopes to suggest that a coherence
mode of justification is no guarantee against moral error, for we are only cohering
our own beliefs and attitudes. One can easily concede this possibility of error without
for a moment conceding Dolinko's hoped for conclusion (that this form of argument
is question-begging). Our particular judgments of visual perception can of course also
be in error on occasion, but that does not eviscerate the power of those judgments
to be good (if not conclusive) reasons for believing the principles abstractable from
them.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIISM
This form of the circularity worry would be a real worry if the
presupposition on which it rests were true. What the worry presupposes
is that the particular judgments from which justification proceeds are
of the content just described. Yet such particular judgments are not that
"the nobleman should be punished because and only because he de-
serves it". Rather, the particular judgments are that the nobleman
should be punished, period. It is making such judgments about a wide
range of particular cases, and abstracting from those cases what they
have in common - the desert of the actors - that then allows the
retributivist to infer the retributivist principle that offenders should be
punished when, but only when, they deserve it. This is not a trivial
inference, for the features of the situations in which we judge punish-
ment to be appropriate could easily be otherwise. It could be the
reformability, or the deterrability, either of the particular offender or of
some class that he instantiates, that is the recurring feature present in
the situations that call forth our judgments about the appropriateness
of punishment in particular cases. As it happens, these are not the
features of relevant situations that recur. The inference that it is the
desert of the offenders that should be generalized from our particular
judgments of punishment is thus not a trivial one. Such particular
judgments thus possess a non-trivial justificatory force for sustaining
the retributive principle.
The third and last form of the circularity worry focuses on desert
itself. The worry here is that "desert" simply means, "a fit subject for
retributive punishment", so that when one abstracts desert from par-
ticular cases (as the feature of those cases that caused judgments of
punishment) one has done no more than start with particular retribu-
tive judgments in order to justify the general retributive judgment.
23
"Desert", however, does not mean, "a fit subject for retributive punish-
ment", any more than "death" means, "a fit subject for concern by a
mortician" or "illness" means, "a fit subject for concern by a medical
doctor".
24
Rather, "desert" means culpable wrongdoing (at least with
respect to blameworthy actions - a broader formulation is needed to
23 For an example of this form of the question-begging worry, see Mackie, "Retributivism:
A Test Case for Ethical Objectivity", supra n. 2, at 679 (Desert' is not such a further
explanation, but is just the general, as yet unexplained, notion of ... retributivism
itself".)
24 These last examples are considered in Moore, Law and Psychiatry, supra n. 1, at
chap. 5.
Nos. 1-2., 1993]
ISRAEL LAW REVIEW
encompass praiseworthy actions). When we abstract desert out as the
recurring feature present in cases of prima facie justified punishment,
we are thus not merely isolating "punishability on retributive grounds".
There are two aspects of desert that are worth drawing out a bit. The
first is wrongdoing. A person does wrong whenever his voluntary act
causes a state of affairs to exist that instantiates a moral norm that
prima facie prohibits such acting and such causing, and there is no
moral justification for this prima facie wrong. Wrongdoing is thus
constructed out of the elements of action, causation, and (lack of) jus-
tification.
=
Culpability is distinct from wrongdoing. A person may do wrong, say
by killing a perfectly innocent victim, yet be non-culpable in the doing
of that wrongful act. He may be insane, for example, or acting under
extreme duress, or he may have mistakenly believed that his victim was
trying to kill him. Likewise, a person may be very culpable and yet have
done no wrong. He may have shot a stump, for example, while believing
that he was shooting his old enemy. Culpability arises from the mental
state with which an action may be done, and the (lack of) excuse for the
otherwise culpable doing of a wrongful action.
To say that a person deserves punishment is to say that he has
culpably done wrong (or, in some cases, that he is culpable even without
wrongdoing). Desert is thus far from an empty concept, connoting no
more than punishability on retributive grounds. Inasmuch as wrong-
doing and culpability are far from empty concepts, neither is desert.
2. I suspect that some critics of my justificatory methods might well
concede the above rebuttals of the three question-begging worries and
yet still think that generalizing from particular judgments to the re-
tributive principle cannot justify the latter principle." To sustain this
form of the objection, however, such a critic would have to believe two
things: first, that there is no more general principle from which the
retributive principle can be deduced; and second, that absent such a
more general principle, the retributive principle is in no sense morally
justified.
We can quickly dispatch with the second of these beliefs, for such a
belief is no more than a return to a rationalist foundationalism in moral
25 On wrongdoing, see Moore, Act and Crime, supra n. 2, chap. 7.
26 In his commentary upon this paper at Stirling my commentator, Dudley Knowles,
appeared to take this view before eventually abandoning it. See Knowles, "Unjusti-
fied Retribution', in this issue, p. 50.
[Is.L.l. Vol. 2.7
JUSTIFYING RETRIBUTIVISM
epistemology. Such rationalism leaves wholly unexplained how first
principles are to be justified, since they cannot be deductively justified
as following from some even more basic, "more first" principles. Al-
though some may find "self-evidence", "God's word", or "transcendental
deductions" to be acceptable starting points in justifying moral belief,
I do not. First principles can be justified only by showing that they
imply, but are not implied by, both secondary principles and particular
judgments that we accept as true. To deny this as a form ofjustification
would be to deny the possibility of justifying any first principle, and to
deny that would be to deny the justifiability of any judgment that is
itself justified by such first principles (which, on the rationalist view of
justification here considered, are all moral judgments).
The history of rationalist justification schemes in ethics should not
make one sanguine about finding indubitable first principles. The
absence of such self-evident first principles is alone enough to dispatch
the objection here considered to my mode of justifying the retributive
principle. If retributivism is a first principle of morals, then it neither
needs nor can have any other form of justification but the form I have
offered. Still, the retributive principle seems to be part and parcel of
yet more general principles, so it is worth pausing briefly to examine
this alternative possibility.
The battleground of theory known as the philosophy of punishment
is littered with the corpses of supposed general principles from which
the retributive principle is supposed to follow. Giving an offender the
punishment he deserves solely because he deserves it has been said to
follow from a principle that: debts to society must be repaid (coupled
with the further idea that crime creates a debt and punishment is a form
of repayment); wrongs must be annulled (coupled with the further idea
that punishment annuls them); God's anger must be placated (coupled
with the further thoughts that God is a retributivist and that human
punishment placates her); wrongdoing must be denounced (coupled
with the further belief that punishment is the appropriate form of
denunciation despite there being other, less draconian forms); etc.
27
David Dolinko correctly observes that in justifying retributivism I have
taken the path of "avoiding the common retributivist metaphors alto-
gether rather than seeking to unpack and develop them".
28
I have taken
27 For a more complete taxonomy, see J.G. Cottingham, "Varieties of Retribution"
(1979) 29 Philosophical Q. 238-46.
28 Dolinko, supra n. 3, at 555.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
this path because I see very little that is correct (or in some cases even
intelligible)
in these views.
2 9
One more general principle that does not seem clouded in metaphor
is a general principle of desert.
30
Such a principle arises because of what
might be called the "secondary" moral rights and duties we all possess.
I have a primary duty not to break (most of) my promises and another
primary duty not to injure or kill (most of) my fellow persons; I have a
primary right to keep (most of) the fruits of my own labors. I also have
what I shall call secondary duties and rights - respectively, a duty
either to perform my promise, even belatedly, or in some other way to
put the promisee in as good a position as he would have been in had I
kept my promise; a duty to correct the injustice that I have caused in
injuring or killing another by making amends in whatever way I can,
including compensation; and a right to have any who deprive me of the
fruits of my own labors either to give those things back (or pay me their
value where they cannot be given back). It is breach of these secondary
duties that warrants the judgment that I ought to be made to keep my
promise or pay its equivalent, or that I ought to be made to compensate
the victims of my violence; it is violation of the duty correlative of my
secondary rights that warrants the judgment that others ought to be
made to give me back my property or its equivalent. We idiomatically
make this "ought" judgment using the word "desert": the promise-
breaker deserves contract sanctions against him, the tort-feasor de-
serves tort law sanctions against him, and the laborer creating a thing
deserves to have those incidents of ownership in that thing that are
distinctive of property rights.
The unity of a principle of desert covering all of these situations does
not lie in the species of legal coercion to be justified, although there is
in fact some overlap in contract, tort, and property remedies. Nor does
the unity of the principle lie in the intrinsic goodness of giving people
what they deserve - for then the term "desert" would mean no more
29 An exception may be Herbert Morris' well-known principle of fairness, which may
well survive the by-now voluminous criticisms made of it. See Morris, "Persons and
Punishmente (1968) 52 The Monist 475-501.
30 This is to take up the suggestion I offered in "The Moral Worth of Retribution", supra
n. 1, at 182, 186.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
than, "intrinsically right".
31
Rather, the unity of the principle of desert
is to be found in the common conditions of culpability and of wrongdoing
(or in their praiseworthy analogues, of "meritability" and of "right
doing"). I refer to the elements of voluntariness, causation, and justi-
fication, and of intention and excuse, out of which wrongdoing and
culpability respectively are constituted. There is a general principle of
desert in that the conditions of wrongdoing and culpability are similar.
32
The retributive principle - that offenders should be punished be-
cause and only because they have culpably done wrong - is an instance
of this more general principle of desert. We all have primary duties not
to do the sort of acts that malum in se criminal statutes prohibit. We
also have secondary duties to allow ourselves to be made to suffer if we
have violated these primary duties. The trigger for these secondary
duties is again our culpability in violating the primary duties that define
wrongdoing.
Although it takes considerably more detail than this to sketch the
unitary principle of desert, my present interest is in a challenge pre-
sented to such a principle, however it may be filled out. I refer to the
objection that desert does not give a sufficient reason to punish. As put
by David Dolinko, who advances a form of this objection: "In general,
the sense of 'fittingness' or 'propriety' inherent in someone's getting
what she deserves is a very poor guide to whether it is permissible to
give her what she deserves".
3
Dolinko's main ground for reaching this conclusion (about the incon-
clusiveness of desert) lies in his ideas of rights. Dolinko correctly
observes that each of us has rights generally not to be made to suffer.
It is thus "entirely possible for a person to deserve treatment which it
would nevertheless violate his rights to be given", as Dolinko also
31 I am doubtlessly regimenting the usage of "desert* here a bit so as to give it this
descriptive content. Some of the ordinary usages of "desert" are such that it can mean
no more than "intrinsically right' or "intrinsically ought to be done". Such broad,
purely evaluative usages of "desert" account for the sense of Mackie and others that
"desert' is an empty concept. See n. 23, supra.
32 For an unpacking of the notions of wrongdoing and culpability in the criminal law
context, see Moore, "A Theory of Criminal Law Theories", supra n. 15. For a
comparison of some of the conditions of culpability in contract and criminal law, see
H. L. A. Hart, "Legal Responsibility and Excuses", in his Punishment and Respon-
sibility (Oxford, 1968).
33 Dolinko, supra n. 3, at 543-44, 558; Dolinko, supra n. 8, at 1627-30.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
observes.
34
One might, for example, believe that lex talionis is what each
offender deserves, so that the torturer deserves to be tortured; yet one
might well believe that the offender has a right not to be tortured
anyway. Dolinko's hoped-for conclusion is that therefore, desert is in
general a poor guide to answering the question of what we should do to
offenders.
3 5
Dolinko's general conclusion does not, of course, follow from his
correct observation that we all have rights not to be deprived of our
liberty, not to be tortured, not to be killed, etc. It would be a crude
caricature of the retributivist to make him monomaniacally focused on
the achievement of retributive justice. The retributivist like anyone else
can admit that there are other intrinsic goods, such as the goods pro-
tected by the rights to life, liberty, and bodily integrity. The retributivist
can also admit that sometimes some of these rights will trump the
achieving of retributive justice (as in my lex talionis example above).
Yet to assert that all such rights trump the achieving of retributive
justice is simply to beg the question against the retributivist, whose
essential thesis is that they do not. Particularly with regard to the
rights to life and to liberty, culpably doing wrong forfeits such rights so
as to achieve retributive justice, which one can accept without thinking
that all rights of offenders are so forfeited.
Dolinko's attempt to remedy this first defect of his analysis intro-
duces a second. Dolinko correctly quotes me for the retributivist point
that "'the moral desert of an offender is a sufficient reason to punish him
or her'."1
6
Because the retributivists must show that giving an offender
what she deserves does not violate any rights not forfeited by the
offender's culpable wrongdoing, Dolinko would admonish me and other
retributivists to "acknowledge that 'criminals deserve to suffer' is simply
not a sufficient explanation of why we are morally permitted to inflict
such suffering
on them".
37
Here Dolinko fails to attend to the context-sensitivity of any talk
about "sufficient conditions", my own included. "Sufficiency" is like
"qualitatively
identical" in that we almost never use such words or
phrases literally. What we always mean when we say of two distinct
particulars that they are qualitatively identical, is that within some
34 Dolinko, supra n. 8, at 1629.
35 Ibid.
36 Ibid., at 1628 (quoting Moore, 'The Moral Worth of Retributivism, supra n. 1, at
180).
37 Ibid., at 1630.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
limited set of properties that the context invariably supplies, the two
share "all" of their properties. Similarly with "sufficient": when we say
that one condition was sufficient for another - as in the counterfactual
statement that a particular fire was sufficient to burn down someone's
house - we mean that within some limited set of conditions this one
by itself was sufficient. Other conditions outside that set - such as the
presence of oxygen in my fire example - are invariably necessary even
while we idiomatically describe a condition within that set as "suffi-
cient".
The retributivist's usage of"sufficiency" is to be so construed. Within
the set of conditions constituting intelligible reasons to punish, the
retributivist asserts, desert is sufficient, i.e., no other of these conditions
is necessary. Of course other conditions outside the set of conditions
constituting intelligible reasons to punish may also be necessary to a
just punishment, such as the condition that the punishment not violate
any non-forfeited rights of an offender. That concession (implicit any-
way in any intelligible usage of "sufficient") in no way detracts from the
distinctive claim of the retributivist: culpable wrongdoing (i.e., desert)
is a sufficient reason to punish. More generally, such desert is a
sufficient reason to attach the contract, tort, and criminal law sanctions
to both promissory and non-promissory wrongs when culpably done, and
to attach property law protection to the meritable right-doing of one who
creates a thing of value by her own labor.
By failing to recognize the context-sensitivity of any use of "suffi-
ciency", Dolinko also fails to recognize that he gives away his objection
here when he admits that '"a person's desert of X is always a reason for
giving X to him, but not always a conclusive reason' because 'consider-
ations irrelevant to his desert can have overriding cogency in establish-
ing how he ought to be treated on balance'."
38
A retributivist can happily
agree with this statement of Joel Feinberg's that Dolinko adopts, and
not for a moment give up his claim that within the set of prima facie
reasons to punish, desert alone is sufficient.
Even without the rights argument, I have the sense that for Dolinko
there remains some "gap" between the retributivist's judgment, "X
deserves punishment", and the retributivist's conclusion, "therefore
punishing X is morally justified".
9
Yet Dolinko has made no showing
38 Dolinko, supra n. 3, at 544 (quoting Joel Feinberg, "Justice and Personal Desert", in
Doing and Deserving (Princeton, 1970), at 60).
39 Ibid., at 545.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
that there is any gap here. As we have seen, the general principle of
desert asserts that the culpable wrongdoing (desert) of X is a sufficient
reason to visit various legal sanctions on K As we have also seen, the
context-sensitive use of "sufficient" here allows for the possibility that
there are overriding reasons (such as non-forfeited rights of X) not to
visit these legal sanctions on X But this last possibility presents a gap
only when it becomes actualized. And what are the overriding reasons
not to make the culpable promise-breaker restore the promisee to the
status quo ante, to make the culpable tort-feasor compensate those
whom he has hurt, and to make the criminal suffer his deserved pun-
ishment? Certainly the general rights of all of us not to suffer such
sanctions provide us no such overriding reasons, since we forfeit such
rights precisely by the culpable wrongdoing that constitutes desert.
It might seem that this difference between Dolinko and myself comes
down to a burden of proof issue, Dolinko urging me to show that there
are no such overriding reasons and I urging Dolinko to show that there
are. If so, surely the burden of proof is on the critic of retributivism on
this issue. After all, once it is admitted that desert constitutes a prima
facie reason to punish, surely it is incumbent on those who urge us not
to punish to show us why we should stay our hand.
It might also seem that I have done little to justify the more general
principle of desert, even if I have shown how the retributive principle
follows from it. But of course, that objection can be raised ad infinitum
for any answer that might be given. That is the problem with thinking
that justification has to be a matter of deducing the principle to be
justified from one that is yet more general. We must give up that picture
of how justification must proceed. Once we do so, then whether we are
justified in believing in retributivism cannot depend on our finding some
yet more general principle with which to justify it.
3. A common objection to a coherentist justification of moral prin-
ciples goes like this: since the data being cohered into more general
principles are our own judgments and feelings, the principles that result
are only a coherent expression of our own sentiments. They can have
no claim to ethical objectivity given the nature of the data out of which
they are constructed. As Joel Feinberg puts it, in response to my
retributive thought experiments: "Is our reaction a reliable guide to
moral
judgment?"
40
40 Joel Feinberg and Hyman Gross, Philosophy of Law (Wadsworth, 4th ed., 1991) 630.
For similar queries about my thought experiments, Bee Sanford Kadish and Stephen
Schulhofer, Cases and Materials on Criminal Law (Little, Brown, 5th ed., 1989) 145.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
We should distinguish a naive from a sophisticated version of this
worry. In its naive form, the thought is that it is necessarily the case
that principles abstracted from our judgments and feelings in particular
cases must be about our sentiments, not about some objective features
of moral reality. This must be the case because the data from which the
principles are abstracted are our own subjective feelings and judgments.
This form of the objection is naive because it confuses the question
of whether our judgments and feelings in moral thought experiments
are subjective, with the question of whether what those judgments and
feelings are about is subjective. Of course, my judgments and feelings
are mine, part of my mental state, and are in that sense subjective. But
this goes no distance towards establishing that what my judgments and
feelings are about are also subjective. My perceptual belief that I am
now confronted by a bird on the branch beyond my window is a subjec-
tive mental state of mine; the bird and the branch, however, are not
mental states of mine but have their own independent existence.
Some beliefs, of course, are about other subjective mental states. I
can believe that I desire to go fishing, that I hate pretension, that I am
in pain, that I am having the visual experience of seeing a bird, etc. In
each case the objects of my mental state of belief is some further mental
state of mine. Yet in form our moral beliefs are not like this. When I
believe that bullfighting is wrong, the object of my belief is not, prima
facie, some further mental state of mine - it is about the institution of
bull-fighting and about a property that institution possesses, namely,
its wrongness.
41
One thus must sophisticate the objection to make it at all interesting.
In its sophisticated form the objection concedes that principles ab-
stracted from our judgments and feelings could be objectively true even
though the data from which the principles are abstracted are (harm-
lessly) our own more particular subjective judgments and feelings.
Indeed, the sophisticate here can also concede that both our particular
judgments and the principles abstracted from them in form purport to
be about independently existing moral qualities.'
2
The sophisticate
urges, however, that in fact such qualities do not exist because the best
41 For the classical argument concluding that our moral beliefs cannot be about our own
moral beliefs, see G. E. Moore, Ethics (1912), and for an update, M.S. Moore, "Moral
Reality', supra n. 16, at 1075-1079.
42 This concession is made by the late John Mackie in his Ethics: Inventing Right and
Wrong (Penguin, 1977), chap. 1.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
explanation why we have the reactions we do (to Kant-like thought
experiments) lies elsewhere.
The sophisticated form of this objection presupposes a certain view
about moral epistemology and metaphysics - a view that I share and
have called "explanationism";
43
a brief description of this view may thus
be helpful to understanding this form of the objection. The view is that
we are justified in trusting our moral intuitions in particular cases to
be evidence of objective moral facts if, but only if, the best explanation
why we have such intuitions includes those moral facts. The view
anticipates that for any natural phenomenon (like our reactions to Kant-
like thought experiments) there will be competing explanations. Each
explanation will presuppose that a somewhat different set of features
exists. If one explanation is demonstrably superior to its competitors,
then the natural phenomenon thus explained is good evidence for the
features of the world that best explain it.
My mode of justifying retributivism supposes that the best explana-
tion for why we have the reactions we do to Kant-like thought experi-
ments is that those reactions are caused by the existing moral qualities
of wrongness and of culpability, the combination of which I ball desert.
My sophisticated objector here flatly joins issue: the best explanation
for the reactions we have to such thought experiments does not lie in
the causal power of any objectively existing moral qualities of desert, but
rather, in certain psychological facts about us and certain sociological
facts about our society. If this latter explanation is best, then we should
not take our reactions to evidence the objective truth of retributivism,
but only certain psychological and social truth about us and our society.
Although Nigel Walker has briefly adverted to this explanationist
objection to my mode of justifying retributivism," the detailed working
43 Michael Moore, "Moral Reality Revisited", supra n. 16, at 2491-2501.
44 Nigel Walker, Why Punish? (Oxford, 1991), chap. 9. Walker focuses on a fact he
believes to be "awkward" for the retributivist, which is 'the fact that by no means
everyone perceives retribution as a duty" in response to my thought experiments.
Walker anticipates that I might explain this fact by reference to a defective upbring-
ing, but urges that I cannot avail myself of this explanation since it suggests that
the perception which I posit "is really a learned reaction to offending rather than an
inborn intuition".
To begin with, Walker gets the competing explanations a bit askew: "Inborn intu-
ition" is not the explanation competing with Walker's social upbringing explanation;
"causation by the desert of offenders" is the competing explanation of our reactions
offered by the retributivist. Secondly and more importantly, Walker is too Manichean
in his competition between explanations. Surely part of why we believe anything
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
out of the objection was done by the late John Mackie.
5
Conceding that
"it is plain that a considerable number of people have what would be
called an intuition that a wrong action in itself calls for the infliction of
suffering or deprivation on the agent",
46
Mackie nonetheless sought to
explain such intuitions as the exclusive result of biological and sociologi-
cal processes. Mackie's sort of explanation is a familiar one. It begins
with an observation about the defensive function served by retributive
attitudes: "resentment of injuries is likely to discourage similar inju-
ries, and so again will benefit the creature that feels and displays
resentment . . .".41 Such defensive function will be favoured by its
survival value in a process of natural selection."
8
Thirdly, such individu-
alized feelings of resentment take their present, generalized form be-
cause co-operation in resentment defends each individual better than
his own resentment against those that injure him could do.
49
Fourthly,
the apparent objectification of these generalized feelings of resentment
is due to the even more superior preventative effects achievable if each
regards his own resentments as being about an objectively existing
desert.
5 0
There is undoubtedly much truth in Mackie's explanation of our
retributive reactions to culpable wrongdoing. Yet as thus far stated
there is nothing in Mackie's explanation that excludes an explanatory
role for an objective moral property of desert. After all, a good survival-
related, evolutionary explanation can be given for our beliefs in the
physical sciences, yet such explanation of those beliefs does not exclude
(e.g., about physics) is due to our educational experiences; yet so long as another part
of the overall explanation as to why we believe (e.g., that protons exist) is that our
perceptual beliefs about cloud chamber phenomena are caused by protons, then we
have good reason to believe such truths of physics.
45 See Mackie, "Retributivism: A Test Case for Ethical Objectivity', supra n. 2; Mackie,
"Morality and the Retributive Emotions", supra n. 2. It is important to see that in
these papers Mackie was not arguing against retributivism in substantive ethics;
rather, he was arguing against moral realism in meta-ethics. Since my mode of
justifying retributivism presupposes moral realism (in that the objective truth of
retributivism causes us to have retributive attitudes), Mackie's attack on moral
realism becomes an attack on my way of justifying retributivism.
46 Mackie, "Retributivism: A Test Case for Ethical Objectivity', supra n. 2, at 683.
47 Ibid.
48 Ibid.
49 Mackie, "Morality and the Retributive Emotions", supra n. 2, at 8-9.
50 Ibid., at 7.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
a supplemental explanation in terms of the reality of physical objects
and their properties.
So Mackie needed to disparage the retributivist's alternative expla-
nation of our retributive attitudes (in terms of the moral property of
desert causing us to have such attitudes), and not merely to propose his
own psycho-social explanation of those attitudes. Mackie's argument
here was twofold, as he himself recognized.
5 1
Mackie's main argument
was a very general one, arguing against objectivist explanations of any
moral beliefs and not just against retributivist beliefs. The argument
was that the "queerness" of moral qualities - what they would have to
be like if they existed, how they would be related to other qualities, how
they would be known - was such that no good explanation could be
proffered if it made use of such qualities.
6 2
I shall here ignore this argument of Mackie's, since it does not
distinctively attack objectivist explanations of retributive moral beliefs
and since I have dealt with it elsewhere in some detail.
5 3
Mackie's
second argument is directed against retributive attitudes more specifi-
cally. Mackie's argument here was based on his conclusion that "all
attempts fail to make moral sense" of the retributive principle.
5 4
Yet
Mackie's argument for accepting this conclusion was surprisingly weak.
The argument consisted of running through the eleven or twelve prin-
ciples that have been invoked in the history of philosophy to justify the
retributive principle, and finding all of them wanting. From this,
Mackie inferred that there was no moral principle with which to justify
("make moral sense") of the retributive principle, and further, that
therefore the retributive principle could not be objectively true (else it
would fit better with other of our moral beliefs).
We should balk at each of these steps of Mackie's. In the first place,
the jury should still be out on Herbert Morris' fairness justification of
retributivism, since each of Mackie's objections to that justification have
spawned their own detailed literature in answer.6 Secondly, Mackie
passes too quickly over the possibility that a generalized principle of
desert reveals that the retributive principle is not isolated from other
51 Ibid.
52 Ibid., See also Mackie, supra n. 42, chap. 1.
53 Moore, "Moral Reality", supra n. 16, at 1086-88, 1117-36; Moore, 'Moral Reality
Revisited", supra n. 16 at 2501-2506, 2513-2526.
54 Mackie, "Morality and the Retributive Emotions", supra n. 2, at 4.
55 See, e.g., Jeffrie Murphy, Retribution, Justice and Therapy (Reidel, 1979); George
Sher, Desert (Princeton, 1987).
[Is.L.F. Vol. 27
JUSTIFYING RETRIBUTIVISM
moral beliefs that we hold, a possibility I explored earlier in this essay.
Thirdly, on a rather more tolerantly pluralistic view of morality's prin-
ciples than that apparently held by Mackie, the retributive principle
may itself be a first principle and thus cannot be made moral sense of,
in Mackie's use of that phrase.
The oddest thing about Mackie's argument, however, lies in none of
these particular queries. Rather, the argument as a whole goes no
distance towards establishing what it was supposed to establish, which
was the inferiority of the objectivist explanation of our retributive
reactions in terms of a moral quality of desert. The ultimate irrelevancy
of Mackie's argument to his intended conclusion stems from his failure
to frame genuinely competing explanations for our attitudes of resent-
ment. His sociological/psychological/biologica explanation is a genuine
candidate to explain these attitudes, but the twelve or thirteen prin-
ciples he refutes are not. Such principles might justify the more par-
ticular retributive judgments and attitudes; however, they do not even
purport to (causally) explain those judgments and attitudes. Indeed, it
would be a kind of category-mistake to think that one timelessly true
proposition of morality - say, that all debts to society should be paid
- could cause another proposition (like that expressed by the retributivist
principle) to be true.
The best face that one can put on Mackie's argument here is that a
moral realist (i.e., one who believes that there are objective moral
truths) must believe that morality is a seamless web of interlocking
principles. With this belief, such a realist must then infer that
retributivism is not an objectively true moral proposition, such inference
being drawn from the supposed isolation of retributivism from other
moral principles. So reconstrued, Mackie's argument is still hostage to
my earlier objections: on a pluralist view of first moral principles, being
"isolated" (in the sense of not following from some yet more general
principle) is no evidence of falsity; and the retributivism principle is not
so isolated, joining natural right theories of property, corrective justice
theories of tort, and promissory theories of contract to form a general-
ized principle of desert.
If we were to propose a better competitor to Mackie's psychological
explanation of our retributive intuitions, such explanation would be
along the following lines. There exists in the world a moral property of
relevance to punishment, namely, desert. Desert is a property of an
actor, consisting of the two moral properties of the wrongness of the act
done and the culpability with which it was done (these properties
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
themselves supervene on the properties of voluntariness, causation, and
lack of justification, for wrongness, and on intentionality or negligence,
and lack of excuse, for culpability).
The properties of wrongdoing and culpability cause most of us to
believe that, when we have culpably done wrong, such acts are evil and
that we are guilty. Given the duty to suffer brought into existence by
our culpable wrongdoing, our belief that we are guilty includes a belief
that we must suffer (i.e., be punished).
If this (moral realist) explanation of our retributive intuitions supple-
ments (it needn't supplant entirely) Mackie's explanation, then our
punitive reaction to Kant-like thought experiments gives us good evi-
dence for the truth of the retributive principle that culpable wrongdoers
must be punished.
56
4. Distinct from any of the foregoing worries is the worry based on
a distrust of the reliance upon emotions to justify retributivism. The
worry is that the particular judgments Kant-like thought experiments
call forth are suspect because they are both caused and accompanied by
strong emotional responses. The worry is not the more precise,
Nietzschean worry that the sort of emotions accompanying retributive
judgments are bad for us;
7
rather, the concern here is the more general
concern about any strong emotional reaction contaminating the ratio-
nality of the judgments it accompanies.
This worry has behind it two mistaken presuppositions. One is a
presupposition about the strength of the connection between the having
of certain emotional responses to a particular situation, on the one hand,
and the truth of certain moral judgments about the situation, on the
other. The second is a presupposition about the lack of cognitive content
to the emotions themselves. If we correct both mistaken presupposi-
tions - by weakening the connection claimed to exist between morality
and the emotions, and by strengthening the case for there being cogni-
56 For the superiority of moral realist explanations of our moral beliefs generally, see
Moore, "Moral Reality Revisited", supra n. 16, at 2491-2533; Nicholas Sturgeon,
"Moral Explanations", in David Copp and David Zimmerman, eds., Morality, Reason
and Truth (Rowman and Allanheld, 1984); David Brink, Moral Realism and the
Foundations of Ethics (Cambridge, 1989), chap. 7.
57 I explored this concern in some detail in "The Moral Worth of Retribution", supra n.
1.
[Is.L.R Vol. 27
JUSTIFYING RETRIBUTIVISM
tive content to the emotions - we will demotivate this fourth worry as
well.
I suspect that those who worry about the rationality of emotion-laden
moral judgments have one of two conceptions of how moral truth is
connected to emotional response. One is the conception of emotivism
(and of certain other forms of non-cognitivism, subjectivism, and con-
ventionalism about morals): the "truth" of a moral judgment is wholly
constituted by the fact that a person (or group of persons) has a certain
emotional response. A second conception is that of non-naturalist
intuitionism: the emotions are the special sense-data of moral truth, so
that the latter wholly depends on emotional responses in the way that
scientific truth wholly depends on perceptual responses for an empiri-
cist.
Both of these conceptions are too strong. The first makes morality
too arational, and the second makes it too queer. Better is a third
conception: our emotions are our main heuristic guide to discovering
moral truths, but they are neither constitutive of such truth nor are they
some special data base from which all moral truths necessarily are
derived.
58
Emotional response stands to moral truth like experience of
consciousness stands to truths about our minds: we happen to be so
constituted that we learn a great deal about morality and mental states
by our emotional and conscious experiences, respectively.
9
But in
neither case should we identify morals or minds with these distinctive
experiences, nor should we unduly privilege these experiences as being
their own kind of special "perception", each autonomous in its own
sphere of knowledge.
How can it be that our emotions have even this weak, "reliable
heuristic" connection to moral truth? This question gets us to the second
mistaken presupposition of the worry about emotions, for the question
more generally asked is how can our emotional responses be reliable
indicators of any truth? Aren't our emotions indicators of our subjective
states, but not of how the objective world is constituted?
58 Ibid., at 199-202.
59 This heuristics view of consciousness is defended by me in "Mind, Brain, and the
Unconscious", in P. Clark and C. Wright, eds., Mind, Psychoanalysis, and Science
(Basil Blackwell, 1988).
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
One of the more interesting developments of the philosophical psy-
chology of the last twenty years has been to suggest otherwise.
60
To
begin with, the words we use to describe our emotions are like "belief'
(and unlike the words we use to describe sensations like pain) in that
emotion-words take intentional objects. We are not simply angry or
fearful, but we are angry at something or fearful of something. This
linguistic fact construes emotions, like beliefs, to be about the outside
world, not about ourselves. Just as a belief that Jones is a thief is a belief
about Jones, so an emotion of anger at Jones is an emotion about Jones.
Moreover, emotions as we understand them obey laws of proportion-
ality. Most of us most of the time feel emotions of a certain kind and
of a certain intensity with regard to certain sorts of situations. It is
because of these laws of appropriate emotional response that psychia-
trists can talk of "displacing" emotions from their real (and appropriate)
objects to the nominal objects towards which we sometimes think our
emotions are directed.
61
Finally, there is reason to think that these laws of appropriate and
proportional emotional response are not simply products of cultural
conditioning alone, but that the reason we feel e.g., guilty, is often in
part because we are in fact guilty of having culpably done some wrong.
The moral fact of the matter often causes our moral beliefs through the
intermediate causing of our emotional responses. Our emotions in such
case become good evidence of the underlying moral landscape.
Consider by way of an example distinct from retributivism the way
in which our emotional responses have been taken to evidence certain
truths about moral dilemmas. A characteristic emotional response to
a resolved moral dilemma is to feel guilty or in some other way regretful
60 The following is a sampling of the now considerable literature on the rationality of
the emotions: Ronald de Sousa, 'The Rationality of the Emotions', in A. Rorty, ed.,
Explaining Emotions (U. of Calif. P., 1980); de Sousa, The Rationality of the Emotions
(MIT Press, 1987); R. Gordon, The Structure of Emotions: Investigations in Cognitive
Psychology (Cambridge U. P., 1987); W. Lyons, Emotion (Cambridge U. P., 1980);
Scruton, "Emotion, Practical Knowledge, and Common Culture', in A. Rorty, supra;
Calhoun, "Cognitive Emotions?', in C. Calhoun and R. Soloman, eds., What Is an
Emotion? (Oxford U. P., 1984); Roberts, "What an Emotion Is: A Sketch* (1988) 97
Philosophical Review 183-209.
61 On the implicit reliance by Freud on these laws of appropriate emotional response,
see David Sachs, "On Freud's Doctrine of the Emotions', in R. Wollheim, ed., Freud
(Anchor, 1975).
[Is.L.l. Vol. 27
JUSTIFYING RETRIBUTIVISM
after one has decided to act in a way favoring one value over another.
2
No matter how certain one might be that one has done the right thing,
still one often feels this emotional "tail" attached to the course of action
chosen. Some would see this emotional experience as the harbinger of
a deep truth about morality: that moral norms do not admit of large
numbers of exceptions to accommodate cases of conflict with other moral
norms, but that such norms are exceptionless and therefore irresolvably
conflicting.63
Our emotions, thus, should not be seen as impediments to the ratio-
nal justifiability of our moral judgments. Far from hampering our
insights into the truth, our emotions are often our best route to discov-
ering that truth. Think of how emotional many people get when the
punishment of an innocent person takes place. Surely this anger and
revulsion is not an impediment to the insight that it is unjust to punish
the innocent; surely these emotions help us to see this important moral
truth. (Indeed, we find the emotions so helpful a heuristic to moral
insight that we are tempted to excuse those who are emotionally blind,
those who used to be called "psychopaths" or "sociopaths".)
5. The fifth and last worry that has been raised about my mode of
justifying the retributive principle from particular judgments has been
a worry about the reach of the retributive principle when justified in this
way. Joel Feinberg raises the query by asking, "is the sort of case
[Moore] treats as a paradigm sufficiently representative of crime in
general for the purposes at hand?"6 David Dolinko subdivides this
worry into two concerns: (1) my "strategy cannot establish a 'duty' to
punish more than a small subset of criminals"; and (2) my strategy
"would rationally justify their punishment even if their conduct were
not criminal at all".
65
Consider the first of these two concerns. Dolinko elaborates the point
as follows:
Moore ... has stacked the deck by giving us a group of especially
savage murders and inviting us to generate, from these examples, an
intuition that "criminals" as such ought to be punished...6
62 This situation is nicely described in Herbert Morris, "Nonmoral Guilte, in F. Shoeman,
Character, Responsibility, and the Enwtions (Cambridge U. P., 1987).
63 Joseph Raz, for example, so takes this evidence. See the discussion in Moore, "Law,
Authority, and Razian Reasons", supra n. 5, at 860-61.
64 Joel Feinberg and Hyman Gross, supra n. 2, at 630.
65 Dolinko, supra n. 3, at 555.
66 Ibid., at 557.
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
Less savage acts, or acts against mala prohibita statutes, or acts that
are declared criminal by positively evil statutes, are crimes to which I
supposedly cannot generalize my results.
A blanket response to Dolinko would proceed from the view that all
criminal laws - whether prohibiting only slightly bad acts, morally
neutral acts, or even morally good acts - obligate citizen obedience, at
least in a reasonably just regime. Then all those who violate such laws
do wrong, even if that same act would not have been wrong absent the
criminalization of the act. Then those who culpably do such acts
culpably do wrong, and deserve punishment just like my savage mur-
derers.
I no longer accept the view that all criminal laws obligate citizen
obedience in a reasonably just regime. The sustained assault of Barry
Smith, Joseph Raz, Heidi Hurd, and others have convinced me other-
wise.
67
Therefore my response to Dolinko must be more piecemeal.
Consider first acts that are less immoral than the brutal murders and
rapes I have previously used to get the blood into the eyes of my readers.
Examples from Dolinko: trading on inside information in the stock
market; giving away ducklings to promote a store sale; or driving on the
wrong side of the road.6 Each of these is a moral wrong, even if not
nearly as serious a wrong as taking another human life or violating a
woman's bodily integrity. This is true of harming financial markets by
insider trading; it is even true of giving ducklings to those who are not
likely to care about them; it is certainly true of driving on the wrong side
of the road once a statute has been put in place solving the co-ordination
problem presented by crowded traffic conditions.
There should be no hesitation in generalizing the retributive intu-
ition from more serious wrongs to these less serious wrongs. Of course,
the amount of punishment should be less because the desert is less. It
is also possible that the intuition of deserved punishment is less strong
for many people in the cases of lesser moral wrongs. But surely, in
testing our commitment to a principle, we should initially focus on those
situations that work our intuition-pumps most vigorously. If the
feature that generates our judgments and that is strongly present in
such cases - say desert - is also present (even if less strongly so) in
67 M. B. E. Smith, "Is There a Prima Facie Obligation to Obey the Law? (1973) 82 Yale
L. J. 950-976; Joseph Raz, The Authority of Law (Oxford, 1979); Heidi Hurd, 'Chal-
lenging Authority" (1991) 100 Yale L. J. 1611-1677.
68 Dolinko, supra n. 3, at 557.
[Is.L.R. Vol. 27
JUSTIFYING RETRIBUTIVISM
other cases, that is reason enough to generalize to those other cases. We
do this all the time in teasing out the nature of different features of the
world. We look at clear cases of causation, for example, in order to
generate concepts and principles that we can then extend to resolve less
clear cases of causation.
6 9
Now consider acts that are morally neutral or morally good, even
after they have been made criminal by statute. Dolinko's example: an
ordinance forbidding giving food to homeless persons.
70
Dolinko defies
me to find an intuition that these offenders should be punished because
and only because they deserve it. Yet I decline the invitation. Such
offenders should not be punished. They have done no wrong and thus,
do not deserve punishment. As I have elsewhere argued,
71
retributivism
justifies punishment only when people deserve it, and desert requires
both wrongdoing and culpability. In a legal regime whose criminal
statutes criminalize morally neutral or even virtuous acts, retributivism
cannot justify punishment (and neither can any other moral principle).
Retributivism presupposes a regime in which the legislature has fol-
lowed what is variously called a "legal moralist", "perfectionist", or
"natural law" theory of criminal legislation. 72
It is important to get some sense of how this presupposition of
retributivism applies to real life regimes. As is well documented in the
literature of the 1960's and 70's in America,
7
" we have admittedly
overcriminalized behaviors. Even so, it is easy to overstate the point.
Much of what is not obviously a moral wrong like mayhem, rape and
murder, is nonetheless a moral wrong.
7 4
It is wrong to deceive people
into parting with their property, wrong to be cruel to animals, wrong to
distress others for your own amusement. Even acts that are morally
neutral become wrong when the criminal law prohibits them, if the law's
prohibition solves either a co-ordination problem or a prisoner's di-
lemma that we each have an obligation to solve. Thus, many of what
are called "regulatory offenses" may nonetheless prohibit actions that
are morally wrong; if such statutes require at least negligence for
69 E.g., H. L. A. Hart and Tony Honore, Causation in the Law (Oxford, 2nd ed., 1985).
70 Dolinko, supra n. 3, at 557.
71 Moore, "A Theory of Criminal Law Theories*, supra n. 15.
72 Ibid.
73 See, e.g., Sanford Kadish, "The Crisis of Overcriminalization" (1967) 374 Annals 157-
170.
74 See, e.g., Judith Shklar, Ordinary Vices (Harvard, 1984).
Nos. 1-2, 1993]
ISRAEL LAW REVIEW
culpability, then a retributive justification for punishment exists for
those who violate such statutes.
This may seem to strengthen Dolinko's second point, which was that
retributivism seemingly justifies punishing people when they culpably
do something morally wrong even if that wrong is not prohibited by the
criminal law. If there are as many moral wrongs as I have just
indicated, that may seem to require the retributivist to punish many
people whose behavior is not criminal.
Yet here again we need to remind ourselves that retributivists are
not monomaniacal about the achieving of retributive justice. Of course
those who culpably do some (slight or great) moral wrong deserve some
(slight or great) suffering, yet there are other intrinsic goods besides
giving culpable wrongdoers their due and sometimes these other goods
override the achievement of retributive justice. One such collection of
goods we have called the principle of legality, which forbids punishment
for acts not made criminal by statute when done.
75
In most cases the
values of liberty, fairness, equality, and utility that justify the principle
of legality outweigh the achieving of retributive justice, so that culpable
actors who do moral wrongs that are not illegal should not be punished.
But not always. As the Israeli trial of Eichmann and the Allied trials
at Nuremberg illustrate, sometimes the wrong that would go unpun-
ished if the principle of legality were observed is so enormous, and the
evil of it not being punished thus so great, that the goods of legality are
themselves overridden by the good of giving the guilty their due.
One might concede that legality stays the retributive punishment of
moral (but not legal) wrongs by courts, yet think that legality can hardly
stay the hand of legislatures in making all morally wrongful behavior
illegal. Given the broad reach of morality I have just suggested, it might
then be thought that this implication of retributivism is unacceptable.
The problem with constructing a reductio ad absurdum argument
against retributivism in this way, is that the conclusion is not at all
absurd. Of course the immorality of conduct, no matter how slight,
constitutes a prima facie reason to criminalize the behavior. What is
wrong ought to be punished, and since legality requires such wrongs
first to be criminalized, a legislature has reason to make what is
immoral also illegal. But again, this reason is subject to being overrid-
75 On the principle of legality and the values for which it stands proxy, see Moore, The
Limits of Legislation" [Fall 1984] USC Cites 23-32; Moore, Act and Crime, supra n.
2, chap. 9.
[Is.L2 . Vol. 27
JUSTIFYING RETRIBUTIVISM
den by other goods that outweigh the good of achieving retributive
justice. In particular, the liberal goods of pluralism, tolerance, and
autonomy (together with epistemic modesty in borderline cases of im-
morality) often outweigh the (minor) retributive good of punishing
minor moral wrongs. It is simply a mistake to think that a perfectionist
theory of legislation cannot be as liberal politically as one likes, depend-
ing on the weight to be given to these liberal goods.
6
III. Conclusion
Once we put aside these five worries about my justificatory methods,
we should return with renewed confidence to the particular judgments
that Kantian thought experiments call forth in us. Of course Dostoevsky's
nobleman should suffer for his gratuitous and unjustified perpetration
of a terrible wrong to both his young serf and that youth's mother. As
even the gentle Alyosha murmurs in Dostoevsky's novel, in answer to
the question of what you do with the nobleman: you "shoot him". You
inflict such punishment even though no other good will be achieved
thereby, but simply because the nobleman deserves it. The only general
principle that makes sense of the mass of particular judgments like that
of Alyosha is the retributive principle that culpable wrongdoers must be
punished. This, by my lights, is enough to justify retributivism.
76 See Moore, "A Theory of Criminal Law Theories", supra n. 15, at 181-83; Moore,
"Sandelian Anti-Liberalism' (1989) 77 Calif. L. R. 539-551; Joseph Raz, The Morality
of Freedom (Oxford U. P., 1987); Robert George, Making Men Moral. Civil Liberties
and Public Morality (Oxford U. P., 1993).
Nos. 1-2, 1993]