views updated


Retributivism is first and foremost a theory of punishment. It answers the question, Why do we have punishment institutions? The answer it gives is very simple: for the retributivist, we are justified in punishing persons when and only when they deserve to be punished. To avoid question-begging circularity, "deserve to be punished" in the above definition cannot simply mean "ought to be punished." Rather, to deserve punishment means to be morally blameworthy. The retributivist thus believes that the sole just end of punishment is to make the morally blameworthy suffer the sanctions we call punishment.

Contasting theories of punishment are the utilitarian and the rehabilitative theories of punishment. Typically, the utilitarian regards punishment as an evil but justifies it by the achievement of the greater good of crime prevention; punishment achieves this greater good through deterrence or incapacitation of would-be criminals (Bentham). Rehabilitationists are often merely utilitarians with a kinder, gentler crime prevention program, reform and education substituting for deterrence and incapacitation; as a truly distinct theory, however, the rehabilitationist regards crime as a disease that is not the fault of the criminal and punishment becomes a cure justified by duties of distributive justice to those unfortunate enough to suffer this disease (Menninger).

In addition to these three dominant theories justifying punishment institutions, there are also "mixed theories" that combine two of these three theories (Moore, 1984). The most widely embraced mixed theory holds that punishment must achieve both the utilitarian goal of crime prevention and the retributive goal of punishing those who deserve it in order to be justified (von Hirsch). Because the desert of the offender is a necessary condition of a just punishment under the mixed theory, this theory is sometimes called "weak" or "negative" retributivism (Mackie). These are misnomers because what is essential to retributivism is that the desert of the offender to be a sufficient reason to punish; a theory that regards such desert as only necessary leaves out what is so distinctive (and so troubling to many) about retributivism.

As a "theory of punishment," retributivism is said to answer the question, Why punish anyone? The question is best interpreted to be a very general one, asking after the justification for the entire criminal law and the institutions that serve it. Retributivism is thus, first and foremost, a theory about the legitimate end served by penal institutions. Retributivism, like other theories of punishment, is a theory about why we should have the criminal law (Moore, 1997). As such, retributivism also purports to answer more discrete questions about criminal law, such as questions about the correct doctrinal triggers for liability and questions about how much offenders should be punished for certain crimes when done with certain levels of culpability. Retributivism also has strong implications for the question of what should be prohibited by penal law; with certain suitable assumptions, a retributivist theory of punishment yields the legal moralist theory of criminal legislation according to which all and only morally wrongful behavior should be forbidden by the criminal law (Moore, 1997). On this latter theory, if a certain sort of behavior is morally wrong, that is a prima facie reason to criminalize it (although other factors may ultimately bar criminalization); if behavior is not morally wrong that is a very good reason not to criminalize it (for no retributive justice is achieved by the punishment of those who do good or at least do no wrong).

There have been famous attempts to restrict the range of questions properly answerable by a retributivist theory. John Rawls (1955) and H. L. A. Hart (1968) urged that retributivism answered the judge's question, Why should this offender be punished? but that the more general questions of why we punish anyone or what should be made criminal were to be answered on the basis of another theory, utilitarianism. The problem for this view was that no good reason can be given for restricting the range of properly answerable questions for either theory. If utilitarianism is a good theory of why and what we should punish, why is it not similarly a good theory of how punishment should be distributed in particular cases? If retributivism is a good theory for why some judge should punish poor old Jones, why is it not an equally good theory for why we should punish anyone like Jones in the relevant respects (namely, equally deserving of punishment)? The fact is that the questions answered by a theory of punishment like retributivism refuse to be cabined in this artificial way. Retributivism is both a general theory of punishment and also a theory about all the more discrete questions about the criminal law, right down to the question of whether and how much each particular offender should be punished.

For a variety of reasons retributivism has probably been the least understood of the various theories of punishment. Part of the bafflement about retributivism stems from its stark simplicity: it essentially asserts that we should punish because and only because culpable wrongdoers deserve it. This simplicity has lead many to seek to divine some other good that giving just deserts causes, inasmuch as this seems to make the theory more satisfyingly complex. For example, it is said that if the state punishes those who deserve it, that prevents vengeful citizens from taking the law into their own hands. This prevention of vigilante justice is then presented as the real good of the retributivism (Marshall). Alternatively, it is urged that if the state punishes those who deserve it, this will satisfy citizens' need for feeling that they are not dupes in restraining themselves from criminal actions. Such punishment thus serves the goods of satisfying a widespread preference that the guilty be punished and it keeps down dissatisfactions (about unpunished free riders) arising among the lawabiding (Stephens).

Retributivism is inconsistent with all of these theories. Essential to retributivism is the thesis that punishing those who deserve it is an intrinsic good, that is, something good in itself and not good because it causes something else. All of these theories make punishment of the guilty merely instrumentally good, that is, good only because it causes: the diminishment of vigilante violence; the satisfaction of citizen preferences for punishment; the maintenance of a sense of social cohesion; or the prevention of crime by those not angered at the sight of others breaking the social contract with impunity. One cannot hold that punishing the guilty is good only because doing so produces these other goods, and still remain a retributivist. Retributivism is a much simpler theory: punishment is justified by the simple moral fact that culpable wrongdoers deserve it.

Retributivism is also sometimes confused with a family of theories that urge that punishment of the guilty is justified because it expresses society's moral outrage at what was done, because it denounces the crime and the criminal or because it communicates to the criminal society's disapproval (Feinberg; Hampton in Murphy and Hampton). On their face, these theories are doubly puzzling: first, why is it plausible to hold that expressing, denouncing, or communicating is much of a good, indeed, so good that the catharsis achieved could justify something as harmful as punishment administered to the offenders? Second, if these are plausible goods, their justification remains unclear since we could as easily undertake dramatic shaming ceremonies where the message is given but the harsh treatment and suffering of offenders is not (Feinberg). In any case, to the extent that these implausible theories reduce the punishment of the guilty to a mere instrumental good in the service of social expression, they are not to be equated with retributivism.

Closer to retributivism are theories that turn punishment into an instrument of victim revenge. Such theories urge that the desert of offenders gives the state the right to punish them, but it is only the desire of the victims for revenge that justifies the state in doing what it has a right to do, namely, punish the guilty. Often adherents of these theories urge that it takes very little to justify the state in punishing those who deserve it, and the slightest satisfaction provided the victim of a crime by punishment of her wrongdoer is reason enough (Murphy, 1990).

Such a view of punishment as an engine for victim-directed vengeance is not utilitarian, but neither is it retributive. A retributivist believes that justice is served by punishing the guilty and thus, the desert of an offender not only gives the state the right to punish him but also the duty to do so. Making victims feel good is no part of retributive justice, although the retributivist may regard it as a welcome side effect of punishment along with crime prevention. Retributive justice is achieved by punishing the guilty even if the victims of such guilty offenders all wish forgiveness and mercy upon their offenders.

How is retributivism to be justified?

Once the essential simplicity of retributivism is fully grasped, it may seem that little can be done in order to justify belief in it (Bedau). Indeed, if justification of some practice as right must consist in showing how the practice produces good consequences, the retributivism cannot be justified (Benn and Peters). In this retributivism and utilitarianism are ultimately alike, for once the utilitarian has shown how a practice maximizes utility there can be no further good that utility causes that makes utility good. It is essential to both theories that retribution and utility be intrinsic goods, and the moment one reduces them to instrumental goodsgoods only because productive of other states of affairsthen the resulting theory can no longer be retributive or utilitarian.

It is thus no criticism of retributivism that it cannot be justified instrumentally, for ultimately all ethical theories are in the same boat; all ethical theories must hold something to be intrinsically good. Retributivism differs from utilitarianism in this respect only in what it regards as intrinsically good.

Fortunately none of this consigns either the utilitarian or the retributivist to some kind of irrationalism about their ethics. Fortunately instrumental modes of justification do not exhaust the possibilities with respect to justifying a moral principle. The trick is to show how retribution is good without relying on the consequences of retribution to make it good.

It is helpful to step away from retributive justice for a moment to see the possibilities here. Consider John Rawls's deservedly famous justification of his two principles of distributive justice (1971). Rawls urged that his two principles of justice were justified in two distinct ways. First, the principles are the best general expression of the mass of our "considered judgments" on more particular matters, such as the irrelevance of inherited wealth for distribution of benefits. Second, the principles are the ones that follow from other more general principles, what Rawls called principles of fairness. Neither of these modes of justification reduced distributive justice for Rawls into a merely instrumental good; this, because neither mode of justification relies on the capacity of distributive justice to produce some other states of affairs. Rather, the distributive principles are shown to be just in that they describe at a general level more particular considered judgments about what justice requires and in that they are the principles that would be chosen in a fair procedure.

Both of these modes of justification are open to the retributivist. Retributivists have long sought to rely on considered judgments in particular cases about what justice requires in order to show how only the retributivist principle can explain such judgments. Imagine, Kant urged, an island society about to dissolve with one vicious murderer yet unpunished. The dissolution of the society removes all of the obvious utilitarian reasons to punish, so that if the facts that make the killer so depraved move one to judge that he should be punished, such judgment can only be explained on retributivist grounds.

A variety of objections have surfaced about this mode of justifying retributivism (Dolinko). The most serious is the charge that our considered judgments in such cases are contaminated by irrational and unsavory emotions. Nietzsche cataloged these as the emotions of fear, resentment, cowardice, envy, sadism, and so on, all lumped under the French term ressentiment (Nietzsche). The objection is that such emotional origins of our retributive judgments make such judgments nonvirtuous to hold and thus unreliable epistemically, whereas our particular judgments of distributive justice can be relied on because stemming from the virtuous emotions of compassion, empathy, and fellow-feeling.

Retributivists have made two responses to this objection. About our third-person judgments about others' wrongful behavior, retributivists have sought to distinguish virtuous emotions of "moral hatred"the hatred anyone should feel if she identifies both with other people who are victims and with morality itselffrom the nonvirtuous emotions of ressentiment that Nietzsche described (Murphy and Hampton; Pillsbury). Our particular judgments about what punishment others deserve are then said to stem from these more virtuous emotions and thus to be reliable epistemically. Similarly, retributivists have sought to show how our first-person judgments about how we should be treated if we have culpably hurt another stem from the virtuous emotion of guilt and thus can be relied on in justifying retributivism (Moore, 1997).

Continuing the analogy to Rawls, the second mode of justification proceeds by showing how the retributive principle is part and parcel of some yet more general principle of justice or fairness. The history of retributivist literature is not reassuring on this matter, for much of what has been said of a general sort is too metaphorical to provide any real grounding for retributivism. Thus, retributivists have argued that unless the guilty receive their due: the blood guilt of the whole people will not be expiated; the moral order will not be restored; the offender's debt to society will remain unpaid; the control of the victim by the crime or the criminal will not be released; and so forth (Fletcher). If one presses these metaphors of blood guilt, moral orders out of whack, debt, control, and the like, they cash out to no more than that culpable wrongdoers deserve to be punished (Moore, 1999). They thus give no more support for that essential retributivist tenet than it gives itself.

The dominant contemporary exception among this mortuary of dead metaphors about retributivism is the attempt to justify the retributivist principle by reference to a general principle of fair play and unjust enrichment. Herbert Morris and others (Scher) have long argued that the prohibitions of the criminal law set a minimum of what is required for a scheme of social cooperation, that criminals unfairly appropriate to themselves the benefits both of their violation and of others' restraint, and that fairness demands some price (in the form of punishment) to be paid for this unjust enrichment. The aim is to show how the retributivist principle follows from a yet more general principle of fairness.

Whether either of these modes of justifying retributivism has succeeded, or can succeed, is not something on which there is any consensus. Issue had been joined at all points between retributivists and their critics, and it is safe to say that no resolution will be seen in the foreseeable future. What is surprising is how much retributivism has made a comeback among contemporary criminal law theorists, political philosophers, and lawmakers. For most of the twentieth century the standard educated view was that retributivism was moribund, but in the last three decades of that century the theory came very much alive.

The institutional implications of retributivism

Retributivism is often associated in the minds of many people with particular institutional arrangements, like the principle of lex talionis (an eye for an eye and a tooth for a tooth), or even more particularly, with the death penalty. While there certainly have been famous retributivists who favored such institutions, nothing essential to retributivism requires them. A retributivist is committed to the idea that the punishment must "fit the crime" in the sense of being proportionate to the degree of wrong done and to the culpability with which it is done; but he need not interpret such principle of proportionate punishment to require that a harm be visited on the criminal just like the harm he visited onto his victim. In addition, a retributivist can easily believe that no one deserves to be put to death at the hands of the state and thus be an opponent, not a proponent, of the death penalty.

Despite this, retributivism does give rise to a distinctive mode of justifying particular institutions within the criminal law such as the death penalty. The retributivist will urge, for example, that the debate about whether the death penalty deters crime is simply irrelevant to the question of whether the death penalty is justified. The only relevant question, according to retributivism, is whether one person can be so morally reprehensible as to deserve to die. If that question is answered affirmatively, then prima facie the death penalty is justified (because the state has the duty to give offenders what they deserve, according to retributivism).

Retributivism has similar implications for other institutions within the criminal law. Indeed, on every doctrinal issue, whether going to liability or to sanction, the retributivist will urge that the relevant policy is retributive, asking in each instance what such a class of persons deserve. Should intentional but nondeliberative killers be punished less than those who premeditate and deliberate about their killings? The relevant question, according to retributivism, is whether this factor makes such a difference in moral blameworthiness that intentional killers should be segregated into two classes (or "degrees") of murderers by this factor alone. Should those who kill under threats of serious harm to themselves be either wholly or partially excused? The relevant question, according to the retributivist, is the degree of moral courage we can fairly expect from people in those coercive circumstances; if we should each die rather than kill an innocent, then those whose behavior fails to live up to that standard are morally blameworthy and deserve punishment in proportion to the degree of their blameworthiness.

Much more determinate institutional implications than these can of course be spun out of retributive theory, but what those implications are depends on the kind of theory of moral responsibility that is accepted by the retributivist. Retributivists differ considerably among themselves on this issue, which is usually termed the "desert-basis" issue. There are three leading views on the touchstone of moral blameworthiness. One view is that there are two ingredients determining our moral blameworthiness, the kind of wrong we do and the culpable mental state in which we do it. On this view, the worse the consequences we bring about by our actions and the less the justification for bringing about such consequences, the more wrongful our actions. The more wrongful the action we either intend, foresee, or risk doing, and the less excuse we have for choosing to act nonetheless, the more culpable we are. The two togetherwrongdoing and culpabilityjointly determine an offender's overall moral blameworthiness.

A second view restricts blameworthiness to culpability alone. On this view what determines our blameworthiness is the degree of wrong we think we are doing in our own mind, not whether we actually succeed in doing such a wrong in the real world. Those who without justification or excuse shoot at another, trying to kill him, are as morally blameworthy if they miss as if they hit and kill their victim.

A third view regards character as the ultimate touchstone of moral blameworthiness. Character consists of our long-term traits, like generosity, courage, compassion. On this view what makes us morally blameworthy is bad character. Wrongful actions and culpable intentions usually evidence bad character, so these items are not irrelevant on this view; such items are merely evidential, however, for what is constitutive of blameworthiness is character and character alone.

The actual institutional implications of retributivism depends heavily on which of these three views of moral responsibility the retributivist adopts. In the Anglo-American penal system the first view has by and large predominated, but in the academic literature it has been seriously challenged by the other two views, particularly the second. All such views can properly be fitted within the retributivist theory.

Michael S. Moore

See also Deterrence; Incapacitation; Punishment; Rehabilitation.


Bedau, Hugo. "Retribution and the Theory of Punishment." Journal of Philosophy 75 (1978): 601620.

Benn, S. I., and Peters, R. S. Social Principles and the Democratic State. London: Allen and Unwin, 1959.

Bentham, Jeremy. An Introduction to the Principles of Morals and Legislation (1823). New York: Hafner, 1948.

Dolinko, David. "Some Thoughts About Retributivism." Ethics 101 (1991): 537559.

Feinberg, Joel. "The Expressive Function of Punishment." Doing and Deserving. Edited by Joel Feinberg. Princeton, N.J.: Princeton University Press, 1971.

Fletcher, George. "The Place of the Victim in the Theory of Retribution." Buffalo Criminal Law Review 3 (1999): 5164.

Hart, H. L. A. Punishment and Responsibility. Oxford, U.K.: Oxford University Press, 1968.

Kant, Immanuel. The Metaphysical Elements of Justice. Translated by John Ladd. Indianapolis: Bobbs-Merrill, 1965.

Mackie, J. "Morality and the Retributive Emotions." Criminal Justice Ethics 1 (1982): 310.

Marshall, Thurgood. Concurring in Gregg v. Georgia, 428 U.S. 153 (1976).

Menninger, Karl. The Crime of Punishment. New York: Viking Press, 1968.

Moore, Michael S. Law and Psychiatry: Rethinking the Relationship. Cambridge, U.K.: Cambridge University Press, 1984.

. Placing Blame: A General Theory of the Criminal Law. Oxford, U.K.: Clarendon Press, 1997.

. "Victims and Retribution." Buffalo Criminal Law Review 3 (1999): 101125.

Morris, Herbert. "Persons and Punishment." The Monist. 52 (1968): 475501.

Murphy, Jeffrie. "Getting Even: The Role of the Victim." Social Philosophy and Policy 7 (1990): 209225.

Murphy, Jeffrie, and Hampton, Jean. Forgiveness and Mercy. Cambridge, U.K.: Cambridge University Press, 1988.

Nietzsche, Friedrich. On the Genealogy of Morals (1887). Translated by Walter Kaufmann. New York: Vintage, 1969.

Pillsbury, Samuel. "Emotional Justice: Moralizing the Passions of Criminal Punishment." Cornell Law Review 74 (1989): 655710.

Rawls, John. "Two Concepts of Rules." The Philosophical Review 64 (1955): 332.

. A Theory of Justice. Cambridge, Mass.: Harvard University Press, 1971.

Scher, George. Desert. Princeton: Princeton University Press, 1987.

Stephens, Sir James. Liberty, Equality, Fraternity. Cambridge, U.K.: Cambridge University Press, 1967.

von Hirsch, Andrew. Doing Justice. New York: Hill and Wang, 1976.