Content uploaded by John Shook
Author content
All content in this area was uploaded by John Shook on Oct 17, 2019
Content may be subject to copyright.
Dewey’s Rejection of Retributivism and His
Moral-Education Theory of Punishment
John Shook
This essay examines the two important arguments that John Dewey’s
social philosophy offers against the retributivist justification for punishment.
The first argument is that retributivism cannot serve as an adequate expla-
nation for the moral practice of punishment, while the second argument
develops the position that a liberally democratic society would naturally
prefer an alternative to the retributive model of punishment that is more con-
sistent with democracy’s concern for the social good. Although Dewey did
not fully develop a systematic theory of punishment, his writings on moral-
ity, education, and democracy provide the foundations for constructing a
Deweyan approach to the justification of punishment. The cornerstone is
Dewey’s theory of moral education for a liberal democracy. This theory points
the way toward an explanation of why and how society should aim at con-
tinually developing the moral responsibility of all citizens. In brief, punish-
ment should play an essential role in moral education by enhancing the ability
of offenders to be more socially responsible in the future. Examination of
Dewey’s arguments against retributivism shows how he provides an alter-
native to both deontological and utilitarian approaches to punishment,
thereby establishing a novel version of the moral-education theory of
punishment.
1. Retribution and the Moral Practice of Punishment
The relationship between punishment, responsibility, and morality can be
expressed in a deceptively simple manner: Punishment minimally requires
the moral disapproval of a person’s conduct (broadly understood), for which
he or she is responsible. Dewey believed that it is essential to distinguish two
divergent notions of responsibility, which in turn ground two very different
approaches to morality, and thus to punishment.1The past-looking or “ret-
rospective” concept of responsibility is involved when we assess the various
relevant factors of the situation leading up to a person’s action to determine
whether that person really performed the action. The forward-looking or
“prospective” concept of responsibility is involved when we attend to the
wider context of the past and future behavior of the person, so that her
future behavior may be adjusted by our interventions. Dewey placed great
weight on this distinction; for example, in his 1932 book Ethics he wrote the
following:
JOURNAL of SOCIAL PHILOSOPHY, Vol. 35 No. 1, Spring 2004, 66–78.
© 2004 Blackwell Publishing, Inc.
Now the commonest mistake in connection with the idea of responsibil-
ity consists in supposing that approval and reprobation have a retro-
spective instead of prospective bearing. The possibility of a desirable
modification of character and the selection of the course of action which
will make that possibility a reality is the central fact in responsibility. The
child, for example, is at first held liable for what he has done, not because
he deliberately and knowingly intended such action, but in order that in
the future he may take into account bearings and consequences which he
has failed to consider in what he has done. Here is where the human agent
differs from a stone and inanimate thing, and indeed from animals lower
in the scale.2
Dewey’s claim that “responsibility” is essentially a normative term of moral
judgment understandably appears false at first reading. Do we not say that
apple trees are responsible for apples, and that the moon is responsible for
tides? But Dewey is not arguing that all uses of the term “responsibility” are
moral uses. Rather, he is arguing that when a morally normative verdict of
responsibility is made on a person’s behavior, this moral verdict is essenti-
ally made for the purpose of affecting future behavior. It is this future-
regarding core to moral evaluation that leads Dewey toward his rejection of
retributivism.
It is not absurd to declare a person responsible for an unintended action,
so long as responsibility is understood in a sufficiently broad sense that does
not require intent. It is this broad sense that Dewey is after to make his point
regarding our treatment of children. Did the child unintentionally hurt the
cat? If punishment requires responsibility, and responsibility requires intent,
then we cannot punish a child for thoughtless behavior. However, as every
parent well knows, a child’s punishments are more often than not earned by
unintentional conduct. Therefore, while the presence or lack of intent may
have a bearing on the degree of moral disapproval, intent is not essential to
the sort of responsibility connected to punishment (and the same is true for
adults, a point to be pursued below). This point alone is of course insufficient
to discredit retributivism; a retributivist could note that negligent behavior,
an act of omission to take greater care, is just as punishable. Dewey quite
agrees and wants to pursue this point further. Since punishment is deserved
for failure to take greater care as well as for intentional acts, then how we
punish should be connected to our more fundamental interest that people
take greater responsibility in general for their conduct, instead of connecting
merely to our interest that people be punished for intentionally responsible
conduct. We punish when people are held responsible in the broad sense and
we desire to modify their future behavior; both conditions (past and forward-
looking) are essentially linked together, and they must be satisfied (at
minimum) to justify punishment. Dewey’s conclusion is that the only way to
make sense of our practice of punishing for failures of responsibility in the
broad sense (for both commission and omission) is to see punishment as nec-
essarily involving future consequences.3The practice of making morally nor-
mative judgments about behavior and applying punishment accordingly
must have originated in, and continues to exist for the sake of, efforts to
Dewey’s Rejection of Retributivism 67
modify people’s habits of conduct to make those people more morally respon-
sible in the future. Having established this conclusion to his satisfaction,
Dewey goes on to argue that our efforts to enhance future responsibility can
be accurately characterized as efforts to form moral character, and thus to
morally educate people.4For this reason, Dewey’s consequentialist approach
to punishment is most appropriately categorized with moral-education
theories of punishment.
The merits and problems of Dewey’s moral-education theory will be
considered at the conclusion of this essay. First, we must examine whether
Dewey’s position on retrospective and prospective views of responsibility
and punishment can be effective against retributivism. The first question
to be asked is, Why does Dewey use an example of punishing a child to
exemplify his approach? After all, we do not regard the punishment of
children as the primary difficulty calling for philosophical justification.
Furthermore, it may be the case that our legal treatments of adults and of
children are so distinct as to amount to a difference of kind, and not just
degree. Perhaps Dewey is correct concerning our treatment of children but
fails to appreciate that a quite different kind of punishment is applied to
adults.
But where lies such an essential difference? For both children and adults,
moral disapproval involves our desire to try to prevent future offenses. If we
disapprove of some event (such as the cat’s suffering) and discover a person’s
responsibility for that event, we naturally transfer our disapproval to that
person’s conduct as well. Our disapproval is of that kind of conduct, and not
merely a disapproval of that particular action. Our view that such conduct is
wrong entails, among other things, our view that such conduct should not
happen (and not merely that this specific instance of conduct should not have
happened). To disapprove of someone’s behavior today, and yet to be indif-
ferent to that same behavior tomorrow under identical conditions, would be
to take an attitude that does not fall within the range of moral evaluation. A
moral evaluation of a person’s conduct as wrong is necessarily connected with
a desire to see that person’s conduct modified for the better in the future. This
is as true for adults as it is for children. Because punishment follows from a
disapproving moral evaluation, and such evaluations necessarily involve the
desire to modify future conduct, punishment must for all people have a
future-looking perspective.
To summarize so far, the foundation for Dewey’s theory of punishment
is his position that the assessment of past responsibility is a necessary first
step for attempting to modify a person’s future responsibility, which explains
how the retrospective and prospective conceptions of responsibility are
related. The proper function of first judging someone’s past responsibility is
to use that information subsequently to engage in a process of altering future
responsibility (hopefully) toward more consistent moral behavior. Of course,
we do make judgments about the past or “causal” responsibility of nonagents
such as an apple tree and the moon. Dewey holds that insofar as we treat
human beings as having only causal responsibility, we do not treat them as
responsible agents, and thus we do not treat them as moral agents. Therefore,
his primary contention is that when we are concerned with morality, we are
68 John Shook
concerned with developing the future responsible agency of a person, and
thus moral education is the relevant practice involved. Our scrutiny of and
verdicts upon a person’s causal responsibility, when made in the appropriate
moral setting, are made for the sake of then attempting to enhance future
responsibility. The connection between responsibility and morality rests on
the fact that future responsibility is successfully enhanced when a person’s
habits of conduct more closely conform to moral expectations. At this early
stage we can see that Dewey clearly prefers something like a deterrence stance
over simple retributivism, since the future conduct of an offender is for
Dewey the central consideration.
What about the connection between punishment and responsibility?
Dewey, like most philosophers, holds that the judgment that a person should
be punished is essentially a normative moral judgment. Given his stance that
normative moral judgments are made ultimately for the sake of trying to
modify persons’ future conduct, he therefore finds that the essential purpose
of punishing an offender is to enhance the offender’s future moral responsi-
bility. This view (let us call it the “prospective” theory of punishment) is
diametrically opposed to any “retrospective” theory of punishment. A retro-
spective theory of punishment instead declares that the essential purpose
of punishment has nothing to do with enhancing or diminishing the
offender’s future responsibility. Instead, the retrospective approach holds that
ascertaining past responsibility for some offense is sufficient to determine
whether, and in what amount, an offender should be punished. Retribu-
tivism, at least in its pure form that wants no compromise with utilitarian-
ism,5is the classic example of a retrospective theory of punishment. The
concept of retribution is vague, and proponents cannot agree on its essentials,
but it is most often defined as involving either (a) the attempt to remove the
stigma, or cancel out the sin, of the offense, or (b) the attempt to restore some
sort of balance by causing the offender to suffer in proportion to the offense.
In both varieties, (pure) retributivism staunchly denies that the future conse-
quences of punishment should determine the form or severity of punishment.
Because retributivism holds that the nature of the crime itself is sufficient
to decide such matters, it is a paradigm type of retrospective theory of
punishment.
From Dewey’s perspective on the proper relationship between assessing
past responsibility and enhancing future responsibility, the retrospective
theory of punishment effectively removes and isolates the assessment of past
responsibility from its proper moral setting. That is the primary reason for
Dewey’s rejection of (pure) retribution: It ignores the proper moral context
that gives our practices of assessing responsibility their meaning and effec-
tively ceases to treat offenders as moral agents.6On Dewey’s moral theory, to
treat a person as an agent necessarily involves, in part, trying where appro-
priate to adjust that person’s future responsibility. There is an inherent rela-
tionship between making moral judgments and acting on those judgments to
attempt to enhance others’ moral responsibility. Dewey explicitly connects the
responsibility of any person with the responsibility of others to sustain and
enhance future moral conduct. Retribution stands condemned as antithetical
to the responsible moral treatment of persons.
Dewey’s Rejection of Retributivism 69
The persons who employ praise and blame, reward and punishment, are
responsible for the selection of those methods which will, with the great-
est probability, modify in a desirable way the future attitude and conduct
of others. There is no inherent principle of retributive justice that com-
mands and justifies the use of reward and punishment independently
of their consequences in each specific case. To appeal to such a principle
when punishment breeds callousness, rebellion, ingenuity in evasion,
etc., is but a method of refusing to acknowledge responsibility. Now the
consequence which is most important is that which occurs in personal
attitude: confirmation of a good habit, change in a bad tendency. The
point at which theories about responsibility go wrong is the attempt to
base it upon a state of things which precedes holding a person liable,
instead of upon what ensues in consequence of it. One is held responsi-
ble in order that he may become responsible, that is, responsive to the
needs and claims of others, to the obligations implicit in his position.
Those who hold others accountable for their conduct are themselves
accountable for doing it in such a manner that this responsiveness
develops.7
In short, Dewey convicts the retrospective theory of punishment generally
and retributivism in particular for endorsing the unintelligent and irrespon-
sible infliction of suffering on people who are not being properly treated as
moral agents. It is contradictory, in Dewey’s view, to morally judge a person
to have behaved irresponsibly and to be deserving of punishment, and then
to carry out the punishment in a manner blind to its effects on that person’s
future capacity for moral conduct. Even more paradoxical, from Dewey’s per-
spective, would be any effort to resist Dewey’s conclusion by challenging the
key premise exposed above: that to make a moral evaluation of disapproval
is to disapprove of that kind of conduct, and not merely a disapproval of that
particular action. A retributivist might prefer to hold that a moral judgment is
only made upon a particular instance of conduct that entails nothing about
judgments on other such conduct under relevantly similar circumstances. But
that sort of particularist moral theory has never been widely accepted and
has more in common with emotivist and subjectivist moral theories, which
actually conflict with the typical moral orientation of retributivism.
The failure of retributivism to promote morally right conduct sufficiently
cannot be overcome by emphasizing retributivism’s ability to satisfy some
abstract principle of justice. Dewey declared that the largest obstacle to
improving our system of punishment is the lingering sanctity attached to ret-
ributivist notions. If the justification of government penal systems is simply
the sacred principle of justice for its own sake, then it is difficult to repress
the idea that there is really no moral justification for penal systems at all:
It is the sacrosanct character thus attributed to the State’s use of force
which gives pungency to the Tolstoian charge that the State is the arch-
criminal, the person who has recourse to violence on the largest scale. I
see no way out except to say that all depends upon the efficient adapta-
tion of means to ends. The serious charge against the State is not that it
70 John Shook
uses force—nothing was ever accomplished without using force—but
that it does not use it wisely or effectively. Our penal measures are still
largely upon the level which would convince a man by knocking him
down instead of by instructing him.8
Dewey rightly distinguished between force and violence, and between mere
force and intelligently constructive uses of force. Intelligently applied pun-
ishment, by enhancing future responsibility in a morally justifiable way, could
never be properly construed as mere violence, whereas retribution is nothing
but violence that lacks moral justification.9Our commitment to morality must
be matched by our commitment to creating and maintaining morally respon-
sible agents.
For Dewey, punishment should be a moral practice that respects the
offender as a moral agent. This stance removes from consideration two other
types of consequentialist approaches, the “incapacitation” and “restitution”
approaches to punishment. The incapacitation approach, in which offenders
are simply prevented from offending again (using banishment, prison, etc.),
is consistent with the principle that punishment should deal with future
conduct. However, its application of simple brute force is inconsistent with
treating offenders as moral agents. Punishment of moral agents would aim
not just at the prevention of future wrong conduct, but rather at the encour-
agement of future right conduct. The restitution approach likewise involves
the offender’s future conduct, by holding that punishment functions to
redress the offender’s harm by forcing the offender to proportionally com-
pensate the victim(s) of his offense. But restitution seeks to control the
offender’s conduct for the short term and on a case-by-case basis; its essen-
tial aim is not to increase the offender’s capacity for responsible conduct.
Dewey is seeking the consequentialist approach that best respects offenders
as moral agents, and this approach clearly conflicts with retributivism.
There remain some consequentialist/deterrence approaches that are not
ruled out by Dewey’s arguments. For example, a prominent consequentialist
approach, rehabilitationism, can be understood as an effort to deal with an
offender’s future moral responsibility. Rehabilitation’s aim of modifying a
person’s character toward adequate law-abiding conduct could (depending
on the methods applied) satisfy Dewey’s requirement that punishment
should be a moral practice that respects the offender as a moral agent. Moral-
education approaches like Dewey’s are not improperly classified under reha-
bilitationism’s umbrella. The conclusion of this essay further discusses the
relationship between Dewey’s moral-education approach and similar theo-
ries. Dewey’s second argument against retributivism is the topic of the next
section.
2. Punishment and the Social Good
Dewey’s allegiance to a prospective approach to responsibility and pun-
ishment can be further explained by setting it within the larger context of his
understanding of the nature of social responsibility in a democracy.10 This
understanding provides the grounds for raising Dewey’s second objection to
Dewey’s Rejection of Retributivism 71
retributivism. Dewey is but a recent example of a philosopher committed to
progressive liberal democracy who has explicitly grounded the purpose of
laws, rights, and responsibilities upon the advancement of the social good.
However, this emphasis on the social good should not be taken to imply that
Dewey was a classical utilitarian and must not license the inference that
Dewey’s consequentialist approach to punishment is not really different from
ordinary utilitarianism (more on this below). Unlike the typical utilitarian,
who asserts that social goods are dependent on the existence of individual-
ized private goods, Dewey doubted whether many important goods are
easily individualized. With the utilitarians Dewey agreed that social norms
should be judged by their ability to facilitate the production of social goods.
Yet Dewey always added the proviso that many important social goods are
valuable only to the extent that most or all members of society participate in,
and subsequently enjoy, these goods’ production. Due regard for the perva-
sive role of social goods in society would help to resolve the traditional util-
itarian problem of gratifying one part of society by degrading another.
Opponents of utilitarianism often use the concept of justice to highlight this
utilitarian problem by claiming that the search for the social good must be
replaced by, or at least balanced against, respect for principles of justice. But
Dewey felt no need to appeal to justice conceived as independent of the social
good. Justice for its own sake, he argued, has never truly been an abiding
principle of democratic life. Citizens of a democracy desire justice, certainly;
but they are rightfully suspicious of claims that some particular law, right, or
morality transcends any and all considerations of the welfare of society. This
is as true for the typical conservative protector of constitutional rights as it is
for the most progressive experimenter with novel freedoms. Furthermore,
citizens of a democracy do care about the impact of enforcing justice on the
future behavior of the punished, and this concern is heightened when the
public costs of recidivism are appraised.11
From the standpoint of Dewey’s philosophy of democracy, justice-
centered theories attempt to ground the existence of justice in some nonso-
cial realm, where it is safe from social evaluation and modification. This
attempt, however, only succeeds in isolating justice from human intelligence:
That something called justice is to be done even though it bring down
the heavens in ruin, is the final word of an ethics which is resolutely irrel-
evant to the circumstances of action and the conditions of life. But the
skill which invented the phrase dulls by the side of the greater cunning
which disguised the fact that the phrase is an ironic summary of every
moral code that leaves consequences out of account. For it is only acci-
dent if abstract ethics does not always bring wreck in its train. Yet those
who recommend this adage of an unswerving morality always wear an
expression of superiority though wilfully disregarding the consequences
of wreaking vengeance under the plea of vindicating eternal justice.12
Dewey does hold that punishment must serve justice. However, for punish-
ment to be morally responsible, the justice that punishment serves must be
deeply connected to the social good. Justice-centered theories often portray
72 John Shook
justice and its attendant values of legal duty, or rights, or righteousness as
nonconventional and exempt from human deliberation. The debate between
deontologists and utilitarians, from Dewey’s perspective, has obscured the
most significant issue: the nature and origin of social responsibility. Social
responsibility is, first, the only kind of responsibility that democratic society
truly cares for; and second, it is grounded in society and its structure, and not
in some nonsocial realm.
Dewey steadfastly repudiated any methodology that involves looking
beyond the social sphere to discern moral or legal foundations.13 For example,
that deontological variety of moral theory that locates moral goodness exclu-
sively within the private space of autonomous motives has forgotten that we
ascertain and judge the particular motive of a person’s act only by judging
that act in the wider context of that person’s patterns of responsible or irre-
sponsible social conduct. All judgments of personal character are judgments
of that person’s social conduct, and vice versa. A “free-will” theory of moral
responsibility postulates an individual’s free will as the cause for responsi-
ble, or irresponsible, moral behavior. But such a theory, instead of clearing
the space for a justification of punishment, actually makes punishment of the
person irrational:
It is a man, a human being in the concrete, who is held responsible. If the
act does not proceed from the man, from the human being in his concrete
make-up of habits, desires and purposes, why should he be held liable
and be punished? Will appears as a force outside of the individual person
as he actually is, a force which is the real ultimate cause of the act. Its
freedom to make a choice arbitrarily thus appears no ground for holding
the human person as a concrete being responsible for a choice. Whatever
else is to be said or left unsaid, choice must have some closer connection
with the actual make-up of disposition and character than this philoso-
phy allows.14
We do esteem the individual’s consciously willful efforts to behave
morally. To this extent Dewey agreed with Kant that good will is a proper
object of moral approbation. However, esteem for such moral efforts of char-
acter is justifiable only because we esteem morally responsible conduct, and
not because willful efforts to perform the good regardless of actual conse-
quences are morally valuable in themselves. Moral consequentialists are
correct to claim that we primarily encourage people to perform moral actions
and not merely to have moral character. Yet moral consequentialism fails to
realize adequately that we esteem morally responsible conduct, and not merely
isolated and thoughtless actions that just happen to meet our moral stan-
dards. Society promotes enhanced responsibility by encouraging the devel-
opment of autonomously responsible character, which is identical to the
development of responsible conduct. If our bad conduct does not match our
good intentions, then society is willing to recognize excusing conditions.
However, society also demands that in the future we take greater care in
understanding existing conditions in order to execute our good intentions
successfully.
Dewey’s Rejection of Retributivism 73
Society naturally demands social responsibility, and its members must
take the moral responsibility for enhancing such responsibility in all citizens.
Justice, according to social-good-centered theories such as Dewey’s, exists for
the sake of promoting harmonious social cooperation promoting the social
good. Justice therefore demands that each member of society receive the
social treatment that corresponds to the degree of socially responsible behav-
ior exhibited. Punishable actions are actions of irresponsible noncooperation
that impede the production of social goods.15 Some form of intervention with
respect to the offender, interrupting and altering the offender’s patterns of
conduct, will be sometimes necessary to modify habits of conduct (less severe
offenses may require only admonishments and warnings). Combining
Dewey’s defense of the prospective theory of punishment with his social-
good notion of justice permits us to sketch the outline of a Deweyan model
of punishment: Punishment is a type of serious social intervention with
respect to the offender that justly aims at enhancing the offender’s future
moral responsibility to cooperate in the production of social goods. This is
not a complete definition of punishment, since the techniques to be used in
such an intervention are not yet delineated. But at minimum the techniques
must be carefully designed to serve the purpose of enhancing the offender’s
future moral responsibility. The techniques typically preferred under ret-
ributivist models may haphazardly and in part accomplish this same goal,
but retributivism is marked by its failure to prioritize the offender’s future
conduct.
Our future responsibility, our capacity to engage in moral conduct more
or less consistently, is the overriding moral concern for a society (especially
a democratic society), according to Dewey. Punishment, because it is a prac-
tical expression of this social concern for moral conduct, must essentially aim
at enhancing future responsibility. Theories of punishment based on abstract
principles of justice, free-will autonomy, or calculations upon the conse-
quences of conduct have obscured this fundamental social concern for future
responsibility. Dewey’s prospective approach to responsibility and punish-
ment is a much-needed third alternative to deontological and utilitarian
approaches. The social good requires social responsibility, which in turn
requires social interventions in others’ conduct as needed. But those inter-
ventions must satisfy justice while at the same time morally treating the
offender as a moral agent.
The social good cannot simply and directly justify punishment, since that
leaves the (utilitarian) problem of justifying punishment of the innocent. On
Dewey’s moral theory, society has the moral duty to punish only the truly
guilty because they are guilty, thus agreeing on that point with retributivism.
However, the duty to punish the guilty simply and directly because they are
guilty cannot fully justify punishment, since that leaves the (deontological)
problem of determining how punishment should proceed in the face of con-
textual considerations that weigh against simple “eye for an eye” or propor-
tional punishment. “Punishment only for the guilty” may sound like a
damaging concession to retrospective retributivism, but from Dewey’s per-
spective, there is a straightforward explanation needing no retributionist
principles: Only the guilty need to have their future conduct modified. On
74 John Shook
Dewey’s moral theory, society has the moral duty to attempt to enhance future
responsibility, if indeed society is committed to any genuinely moral stan-
dards. This prospective theory is the start of an explanation as to why our
practices of assigning responsibility, placing moral blame, and determining
punishment can be best understood as social practices that together aim at
enhancing future responsible conduct for all members of society.
3. Criticisms of Dewey’s Moral-Education Approach
As a type of moral-education theory, Dewey’s prospective approach to
punishment must answer the usual criticisms directed toward such theories.16
Afull defense against the contentions of rival theories of punishment cannot
be mounted here for lack of space and the absence of a detailed Deweyan
model of punishment.17 Still, some preliminary observations concluding this
essay might add plausibility to Dewey’s approach. To begin, the moral-
education theory is sometimes characterized, by both proponents and detrac-
tors, as opening the way toward an alternative to punishment, and not as
offering a different justification of punishment. However, such characteriza-
tions presume a retributivist essence to the nature of punishment that Dewey
rightly calls into question. A variation on this point is the claim that punish-
ment must be a harm to the offender, not a benefit, and thus a moral-
education theory is really just about replacing punishment. From the Deweyan
perspective, merely causing harm to the offender is insufficient to morally
justify punishment. Therefore, regardless of whether the offender or society
finds the moral-education approach’s process of punishment to be harmful,
that process remains justified through its attempt to increase the offender’s
future responsibility. The point of punishment is modification of future
conduct, not harm for its own sake. Those unable to countenance a process of
punishment that might benefit an offender in the long run should reconsider,
as Dewey urges in his antiretributivist arguments, how our genuine moral
duties and democratic goals suffer at the hands of retributivism.
Another potential difficulty is that a moral-education theory is often
defended by appeal to principle(s) of human right or autonomy that are them-
selves quite controversial and need prior defense. Dewey’s moral-education
approach, however, does not rely on abstract principles of right or auton-
omy.18 His approach does rely on the assumptions that societies encourage
moral responsibility and that actual punishment techniques can be designed
that will (often enough) succeed in moral education. This opens the door for
the objection that the specific modes of punishment that a moral-education
approach can endorse are neither severe nor effective enough to prevent
future offenses. Like the rehabilitation theory of punishment, the moral-
education theory has been reproached for being too soft-hearted and gener-
ous toward the offender. How can society show that it takes offenses
seriously, retributivists complain, unless stern suffering is imposed?
Society’s reasonable need to promote and display its moral commitments
can be satisfied by a moral-education theory if we understand with Dewey
that our “moral commitment” is our commitment to moral agency (the doer),
and not merely to moral conformity (the deed). This understanding of moral
Dewey’s Rejection of Retributivism 75
commitment helps to answer the counterpart question of whether moral-
education theory might instead be far too severe in its pursuit of moral char-
acter. It appears to set no boundaries on the invasiveness and length of
punishment to achieve the required responsibility, which conflicts with the
proportionality principle, our abhorrence of torture and brainwashing, and
our fear of unrestrained government power. Three Deweyan responses are
available. First, proportionality would be respected, since the educational
effort required should match the degree of increased responsibility needed.
Stealing should not be met with decades of reform to achieve moral saint-
hood. Second, Dewey’s moral-education approach is grounded on respect for
all persons, even offenders, as moral agents, so that morally unacceptable
punishment techniques are ruled out. Third, even where the offense is serious
and the offender has prolonged resistance to education, punishment should
have a preset, proportional time limit. Without definite sentence terms, pun-
ishers might not be sufficiently motivated to maximize their efforts, and the
punished would be more likely to resent and resist the punishment process.
Furthermore, unrestrained government power is incompatible with the
democratic component supporting Dewey’s moral-education approach.
Many rehabilitationists have endorsed these three responses as well. A
rehabilitationist theory can be pragmatically distinguished from Dewey’s
moral-education approach if that theory promotes rehabilitation for some
reason(s) other than respect for the offender’s moral agency. To the extent that
such a rehabilitation theory fails to display such respect, it will be unable to
respond adequately to one or more of the three concerns raised above.
However, it should not be difficult for a rehabilitationist to see the cogency
of Dewey’s position. A deeper problem still lurks: Perhaps rehabilitationism
and moral education may stand together only to fall together. Retributivists
often claim that other theories, especially rehabilitationist and moral-
education theories, cannot appreciate the kind of justice required for mur-
derers (for example), either because (a) such serious offenders could never be
sufficiently rehabilitated, or (b) trying to rehabilitate them simply fails to take
their crime seriously by attempting to benefit those who deserve only grave
harm. A reply to (b) has been supplied already. As for (a), a moral-education
approach should acknowledge that for the most serious crimes like murder,
some offenders may never earn our trust in their responsibility and thus never
rejoin normal society. A moral-education approach would not regard life in
prison as an unacceptable outcome. Its rejection of capital punishment also
has the advantage of displaying society’s deep respect for human life and
moral agency, while at the same time helping the truly innocent by provid-
ing for the possibility that a sentence could be overturned by new evidence.
Punishment should aim at teaching offenders and potential offenders
why certain types of conduct are wrong and how to avoid choosing such
conduct in the future. Another close cousin of moral-education theory, the
reintegration model of punishment, also agrees on this principle and shares
many essential features with Dewey’s approach.19 Together with rehabilita-
tionism, this family of consequentialist theories rejects retributivism for its
inability to provide deterrence strategies that are both intelligent and morally
sound. Accordingly, a retributivist’s claim that only her theory can supply
76 John Shook
sufficient justification for deterrence can no longer be convincing. Pure ret-
ributivism cares little for future consequences, and if a retributivist wishes to
join forces with the consequentialists instead, then the issue must be joined
on the empirical ground of careful studies comparing the actual deterrence
provided by the various models. Whether Dewey’s own model of moral edu-
cation would perform well remains to be seen. Only empirical study of actual
punishment techniques is relevant here, and the current conflicted state of
legal theory on punishment indicates that alternative models and modes of
punishment do deserve trial.
Aearly version of this essay benefited from discussion at the 29th Annual Confer-
ence on Value Inquiry in Tulsa, Oklahoma, in 2001. I am deeply indebted to my col-
league Eric Reitan for many helpful suggestions along the way.
Notes
1Auseful examination of Dewey’s work on responsibility and freedom is that by Abraham
Edel, Ethical Theory and Social Change: The Evolution of John Dewey’s Ethics, 1908–1932
(New Brunswick, N.J.: Transaction, 2001), chap. 10.
2John Dewey, The Later Works of John Dewey, vol. 7: Ethics, ed. Jo Ann Boydston (Carbon-
dale: Southern Illinois University Press, 1985), 303.
3In H. L. A. Hart’s terminology, a “consequentialist” justification of punishment is defined
as “essentially forward-looking; it considers the future good we can do to society
including the criminal.” Hart, Punishment and Responsibility (Oxford: Oxford University
Press, 1968), 159. Hart’s definition is designed with a view toward grappling with util-
itarianism. This definition is not used in this essay, and Dewey’s theory (as I show) is
not another utilitarian effort. Dewey is here arguing that the justification for punish-
ment must involve consideration of its consequences; and so it is a consequentialist
approach, but in a broader sense than Hart’s.
4See Suzanne Rice, “Dewey on Virtue, Character, and Moral Education,” Review Journal of
Philosophy and Social Science 26 (2000): 75–89.
5An example of this unification quest is explored by Johann Klaassen, “Punishment and
the Purification of Moral Taint,” Journal of Social Philosophy 27 (1996): 51–64.
6This argument deserves more elaboration than can be afforded here, and it is but one way
to attack retributivism. For example, as Russ Shafer-Landau has argued, retribution has
no resources for determining why harm, and what amount of harm, is deserved by the
offender. See Shafer-Landau, “The Failure of Retributivism,” Philosophical Studies 82
(1996): 289–316.
7Dewey, The Later Works of John Dewey, vol. 7: Ethics, 304–5. Dewey’s most fervent diatribe
against retributivism (too long for quotation here) is found in his 1922 Human Nature
and Conduct, in Dewey, The Middle Works of John Dewey, vol. 14, ed. Jo Ann Boydston
(Carbondale: Southern Illinois University Press, 1983), 1–230, at 17–18.
8John Dewey, “Force and Coercion,” in The Middle Works of John Dewey, vol. 10, ed.
Jo Ann Boydston (Carbondale: Southern Illinois University Press, 1980), 244–51, at
248.
9See Joseph Betz, “Violence: Garver’s Definition and a Deweyan Correction,” Ethics 87
(1977): 339–51.
10 On Dewey’s political theory see Robert Westbrook, John Dewey and American Democracy
(Ithaca, N.Y.: Cornell University Press, 1991), 319–73, and Michael Eldridge, Trans-
forming Experience: John Dewey’s Cultural Instrumentalism (Nashville, Tenn.: Vanderbilt
University Press, 1998).
11 Dewey’s position can also usefully illuminate the problem of past crimes that confronts
emerging democracies. A similar view, that the retribution model only increases polit-
ical tensions by prohibiting the growth of mutual trust, is advanced by Aleksandar
Dewey’s Rejection of Retributivism 77
Fati , “Retribution in Democracy,” in Political Dialogue: Theories and Practices, ed. S.
Esquith (Amsterdam: Rodopi, 1996), 335–55.
12 John Dewey, “Fiat Justitia, Ruat Coelum,” in The Middle Works of John Dewey, vol. 10, ed.
Jo Ann Boydston (Carbondale: Southern Illinois University Press, 1980), 281–84, at
281.
13 Dewey never found persuasive any appeal to transsocial rights. He was particularly
skeptical about that sort of alleged rights that have the curious dual properties of being
sufficiently general to be possessed by all people simply because they are human, yet
particular enough that we can detect when someone’s behavior has caused one or more
of those rights to evaporate. Daniel McDermott relies on these suspicious dual prop-
erties in his attempt to explain why we can harm people simply because a “human”
right has somehow vanished. McDermott, “The Permissibility of Punishment,” Law and
Philosophy 20 (2001): 403–32.
14 John Dewey, “Philosophies of Freedom,” in The Later Works of John Dewey, vol. 3, ed. Jo
Ann Boydston (Carbondale: Southern Illinois University Press, 1984), 92–114, at 94.
15 The necessity of punishment for developing stable group cooperation is supported by
game-theoretic analyses of reciprocity. Robert Boyd and Peter Richerson conclude that
if sufficiently heavy punishments are enforced, moralistic strategies (those that coop-
erate, punish noncooperators, and punish those who do not also punish noncoopera-
tors) can be evolutionarily stable strategies. Boyd and Richerson, “Punishment Allows
the Evolution of Cooperation (or Anything Else) in Sizable Groups,” Ethology and Socio-
biology 13 (1992): 171–95.
16 For a liberal formulation of a moral-education theory and some responses to criticism,
see Jean Hampton, “The Moral Education Theory of Punishment,” Philosophy and Public
Affairs 13 (1984): 208–38. See also Jeffrie Murphy, “Retributivism, Moral Education and
the Liberal State,” Criminal Justice Ethics 4 (1985): 3–10; and Ido Weijers, “Punishment
and Upbringing: Considerations for an Educative Justification of Punishment,” Journal
of Moral Education 29 (2000): 61–73. Moral-education theory and reintegration theory
have many connections with the “penance” model of punishment and also to “restora-
tive” justice. See R. A. Duff, Trials and Punishments (Cambridge: Cambridge University
Press, 1986); Robert Lipkin, “Punishment, Penance and Respect for Autonomy,” Social
Theory and Practice 14 (1988): 87–104; Wesley Cragg, The Practice of Punishment: Towards
a Theory of Restorative Justice (London: Routledge, 1992); Aleksandar Fati , Punishment
and Restorative Crime-Handling: A Social Theory of Trust (Aldershot, England: Avebury-
Ashgate, 1995); Francis Schweigert, “Learning the Common Good: Principles of Com-
munity-Based Moral Education in Restorative Justice,” Journal of Moral Education 28
(1999): 163–83; and Kent Roach, “Changing Punishment at the Turn of the Century:
Restorative Justice on the Rise,” Canadian Journal of Criminology 42 (2000): 249–80.
17 Such a model is developed in a separate paper that delineates the necessary steps in the
intervention process of moral education and defends this process against further
objections.
18 See John Dewey, ACommon Faith, in The Later Works of John Dewey, vol. 9, ed. Jo Ann
Boydston (Carbondale: Southern Illinois University Press, 1986), 21–39, and “Democ-
racy and Human Nature,” in The Later Works of John Dewey, vol. 13, ed. Jo Ann Boyd-
ston (Carbondale: Southern Illinois University Press, 1988), 136–55. Dewey’s rejection
of autonomy for its own sake circumvents Russ Shafer-Landau’s criticism that the
theory of punishment as moral education is unstably grounded on respect for moral
autonomy. Shafer-Landau, “Can Punishment Morally Educate?” Law and Philosophy 10
(1991): 189–219.
19 See John Braithwaite, Crime, Shame and Reintegration (Cambridge: Cambridge University
Press, 1989); John Braithwaite and Philip Pettit, Not Just Deserts (Oxford: Oxford Uni-
versity Press, 1990); John Braithwaite, “Shame and Criminal Justice,” Canadian Journal
of Criminology 42 (2000): 281–98; and Eric Reitan, “Punishment and Community: The
Reintegrative Theory of Punishment,” Canadian Journal of Philosophy 26 (1996): 57–81.
c
¢
c
¢
78 John Shook
