3
Justifications of the Practice: Utilitarian and Retributive

In the previous chapter we considered arguments of radical critics who seem to challenge the whole practice of legal punishment. Now we shall consider theorists whose energies are devoted, not to generating in us skepticism of the practice, but to justifying the practice so that we may be at home in it. As in the last chapter, we will have to be sensitive to the difficulty of justifying practices as a whole: it isn't clear how to tell who is right or wrong about either what the justification for a practice is or whether it is in fact justified.

philosophers and theorists have long debated why it is that we punish and what principles are involved when we punish. Two competing accounts—utilitarian and retributive—have dominated the discussion, though neither has prevailed.^[1] There are distinct versions of both utilitarian and retributive accounts, and sometimes the differences among retributivists seem greater than the differences between some utilitarians and some retributivists. We can, however, describe the two accounts in general terms, as follows: the utilitarian argues

[1] Despite what the Supreme Court declared in Williams v. New York , 337 U.S. 241, 248 (1949): "Retribution is no longer the dominant objective of the criminal law."

that we should punish only when doing so would augment social utility; the retributivist objects, saying that we must punish those who do wrong, even if doing so diminishes social utility, because justice demands that we punish. Utilitarians are consequentialists, always forward-looking, insisting that an action or a practice is justified only if its future benefits outweigh its future cost.^[2] Retributivists are not forward-looking in this way. That an action conforms to a principle of right or justice is, for the retributivist, usually sufficient justification for that action.^[3]

Both utilitarians and retributivists claim to give an account of why we punish, of the principle immanent in the existing practice. Both are in a position to be immanent critics: they can use their interpretation of the purpose of and principle immanent in the practice (to augment social utility, to mete out justice) to criticize the actual practice when it diverges from this principle or fails to live up to its purpose. In chapter four, when we step inside the practice and take up, not the justification of the practice as a whole, but particular problems within the practice, we shall see how the theorist's advice to the judge, prosecutor, or sentencing commission member differs depending on the conception the theorist has of why we punish at all. The debate between utilitarians and retributivists matters practically. In chapter four I shall defend a version of retribution. The task of this chapter is to understand both the different retributive accounts of why we punish—this will allow me to distinguish the version I shall defend from others which I find unpersuasive but which most people probably have in mind when thinking about retribution—and various utilitarian versions as well, and to see how the different theories

[2] There are utilitarian and nonutilitarian consequentialist theories. For an example of the latter, see Plato, Protagoras 324a–b. Cf. lgor Primoratz, Justifying Legal Punishment (Atlantic Highlands, N.J.: Humanities Press International, 1989), pp. 9–10.

[3] This is not to say that retributivists are never forward-looking. I shall argue that the most persuasive retributive account is forward-looking in some sense: see chapter 3, section 3, and chapter 4, section 4, below.

clash. We shall examine classic and contemporary accounts, first of committed utilitarians, then of committed retributivists.

1. Utilitarians

When used without care, the label "utilitarian" can apply to anyone, even self-declared opponents of utilitarianism. For example, in the previous chapter we saw that for Marx, punishment is an instrument used by the ruling class to defend itself so that it may maintain itself: punishment, that is, is useful in preserving the ruling class. But Marx is no utilitarian. His aim is not to justify the practice by appealing to principles of utility. Nor should we necessarily call utilitarians all those who, by showing its use for society, do mean to justify the practice of legal punishment.^[4] Unless I indicate otherwise, I shall reserve the label "utilitarian" for those who justify the practice of legal punishment by appealing to some principle of utility or some calculation of net utilities. Probably all practices have or once had a use or function. But whether the practice is justified by some principle of utility is another matter.

1.1 Jeremy Bentham

The classic exponent of utilitarianism is Jeremy Bentham, and one of his greatest works lays out a utilitarian justification of legal punishment.^[5] For Bentham, the principle of utility is the ground of all moral actions. It is a natural principle that lacks any further ground,^[6] and it is not to be questioned: "Systems

[4] For example, René Girard or Hans von Hentig—see chapter 2, Section 3. Girard argues that legal punishment is a functional equivalent of ritual sacrifice and serves as an outlet for violence, an outlet necessary to the survival of society. Von Hentig argues that punishment serves as a means of selection and security.

[5] Jeremy Bentham, An Introduction to the Principles of Morals and Legislation (1789; reprinted New York: Hafner Press, 1948). In citing this I refer to Bentham's own chapter/section numbers.

[6] Ibid., ch. 1, section 11.

which attempt to question it deal in sounds instead of sense, in caprice instead of reason, in darkness instead of light."^[7] For Bentham, our privileged guide is

that principle which approves or disapproves of every action whatsoever, according to the tendency which it appears to have to augment or diminish happiness of the party whose interest is in question.^[8]

Bentham believes that human beings all, implicitly or explicitly, consent to this principle, which calls on each of us to calculate the pleasures and pains that result from an action we contemplate taking:

Gross ignorance, they will say, never troubles itself about laws, and passion does not calculate…. [But] men calculate, some with less exactness, indeed, some with more: but all men calculate. I would not say, that even a madman does not calculate.^[9]

Bentham claims that utilitarian calculation underlies not only human actions in general but legal punishment in particular:

The business of government is to promote the happiness of the society, by punishing and rewarding…. In proportion as an act tends to disturb that happiness, in proportion as the tendency of it is pernicious, will be the demand it creates for punishment.^[10]

Bentham maintains that the purpose of punishment is to discourage crimes, which he calls acts of "mischief." A crime produces a "primary mischief," which is sustained by an assignable individual or multitude of individuals, and a "secondary mischief," which is the extension of mischief to unassignable

[7] Ibid., ch. 1, section 1.

[8] Ibid., ch. 1, section 2.

[9] Ibid., ch. 14, section 28.

[10] Ibid., ch. 7, section 1.

individuals or to the whole community.^[11] If I am robbed, I sustain primary mischief, and so do family members who care about me or rely on me for support. Society sustains secondary mischief by my being robbed, because the level of "danger" and "alarm" have been increased.^[12] The "danger" lies in the suggestion to others of the feasibility of robbing. "Alarm" refers to the increased. fear we all suffer from the prospect of being victims of robbery.^[13] It is to prevent such mischiefs that we punish. For Bentham, punishment is "an artificial consequence annexed by political authority to an offensive act."^[14] We punish in order to augment the total happiness of the community by excluding mischief, which tends to subtract from that happiness.^[15]

Punishment is itself a mischief, or evil, since it inflicts pain, and on the principle of utility "it ought only to be admitted in as far as it promises to exclude some greater evil."^[16] Punishment does this by reformation, disablement, and compensation, but maily by "example"—by which Bentham means deterrence.^[17] Compensation, or the providing of a "pleasure or satisfaction to the party injured," is not the primary purpose of punishing, because "no such pleasure is ever produced by punishment as can be equivalent to the pain."^[18] Bentham sees as primary instead the deterrent function of punishment: "Example is the most important end of all, in proportion as the number of the persons under temptation to offend is to one."^[19] Bentham often appeals to the deterrent effects of punishing, for example, in justifying the practice of not punishing retroactively: we cannot deter by punishing someone for an act

[11] Ibid., ch. 12, section 3.

[12] Ibid., ch. 12, section 5.

[13] Ibid., ch. 12, section 8.

[14] Ibid., ch. 12, section 36.

[15] Ibid., ch. 13, section 1.

[16] Ibid., ch. 13, section 2.

[17] Ibid., ch. 13, section 2; ch. 15, section 14.

[18] Ibid., ch. 13, section 2.

[19] Ibid., ch. 13, section 2, note.

he could not have known was mischievous.^[20] Similarly, punishing infants or the insane or intoxicated is not warranted, for they could not be deterred.^[21]

Bentham, then, gives what is essentially a deterrence-based justification of the infliction of punishment—we inflict punishment to deter future mischief—that is premised on the more general claim that mischief detracts from our happiness, and the increase of happiness should be the ultimate end of all ethical action.^[22] Bentham justifies punishment by showing, not that it servesjustice, but that it promotes the good.^[23]

Once he has established that the purpose of punishment is to yield the good by excluding mischief—which is painful and therefore evil—Bentham argues that we should employ punishment in particular cases only when it lives up to this purpose; he uses his account of the principle immanent in the practice to criticize the actual practice when it diverges from the principle. In Bentham's view, we should punish only when the principle of utility warrants punishment. Therefore, he argues, we should not punish where doing so would be groundless for want of mischief to deter; nor where punishing is inefficacious; nor, as we saw above, "where it cannot act so as to prevent the mischief"; nor where punishing is "unprofitable" or "too expensive"; nor where we could stop the mischief in some other, cheaper way.^[24]

[20] Cf. ibid., ch. 13, section 7.

[21] Ibid., ch. 13, section 9. H. L. A. Hart has sharply criticized this point: "Plainly it is possible that the actual infliction of punishment on the insane or children may deter normal persons" ("Prolegomenon to the Principles of Punishment," in Stanley Grupp, ed., Theories of Punishment [Bloomington: Indiana University Press, 1971], p. 369).

[22] Bentham defines ethics as "the art of directing men's actions to the production of the greatest possible quantity of happiness" (Introduction , ch. 17, section 2).

[23] "Now, pleasure is in itself a good; nay, even setting aside immunity from pain, the only good; pain is in itself an evil; and, indeed, without exception, the only evil; or else the words good and evil have no meaning" (ibid., ch. 10, section 10).

[24] Ibid., ch. 13, section 3. Bentham infers so many "cases unmeet for punishment" that he runs the risk of undermining his own argument justifying punishment for its deterrent effect. Most modern deterrence theorists emphasize that if punishment is to deter, the potential criminal must be reasonably certain that his crime will be met with punishment. Bentham would have us factor into our calculation so many variables that no person could know for certain whether in the end the action he weighs would be deemed punishable.

Bentham also uses his principle of utility to formulate rules for how we should go about punishing. For example:

The value of the punishment must not be less in any case than what is sufficient to outweigh that of the profit of the offence.^[25]

The greater the mischief of the offence, the greater is the expense, which it may be worth while to be at, in the way of punishment.^[26]

Bentham gives utilitarian accounts of aspects of legal punishment usually justified by retributive principles. For example, he says that punishment should share the characteristic of the offense, for in this way it is an analogy and will be efficacious.^[27] Retaliation, therefore, "in the few cases in which it is practicable, and not too expensive, will have one great advantage over every other mode of punishment."^[28] Bentham thus gives something approaching a utilitarian justification of the lex talionis. Bentham also finds a utilitarian ground for another retributive principle—that we punish in order to express society's moral disapproval of crimes. He suggests that by expressing reprobation for a crime—by using "solemnities"—we can increase the apparent magnitude, without needlessly increasing the cost (level of mischief) of the punishment.^[29] Bentham argues, not, as do some retributivists, that we punish in order to condemn; but, rather, that by punishing in a way that expresses

[25] Ibid., ch. 14, section 8.

[26] Ibid., ch. 14, section 10.

[27] Ibid., ch. 15, section 7.

[28] Ibid., ch. 15, section 8.

[29] Ibid., ch. 15, section 9.

condemnation we can achieve our purpose—more total pleasure and less total pain—at a lower cost.

The retributivist may object to Bentham's justificatory project by claiming that Bentham does not really justify punishment—he does not show the justice of the practice; rather, he gives reasons why punishment is good. The retributivist might claim that questions of the good are separate from questions of justice or of what is right, and that before we decide how to obtain what is good we must know how justice or right limits what actions we might take to obtain the good.^[30] Some retributivists infer from the utilitarian's emphasis on the good that the utilitarian ignores questions of right, that in principle he justifies the manifest in justice of punishing the innocent if doing so would promote social utility.^[31]

Bentham does not argue that we should punish an innocent person, even if doing this would augment the total happiness of the community.^[32] But he does argue that we should not punish in cases where this would be inefficacious, or unprofitable, or too expensive,^[33] and some retributivists would reply that justice demands that we punish even in such cases. The dispute will remain obscured unless we make the distinction between demanding we punish only for an offense (a "negative retributive principle") and demanding we always punish for an offense (a "positive retributive principle").^[34] At one point

[30] A major topic of contemporary moral philosophy is the question of whether the right is prior to the good. Michael Sandel challenges what he takes to be the claim of John Rawls that justice should have absolute priority over all particular conceptions of the good, by arguing that we give up the politics of rights for a politics of the common good. Sandel describes Rawls as a "deontological liberal" in holding to the moral priority of justice (Sandel, Liberalism and the Limits of Justice [Cambridge: Cambridge University Press, 1982], esp. Introduction and ch. 1; cf. John Rawls, A Theory of Justice [Cambridge, Mass.: Harvard University Press, 1971], esp. chs. 1–2.

[31] See chapter 5, section 4.

[32] Bentham says explicitly, if not emphatically, that we punish only in response to "an offensive act" (Introduction , ch. 12, section 36; cf. ch. 15, section 25, and ch. 13, section 3).

[33] Ibid., ch. 13.

[34] Recently such a distinction has been suggested by J. L. Mackie, Persons and Values (Oxford: Clarendon Press, 1985), pp. 207—8. Cf. C. L. Ten, "Positive Retributivism," Social Philosophy and Policy , vol. 7, no. 2 (Spring 1990), pp. 194–208.

Bentham appears implicitly to assent to the positive retributive principle that we must punish for an offense. Bentham writes that the deterrent effect of punishment

depends altogether upon the expectation it raises of similar punishment, in future cases of similar delinquency. But this future punishment, it is evident, must always depend upon detection. If then the want of detection is such as must in general appear too improbable to be reckoned upon, the punishment, though it should be inflicted , may come to be of no use.^[35]

Here Bentham argues that punishment would be useless, and therefore by his own principle ought not to be inflicted, in cases in which punishing would not deter future mischief. However, Bentham, in reflecting on such cases, writes, not that we ought not to punish, but, rather, that punishing would be of no use "though it should be inflicted." This might indicate that he recognizes implicitly that there is some ethical demand (based on utility!) for punishing offenses, regardless of the bearing on utility of actually inflicting punishment in these cases. Of course the "should" here is ambiguous in the English of Bentham's time. It may mean "even were it to be inflicted" rather than "it ought to be inflicted," and so I do not think we should make too much of this point. In one other passage, referring to the accidental punishment of a person innocent of an offense, Bentham uses the phrase "justly punished" to describe the punishment deserved by someone guilty of an offense,^[36] thus implicitly acknowledging the negative retributive principle that we may punish only those who commit offenses—justice demands this.

Bentham's position seems to be this: we may justly punish only those guilty of a crime. But though some will say justice demands that we punish all who commit crimes, in some cases

[35] Bentham, Introduction , ch. 17, section 13, my emphasis.

[36] Ibid., ch. 15, section 25.

the principle of utility dictates that we should refrain from carrying out the act of punishment.

Bentham did allow that "lots of punishment" are "variable."^[37] Perhaps Bentham's response to the retributivist objection that it is unjust not to punish someone who commits an offense would be to say that we should always punish, but where the punishment we would ordinarily prescribe for a particular crime would be too great a mischief to be justified by utility, we should adjust the punishment to a lower level of mischief, thereby satisfying the demands both of retribution and of utility. The problem with this argument is that it is not consistent with Bentham's principle of utility, which insists that we punish only when punishment will be efficacious. By insisting on this point, Bentham stands opposed to the positive retributive principle that we must punish all offenses. But Bentham himself implicitly accepts the weaker negative retributive principle that we must punish only for an offense, and there is some evidence (though weak) that he might recognize implicitly even the stronger principle which he explicitly opposes.

1.2 Cesare Beccaria

We have seen a tension in Bentham's account between his utilitarianism and his very tentative and implicit acknowledgment of the demands of right or justice. This tension also surfaces in the work of another classic theorist of punishment, Cesare Beccaria. Beccaria is not properly called a utilitarian, for he does not appeal systematically to some principle of utility, as does Bentham. But in his famous work An Essay on Crimes and Punishments , Beccaria justifies legal punishment, and also its limits, by appealing to the idea of social utility.^[38] Beccaria combines elements of both rights-based and utilitarian

[37] Ibid., ch. 15, section 2.

[38] Cesare Bonesana Beccaria, An Essay on Crimes and Punishments (Philadelphia: William P. Farrand and Co., 1809). Beccaria's book was important for its plea for humanity in punishment and for its opposition to the death penalty.

theories. Like Hobbes, Beccaria argues that the natural condition of man is a continual state of war, and to escape it we sacrifice part of our liberty in order to enjoy the rest in peace and security; this reservoir of liberty has to be defended, and punishment is the means. Punishment is necessary to restrain passions and preserve our lives.^[39] Armed with this principle that we punish to preserve the safety of society, Beccaria goes on to ask of all occasions on which we consider whether to punish, whether punishing is reay useful or necessary for the safety or good order of society.^[40] As for Bentham, for Beccaria punishment is an evil, and we are to use it only when the principle of social utility dictates that we should:

The degree of the punishment, and the consequence of a crime, ought to be so contrived as to have the greatest possible effect on others, with the least possible pain to the delinquent—for mankind, by their union, originally intended to subject themselves to the least evils possible.^[41]

Beccaria gives essentially a deterrent theory of punishment: the intent of punishment is not to torment or to undo past crime, but to deter future injury to society, and punishment ought to be chosen to maximize its deterrent effeect.^[42]

But Beccaria's utilitarianism is combined with his rights-based social-contract theory. Beccaria argues that men give up some of their natural liberty and agree to obey the laws and be punished for violating them, only because doing so is necessary to prevent a state of war.^[43] Beccaria objects to the death penalty by appealing to his rights-based theory: the sovereign has no right to impose the death penalty, because we never gave to others the right of taking away our lives—in agreeing to the social contract, we each sacrificed only a small

[39] Ibid., chs. 1, 4.

[40] Ibid., ch. 11.

[41] Ibid., ch. 19.

[42] Ibid., ch. 12.

[43] Ibid., ch. 2.

portion of liberty to the public good. Beccaria, however, also appeals to principles of utility in opposing the death penalty. arguing, for example, that the death penalty does not deter.^[44] Hobbes had similarly said that we can never be understood to lay down our right to defend our lives or to resist being wounded, chained, or imprisoned.^[45] Hobbes, however, does not argue that the sovereign has no right to punish; his sovereign may do what he pleases, even punish an innocent person without this being an injustice.^[46] Hobbes argues, rather, that the sovereign may with right execute me, but I retain the right to try to escape. Unlike Hobbes, Beccaria is not content with declaring the right of the convicted person to try to flee from the executioner, leaving the prisoner a corpse with rights. Beccaria, as we have seen, gives other reasons in arguing against state execution.

Both Bentham and Beccaria justify legal punishment by appealing to some principle of social utility. Both recognize, implicitly or explicitly, limits to their justificatory principle. Bentham implicitly acknowledges that we may justly punish only for a mischievous offense; and Beccaria argues that we may punish only in a way that does not violate the individual rights we carry with us into civil society.

1.3 Richard Posner

The legacy of Bentham and Beccaria has carried over to the present.^[47] It is fair to say that utilitarianism is favored among

[44] Ibid., ch. 28. Of course, Beccaria does not imply that the murderer should not be punished by some other form of punishment. Whether we punish is a separate question from how we punish—a point on which I elaborate in the next chapter.

[45] Thomas Hobbes, Leviathan , ed. Michael Oakeshott (New York: Collier Books, 1962), ch. 21.

[46] Ibid., ch. 14.

[47] One of the leading utilitarian (in a loose sense) accounts of legal punishment is given by James Q. Wilson, Thinking About Crime (rev. ed. New York: Random House, 1985). Wilson opposes the "sociological approach" that tries to trace the causes of crime; he thinks we ought to deter a criminal by increasing the costs and reducing the benefits of crime. Wilson thinks individuals choose the path of crime, or at least that we should assume this as a matter of policy (pp. 45–51. In Wilson's view, crime will decline if it becomes less profitable compared to other ways of spending one's time: Would-be offenders are reasonably rational and respond to their perception of the costs and benefits attached to alternative courses of action" (p. 118).

the competing theories and justifications of legal punishment. In this section I shall take as a contemporary example of this legacy the work of Richard Posner, even though for a technical reason Posner distinguishes his account from utilitarianism.^[48] Posner takes utilitarianism to the extreme in accounting for many aspects of the practice of legal punishment. His argument is interesting and highly controversial, and because of its extreme character offers us an instructive utilitarian foil to retributive justifications. Also, Posner, with the "law and economics movement" of which he is a leading advocate, has been an influential voice in public policy debates, and so consideration of his argument is particularly appropriate, given our concern ultimately with the application of theory to practice.^[49]

[48] Richard Posner, "An Economic Theory of the Criminal Law," Columbia Law Review , vol. 85, no. 6 (October 1985), pp. 11 93–1231. The technical reason Posner distinguishes his view from utilitarianism has to do with the difference between the economic concept of value and the moral concept of utility. In Posner's view, we punish to further economic efficiency, and so my stealing a car is wrong because it is an inefficient method of allocating resources or because it moves resources from a more to a less valuable employment. Posner defines value as "a function of willingness to pay." "Since I am unwilling (because unable—but it does not matter why) to pay my neighbor's price for the car, it follows that the car would be less valuable in an economic sense in my hands than in his." Posner adds, "The car might, of course, confer more utility (pleasure, satisfaction) on me than on my neighbor, but there is a difference between utility in a broad utilitarian sense and value in a (perhaps narrow) economic sense, where value is measured by willingness to pay for what is not yours already" (1196, and note 9). One objection to Posner's view is that it doesn't take seriously the discrepancy between willingness and ability to pay. In a (broad) utilitarian theory we can allow that a Stradivarius in the hands of a low-income prodigy is a more valuable allocation of resources than a Stradivarius locked in the vault of a rich investor who can't play a note.

[49] See, for example, New York Times , October 10, 1988, on the role of the law and economics movement in the debate about corporate penalties.

Posner argues that the criminal law exists to be functionally efficient by serving the goal of wealth maximization. He gives a comprehensive rationale or theory of the criminal law and criminal law doctrine, explaining how this law and doctrine serve to enhance market efficiency. Posner is not obviously engaged in any justificatory activity, at one point saying that he hasn't enough time to consider competing theories:

lf this Article were not already so long, I would go on and compare the economic approach with its principal rival, the "moral" theory of criminal law, which argues that the criminal law should only punish morrally blameworthy conduct. Whatever the normative merits of this approach, I doubt that it is as good a positive theory of criminal law as the economic, since in so many areas conduct is punished that is not blameworthy in the moral sense.^[50]

Posner argues that

[t]he major function of criminal law in a capitalist society is to prevent people from bypassing the system of voluntary, compensated exchange—the "market," explicit or implicit—in situations where, because transaction costs are low, the market is a more efficient method of allocating resources than forced exchange.… Most of the distinctive doctrines of the criminal law can be explained as if the objective of that law were to promote economic efficiency.^[51]

[50] Posner, "An Economic Theory of the Criminal Law," pp. 1230–31

[51] Ibid., p. 1195. We might think that to say, as does Posner, that our practice can be explained "as if" x were the principle immanent in it is not to say that x is the principle immanent in the practice. But this distinction between "as if" and "is" holds only if we think principles have ontological status. I believe, rather, that to say a principle inheres in a practice is to make an interpretive claim that reflects not the ontology but one's understanding of a practice.

With this principle—that we punish in order to promote market efficiency—Posner accounts for numerous aspects of the criminal justice system. He accounts for why in most cases we require as a condition for punishment that there be a certain state of mind, or intent, of the accused, a requirement he says is "puzzling to the economist."^[52] His answer is that criminal intent identifies "pure coercive transfers," and this is economically important. Where there is intent, we can infer that the criminal invests resources to bring about a wrong, and that the wrong did not emerge accidentally. We should expend the resources needed to punish only where other resources are being expended by the criminal. Otherwise all sorts of serious social costs would be incurred by people avoiding lawful activity in order to prevent the appearance of being engaged in unlawful activity:

If I take from a restaurant an umbrella that I mistakenly think is mine, I am not a thief, if I know the umbrella is not mine and take it anyway, I am. The economic difference is that in the first case I would have to expend resources to avoid taking the umbrella and the probability of my taking the wrong umbrella is low … the risk of overdeterrence through a criminal penalty is great.^[53]

Therefore the requirement of intent is justified on economic grounds. Implicit in Posner's account is Bentham's own rule that we should not punish where doing so is unprofitable or too expensive.

Posner's claim that we punish in order to prevent the bypassing of markets may seem plausible in accounting for why we punish those who steal or those who create monopolies.^[54] But much of our criminal law is concerned with crimes of

[52] Ibid., p. 1221.

[53] Ibid.

[54] Though retributivists would argue that the thief who steals my property has violated my person, has done something that is morally wrong. The retributivist might appeal to the sense of indignation and the feeling of being violated that one experiences when robbed even of something trivial.

violence. Can Posner's principle account for why we punish the murderer or rapist? Posner thinks so.

Crimes of passion often bypass implicit markets—for example, in friendship, love, respect … less obviously, crimes of passion often bypass explicit markets too.… Someone who gets his satisfaction in life from beating up other people, without compensating them, rather than from engaging in trade with them, is thus bypassing explicit markets.^[55]

Posner also claims that the rapist bypasses markets: "if [the rapist] spent his time raping rather than dating women he would be bypassing an implicit market."^[56] Posner's claim is that having laws against murder and rape is more efficient than allowing these "coercive transfers." If we did not punish murderers and rapists and allowed such coercion we

would create incentives for potential victims to spend heavily on self-protection and for potential aggressors to spend heavily on overcoming the victims' self-protective efforts. All this spending would yield little if any net social product.^[57]

Posner realizes how perverse his argument sounds and at least pauses to acknowledge that of course "rape is a bad thing." He says his point is that "economic analysis need not break down in the force of such apparently noneconomic phenomena as rape."^[58] At this point his argument is far removed from not only a justification but also an explanation of some aspects of legal punishment—he gives a rationale which is not in the least bit persuasive as an account of the reasons we punish crimes of violence.

[55] Posner, "An Economic Theory of the Criminal Law," p. 1197. One of Posner's examples "of a lawful market alternative to battery" is professional boxing (p. 1198, note 12).

[56] Ibid., p. 1198.

[57] Ibid.

[58] Ibid., p. 1199.

Posner's account is troubling, especially to the retributivist. Posner sees the criminal justice system as an instrument we can fine-tune to promote economic efficiency. For example, he claims that we adjust the probability of apprehension (for example, by increasing or reducing the number of police) with the length of sentence, in order to approach an efficient use of resources. When we make more arrests and reduce the length of sentence, we trade litigation and pre-trial detention costs for prison costs.^[59] The retributivist objects to the very idea of adjusting the level of arrest—it is unfair that who gets apprehended and punished depends on the decision of policy-makers concerned with economic efficiency. Posner acknowledges this claim that selective enforcement of laws "creates ex post inequality among offenders,"^[60] but he sees nothing wrong with this. The criminal justice system, in his view, is like a lottery, and lotteries aren't unfair:

Nor is it correct that while real lotteries are voluntary the criminal justice "lottery" is not. The criminal justice['sl is voluntary: you keep out of it by not committing crimes.^[61]

But those holding to the positive retributive principle find selective enforcement unacceptable, not merely because it is unfair, but because it is premised on the mistaken view that it is not inherently wrong to allow some crimes to go unpunished.

When we turn to the contemporary utilitarian-retributive debate (chapter five, section 4), we shall see that an important retributive objection to utilitarianism is that in principle the utilitarian, who justifies an action on the basis of whether it augments social utility, is committed to punishing an innocent person if the principle of utility recommends doing so. None of the utilitarians we have discussed justifies the punishing of

[59] Ibid., p. 1213.

[60] Ibid.

[61] Ibid.

an innocent person for the sake of augmenting social utility. Even Bentham implicitly rules this out. But neither Bentham nor Posner accepts the claim that seems fundamental to the retributivist position: that there are such things as wrongs deserving of punishment. For Posner, the reason we call (and ought to call) certain activities wrongs, or crimes, is that they bypass the market and therefore promote economic inefficiencies; for Bentham, the reason is that these activities, or mischiefs, take away from society's total happiness. Each holds that the reason we have the whole practice of legal punishment is to augment social utility, however understood; each is therefore willing in particular cases to say "don't punish," if the principle each sees underlying the practice as a whole determines that we should not punish. Each explicitly rejects the positive retributive principle that we must always punish for an offense, a principle to which many retributivists hold. We might think that one is not a retributivist unless one holds to this principle. Our task now is to see just what it means to be a retributivist.

2. Retributive Justifications of Legal Punishment

There is a great deal of confusion about precisely what it is to be a retributivist. One of today's leading retributivists himself does not think "retribution" is a helpful term: Andrew von Hirsch notes that the O.E.D. definition of "retribution" as "return of evil" and the declaration in the 1972 Model Sentencing Act that "sentencing should not be based upon revenge and retribution" illustrate how, retribution is often confused with vindictiveness.^[62] The report of the Royal Commission on Capital Punishment clarifies how retribution can be understood differently:

Discussion of the principle of retribution is apt to be confused because the word is not always used in the same

[62] Andrew von Hirsch, Doing Justice (Westford, Mass.: Northeastern University Press, 1986), pp. 45–46.

sense. Sometimes it is intended to mean vengeance, sometimes reprobation. In the first sense the idea is that of satisfaction by the State of a wronged individual's desire to be avenged; in the second it is that of the State's marking its disapproval of the breaking of its laws by a punishment, proportionate to the gravity of the offense.^[63]

Ambiguity exists not only in the meaning of "retribution" but in the burdens of a retributive justification of punishment. Some think that the amount of punishment need not be a just amount for the punishment to be justified on retributive grounds.^[64] Others argue that any deviation from the just amount of punishment is unjust and therefore state punishment which can never hope to mete out punishment with such precision, cannot be justified by an appeal to the principle of retribution.^[65] From our previous discussion we know of other possible criteria for being a retributivist. To be a retributivist, must I hold that we must punish all wrongs (positive principle), or merely that we must never punish an innocent person (negative principle)? To be a retributivist, must I hold that we punish only morally blameworthy conduct? Our task in this section is to answer these questions by examining various accounts of punishment commonly regarded as retributive.

We shall consider four sorts of retributive accounts, which I shall refer to as (1) revenge (associated by some with the lex talionis ); (2) condemnation; (3) deontological theories; and (4) just deserts.

[63] Report of the Royal Commission on Capital Punishment (London: H.M.S.O., 1953), pp. 17–18, cited in Joel Feinberg, "The Expressive Function of Punishment," in his Doing and Deserving: Essays in the Theory of Responsibility (Princeton: Princeton University Press, 1970), p. 101.

[64] C. W. K. Mundle, "Punishment and Desert," in Grupp, ed., Theories of Punishment , p. 66.

[65] A. C. Ewing, The Morality of Punishment (London: Kegan Paul, 1929), cited in Mundle, "Punishment and Desert," pp. 66–67.

2.1 Revenge

Oliver Wendell Holmes wrote that retribution is "only vengeance in disguise."^[66] Most of us, when we hear someone declare that we ought to punish for the sake of retribution, associate retribution with revenge or retaliation (which we wrongly conflate).^[67] We think of the biblical expression of the lex talionis :

And if any mischief follow, then thou shalt give life for life, eye for eye, tooth for tooth.^[68]

Breach for breach, eye for eye, tooth for tooth: as he hath caused a blemish in a man, so shall it be done to him again.^[69]

We have been taught to resist the urge to retaliate:

Ye have heard that it hath been said, An eye for an eye, and a tooth for a tooth: But I say unto you, That ye resist not evil: but whosoever shall smite thee on thy right cheek, turn to him the other also.^[70]

Christianity has denounced retaliation, and, consequently, in the minds of many, retributive accounts of punishment. Of course, many of us do not easily heed Christ's words. We hardly need the survey research of social scientists (though

[66] Oliver Wendell Holmes, The Common Law (Boston: Little, Brown, 1923), p. 45.

[67] The justice of retaliation, or lex talionis , was deeply entrenched in the moral sensibilities of the Greeks and other archaic societies: see Gregory Vlastos, "Socrates' Contribution to the Greek Sense of Justice," Archaiognosia , vol. 1, no. 2 (1980), pp. 304ff. It was thought of as repayment, and the metaphor of paying back a debt was often used to characterize this sense of justice. Revenge is different, measuring punishment by the feelings of the victim.

[68] Exodus 21:24.

[69] Leviticus 25:20.

[70] Matthew 5:38–39.

the data are available) to tell us that revenge is an urge deeply seated in us.^[71]

The law of the talio , or of retaliation, is not necessarily connected to the idea of revenge. To see punishment as revenge is to focus on the motivations of the punisher, whereas the lex talionis is a law of equivalence that dictates what punishment is commensurate with the crime; it is not a theory of motivation. Both the talio and the view that we punish to avenge have been discredited by most modern retributivists.

Retributivists needn't commit to the lex talionis . One of the most famous and important retributivists, Hegel, is sharply critical of the lex talionis , using Blackstone's example to make his point: "an eye for an eye, a tooth for a tooth—and then you can go on to suppose that the criminal has only one eye or no teeth."^[72] In any case, what amount of punishment we inflict is one question; why we punish is another. No retributivist of repute takes the lex talionis as the justification for punishing at all. One modern retributivist, Joel Feinberg, calls "incoherent" the version of retributive theory that insists "that the ultimate justifying purpose of punishment is to match off moral gravity and pain, to give each offender exactly that amount of pain the evil of his offense calls for, on the alleged principle of justice that the wicked should suffer pain in exact proportion to their turpitude."^[73]

Nor need retributivists commit to the idea that we punish to avenge. The most persuasive retributivists distinguish their

[71] See, for example, the very dated but still interesting study by F. C. Sharp and M. C. Otto, "A Study of the Popular Attitude Towards Retributive Punishment," International Journal of Ethics , vol. 20, no. 3 (April 1910). The authors conclude that revenge is deeply built into the values of those surveyed, "contrary to traditional Christian ethics."

[72] G. W. F. Hegel, Philosophy of Right , trans. T. M. Knox (1821; London: Oxford University Press, 1952), par. 101, Remark, p. 72. Hegel opposes, not the idea that the severity of the punishment we inflict should be equivalent in value to the severity of the crime, but only the strict equivalence established by the lex talionis .

[73] Feinberg, Doing and Deserving , p. 116.

view from revenge theories of punishment. For Hegel, the judge who oversees legal punishment is not an avenger. Whereas revenge can be arbitrary and further the wrong, the judge of a rational modern state must be "cold, heartless, and have only the interests of the law," and this presupposes the education or cultivation (Bildung ) of a modern state.^[74] says that the word Gerechtigkeit (justice) comes from the word Rache (revenge) and that in uncivilized (ungebildeten ) states justice is revenge,^[75] but in rational modern states revenge is too contingent and arbitrary and subjective to serve justice or right.^[76] In an earlier work Hegel explains that if we rely on the sufferer or his next of kin to punish, then right is mixed with arbitrariness; legal punishment depends, rather, on a third party.^[77]

Revenge, as Hegel describes it, is subjective; it derives from feelings of anger and resentment within an individual. Not all retributivists, however, take anger to be a subjective measure residing only within individuals. Walter Berns defends punishment as an expression of anger, but the anger he means resides not merely within the hurt victim but within society. The anger Berns thinks punishment expresses is a righteous anger, an anger "somehow connected with justice." For Berns, this anger is not "a selfish indulgence," but "may more accurately be called a profound caring for others."^[78]

[74] G. W. F. Hegel, Vorlesungen über Rechtsphilosophie (1818–1831), 4 vols., ed. Karl-Heinz Ilting (Suttgart-Bad Cansatt: Friedrich Fromman, 1973), vol. 4, p. 556.

[75] Ibid., vol. 4, p. 294.

[76] G. W. F. Hegel, Grundlinien der Philosophie des Rechts , in Hegel, Werke in zwanzig Bänden , ed. Eva Moldenhauer and Karl Michel (Frankfurt am Main: Suhrkamp, 1970), vol. 7, par. 102.

[77] G. W. F. Hegel, Philosophische Propaedeutik , in Hegel, Werke , vol. 4, part 1, par. 21.

[78] Walter Berns, "The Morality of Anger," in Hugo Bedau, ed., The Death Penalty in America (Oxford: Oxford University Press, 1982), pp. 334–35. See also Stanley Brubaker, "Can Liberals Punish?" American Political Science Review , vol. 82, no. 3 (September 1988): "[P]unishment expresses and satisfies righteous anger" (p. 825).

Anger is expressed or manifested on those occasions when someone has acted in a manner that is thought to be unjust, and one of its origins is the opinion that men are responsible, and should be responsible…. We can become angry with an inanimate object (the door we run into and then kick in return) only by foolishly attributing responsibility to it, and we cannot do that for long, which is why we do not think of returning later to revenge ourselves on the door…. Anger recognizes that only men have the capacity to be moral beings and, in so doing, acknowledges the dignity of human being.^[79]

By punishing to vent our anger,

we demonstrate that there are laws that bind men across generations as well as across (and within) nations, that we are not simply isolated individuals, each pursuing his selfish interests and connected with others by a mere contract to live and let live.^[80]

Berns tends to speak of punishment as justified "revenge." But rather than see his retributivism as a revenge theory, I think we should invoke a distinction Hegel suggests, between revenge, which is subjective and appeals to an individual's feelings of hurt; and righteous anger, which reflects a social judgment. We can then understand Berns's retributivism as advocating, not the vindictive satisfaction of personal desires to avenge, desires of the sort that fuel violently destructive blood feuds, but, rather, the satisfaction of the demands of justice and right. Berns's retributivism, then, more properly belongs to the next variety of retributive theories we shall consider.

Some may think our lust for revenge explains why we punish. But few retributivists of repute take revenge to be the principle we use to guide us in our practice, and most reject the view that the purpose of punishment in a modern state is

[79] Ibid., p. 334.

[80] Ibid.

to satisfy the desire to avenge. Retribution as revenge is not a compelling account of legal punishment, and it is not the version of retributivism I shall defend.

2.2 Condemnation

According to a second version of retributivism, we do not punish to deter, incapacitate, reform, or satisfy a private desire for vengeance; rather, punishment is justified as an expression of society's condemnation of the offensive act.

Henry Hart emphasizes the condemnatory or reprobative function of punishment in answering the question he poses of what is the distinctive function of criminal (as opposed to other) law. Some utilitarians argue that there is no essential difference in purpose between criminal and tort law; to Richard Posner, for example, the only time there is a justification for invoking criminal rather than civil remedies is when the latter "bump up against a solvency limitation."^[81] Hart claims there is an essential difference; in his view, criminal law, unlike other law, reflects "the judgment of community condemnation which accompanies and justifies its imposition."^[82] Whereas a tort is an injury to a private person, a violation only of private law, a crime is an affront to the social morality articulated in criminal or public law. Crime is

conduct which, if duly shown to have taken place, will incur a formal and solemn pronouncement of the moral condemnation of the community.^[83]

Hart concurs with the view that

The essence of punishment for moral delinquency lies in the criminal conviction itself. One may lose more money on the stock market than in a court-room; a prisoner of war camp may well provide a harsher environment

[81] Posner, "An Economic Theory of the Criminal Law," p. 1204.

[82] Henry M. Hart, Jr., "The Aims of the Criminal Law," Law and Contemporary Problems , vol. 23 (Summer 1958), p. 404.

[83] Ibid., p. 405.

than a state prison…. It is the expression of the community's hatred, fear, or contempt for the convict which alone characterizes physical hardship as punisment.^[84]

Hart takes issue with humanitarian theories that see treatment as the aim of punishment, for such theories fail to see that the essential aim of punishment is to express condemnation:

Today "treatment" has become a fashionable euphemism for the older, ugly word ["punishment"]. This bowlderizing of the Constitution and of conventional speech may serve a useful purpose in discouraging unduly harsh sentences and emphasizing that punishment is not an end in itself. But to the extent that it dissociates the treatment of criminals from the social condemnation of their conduct which is implicit in their conviction, there is danger that it will confuse thought and do a disservice.^[85]

Like all the utilitarians we considered, Hart uses his account of the essential aim of punishment to criticize the existing practice. He says his thesis is "that a sanction which ineradicably imparts blame, both traditionally and in most of its current applications, is misused when it is thus applied to conduct which is not blameworthy."^[86]

Joel Feinberg also emphasizes the reprobative function of legal punishment. Feinberg contends that

[p]unishment is a conventional device for the expression of attitudes of resentment and indignation, and of judgments

[84] George K. Gardner, "Bailey v. Richardson and the Constitution of the United States," Boston University Law Review , vol. 33 (1953), p. 193, cited by Hart, "The Aims of the Criminal Law," p. 405. Neither Hart nor Feinberg, who also cites this passage, observes that contempt and fear are separate reasons for punishing. Fear may motivate the utilitarian more than contempt or hatred does. Punishing in order to condemn is similar in some ways, different in others, from punishing out of revenge, but the equation of condemnation with the expression of hatred or contempt tends to stress only the similarities.

[85] Hart, "The Aims of the Criminal Law," p. 405.

[86] Ibid., p. 405, note 13.

of disapproval and reprobation, on the part either of the punishing authority himself or of those "in whose name" the punishment is inflicted. Punishment, in short, has a symbolic significance largely missing from other kinds of penalties.^[87]

But Feinberg distinguishes his view from Hart's. Hart suggests that condemnation alone may count as punishment, even absent the added consequences of unpleasant physical pain:

[O]therwise, it would be necessary to think of a convicted criminal as going unpunished if the imposition or execution of his sentence is suspended.^[88]

For Feinberg, what makes something legal punishment is its reprobative force and its being "hard treatment."^[89] Whereas Hart suggests that hard treatment is a distinct and nonessential feature of punishment—the convicted criminal whose sentence is suspended is still punished, still stigmatized—Feinberg points out that in some cases it is only the actual infliction of hard treatment that constitutes punishment: "it does not always happen that the convicted prisoner is first solemnly condemned and then subjected to unpleasant physical treatment."^[90] Hard treatment itself brings shame and ignominy and stigma: "[C]ertain forms of hard treatment have become the conventional symbols of public reprobation."^[91]

The condemnation theory of punishment might seem to be nothing but a theory of public vengeance. Feinberg suggests that punishment is in part "a symbolic way of getting back the criminal, of expressing a kind of vindictive resentment."^[92] But in the condemnation theory, punishment also expresses judgments of community disapproval, which needn't be motivated by resentment or even by anger. The condemnation

[87] Feinberg, "The Expressive Function of Punishment," p. 98.

[88] Hart, "The Aims of the Criminal Law," p. 405.

[89] Feinberg, "The Expressive Function of Punishment," p. 98.

[90] Ibid., p. 99.

[91] Ibid., p. 100.

[92] Ibid.

theory is not merely a theory of public vengeance, though it is that as well. It understands legal punishment to serve other purposes besides the venting of public anger, purposes which are distinct also from those declared by the utilitarian. One purpose of punishment, in Feinberg's view, that presupposes its expressive function is the authoritative disavowal of what was done. Another is the vindication of the law:

A statute honored mainly in the breach begins to lose its character as law, unless, as we say, it is vindicated (emphatically reaffirmed); and clearly the way to do this (indeed the only way) is to punish those who violate it.^[93]

The retributivist notes that only by punishing do we affirm right. Unless we punish a wrong, it will be held to be valid.^[94] Hegel suggests that to have a law the violation of which is to be called a "wrong" or "crime" logically requires punishing those who violate it:

[P]unishment, as we have seen, is only crime made manifest; i.e. is the second half which necessarily presupposes the first. Prima facie, the objection to retribution is that it looks like something immoral, i.e. like revenge, and that thus it may pass for something personal. Yet it is not something personal, but the concept itself, which carries out retribution.^[95]

[I]t would be impossible for society to leave a crime unpunished, since that would be to posit it as right.^[96]

A third function of punishment noted by Feinberg is the absolution of others: "Quite often the absolution of an accused hangs as much in the balance at a criminal trial as the inculpation of the accused." Of course, the state could do this job

[93] Ibid., p. 104.

[94] Hegel, Philosophy of Right , par. 99.

[95] Ibid., par. 101 addition.

[96] Ibid., par. 218 addition.

without punishing, "but when it speaks by punishing, its message is loud and sure of getting across."^[97]

All of these functions are nonutilitarian reasons for punishing. They are forward-looking or consequential in some sense, in that they appeal to some future good, be it the vindication of right, absolution of someone wrongly accused, or whatever good we associate with the venting of righteous anger; but it is not on the basis of a calculation of utilities, pleasures and pains, or effects on economic efficiency that they are regarded as good reasons. They are moral or ethical reasons for punishing that presuppose that the purpose of legal punishment is to express public condemnation of certain actions we call crimes.^[98]

2.3 Deontological Retributive Theories

It might be surprising to hear a retributive theory characterized as forward-looking or consequential, for retributivism is usually characterized as nonconsequential, as "deontic." According to one commentator, "[r]etributivism is a deontological theory, different in its logical behaviour from its teleological counterpart."^[99]

A teleological theory would consider the punishment as the means to some good, either general or individual. But the obligation laid upon us by "This is a just punishment" asserts the independent moral value of the punishment itself, considered apart from, and even to the frustration of, some prudential value to be derived from its effects…. In teleological theory we may ask of each action in a sequence "why?" until we reach an answer that is considered prudentially sufficient…. In a deontological

[97] Feinberg, "The Expressive Function of Punishment," p. 105.

[98] In chapter 4, section 4, I shall defend a consequentialist retributivism. The sense in which this retributivism is con sequentialist is rather weak; but it is nevertheless forward-looking enough to be distinguished from the deontic theory we shall consider in the next section.

[99] Mary MacKenzie, Plato on Punishment (Berkeley: University of California Press, 1981), p. 29.

theory, on the other hand, "why?" questions terminate in a judgment which is considered to be morally sufficient—maybe from an intuitionist point of view.^[100]

In this view, where retributivism is seen as a deontological theory that is mutually exclusive of teleological theories, the retributivist insists that we punish, not for any consequences, such as to deter future crimes, or to reform or incapacitate the criminal, but, rather, for the sake of punishing, because punishing is in itself just or right—regardless of the good it may yield.

Does anyone hold such a view? If anyone does, it is Kant,^[101] and it is to his views about legal punishment that we now turn.

The conventional understanding of Kant is that he is a retributivist who opposes all utilitarian justifications of punishment, instead holding that we punish only because justice demands this.^[102] I shall argue that this understanding gets Kant wrong. Kant, it is true, rejects consequentialism in thinking about moral actions, but Kant also thinks law and morality are separate spheres: the justification for a moral action has a different character from the justification for a legal action. Kant's theory of legal punishment is not deontological.

In his Lectures on Ethics , Kant distinguishes moral from pragmatic laws; the latter comprise statute and common law.^[103] Pragmatic laws constrain actions related to other people;

[100] Ibid.

[101] MacKenzie, who makes the claim that retributivism is a deontological, not a teleological, theory, thinks Kant's is the paradigm of a deontological theory of punishment (ibid., p. 29, note 39).

[102] For example, Mitchell Franklin says that in Kant's view it is not justified to punish in order to deter ("The Contribution of Hegel, Beccaria, Holbach and Livingston to General Theory of Criminal Responsibility," in Philosophical Perspectives on Punishment , ed. Edward H. Madden, Rollo Handy, and Marvin Farber (Springfield: Charles C. Thomas, 1968), p. 102. See also Brubaker, "Can Liberals Punish?" p. 826.

[103] Immanuel Kant, Lectures on Ethics , trans. Louis Infield (New York: Harper Torchbooks, 1963).

unlike moral laws, they demand compliance regardless of one's moral disposition.^[104] Whereas pragmatic laws are made by governments, moral laws have no author.^[105] Kant then makes the following distinction, which seems so often ignored: the punishment imposed, by a being who is guided by moral standards is retributive,^[106] but punishment for the violation of (pragmatic) law is imposed to deter or reform:

AII punishments imposed by sovereigns and governments are pragmatic. They are designed either to correct or to make an example.^[107]

Ruling authorities do not punish because a crime has been committed, but in order that crimes should not be committed.^[108]

Kant thinks that the risk of punishment should not be our ground for avoiding evil deeds. We use rewards and punishment in order to make up for our lack of morality,^[109] not to inspire moral action. Kant thinks it is wrong to use the threat of punishment to inspire moral action, but that through punishment man acquires the habit of doing good deeds.^[110]

In his often neglected essay On the Old Saw: That May Be Right in Theory But It Won't Work in Practice , Kant gives a hypothetical example in which he appeals implicitly to deterrence as the justification for legal (but not moral) punishment.^[111] Kant supposes that one man on a life raft pushes the other off to save his own life. Kant says the man does not have a duty to save his own life; rather, he has an unconditional

[104] Ibid., p. 48.

[105] Ibid., pp. 51–52.

[106] Ibid., p. 55.

[107] Ibid.

[108] Ibid., p. 56.

[109] Ibid.

[110] Ibid., p. 57.

[111] Immanuel Kant, On the Old Saw: That May Be Right in Theory But It Won't Work in Practice , trans. E. B. Ashton (Philadelphia: University of Pennsylvania Press, 1974).

duty not to take the life of someone else who is not causing the danger threatening his life. (Kant does not consider the objection that the other man, by consuming what food and drink are available, is indirectly causing a danger to the life of the first.) But, in a footnote, Kant defends "law professors" as

quite consistent in making legal allowance for such emergency acts. For the authorities can't attach any punishment to this injunction, because that punishment would have to be death, and it would be an absurd law that threatened death to one who refuses to die voluntarily in a dangerous situation.^[112]

Kant's reasoning is that state laws, by threatening us with sanctions, are intended to prevent us from acting in certain ways. The point of these laws is to deter. Consequently, a law that imposes a punishment that could not deter the action the law proscribes is absurd. In his Metaphysics of Ethics Kant repeats the lifeboat example and makes the further distinction that the rescued person's killing is not inculpable (unsträflich ) but is impunible (unstrafbar ).^[113] here it is clear how important is Kant's separation of law (Legalität —whether an action accords with a law without further consideration of motivations) and morality (Moralität ). Legal duty is external duty, whereas ethical or moral duty is internal duty.^[114] Legal duty binds by force or coercion (Zwang ).^[115] In the lifeboat example there is a moral, not a legal, duty not to kill the other person. The rescued

[112] Ibid., p. 68, note.

[113] Immanuel Kant, Metaphysik der Sitten , in Kant, Werke in Sechs Bänden , vol. 4, ed. Wilhelm Weischedel (1798; reprinted Darmstadt: Wissenschaftliche Buchgesellschaft, 1963), AB41–42.

[114] Ibid., AB15.

[115] Ibid., AB16. Kant also makes the distinction between recht and gerecht (right and justice). He says that legal action is either gerecht or ungerecht ; moral action is either recht or unrecht (AB23). Kant's distinction is not unlike Hobbes's distinction between injustice and iniquity: Hobbes writes in Leviathan that the sovereign can do no injustice, but he may commit iniquity (ch. 18).

person is to be morally condemned but not legally punished.

Kant's theory of legal punishment does not rule out utilitarian considerations, but does it make sense to call his a retributive theory? If so, it is not because Kant thinks we punish to avenge: "[T]o insist on one's right beyond what is necessary for its defence is to become revengeful … such desire for vengeance is vicious."^[116] There are other reasons why we call Kant a retributivist. Kant holds to the principle of equality in punishing. In the Metaphysics of Ethics Kant writes: "[W]hatever undeserved evil you inflict on someone, you do to yourself…. [I]f you strike him, you strike yourself; if you kill him, you kill yourself."^[117] Kant draws the conclusion that my action should literally be turned back on me—if I kill another, I should be killed.^[118] Kant thus adopts a version of the lex talionis , though he carefully distinguishes this view from one of private revenge.^[119] He also allows that the principle of equality can be valid, if not in the letter, then in its effect; for example, justice can require a nobleman to apologize publicly if he insults someone, of lower class, since a fine would have no impact on him.^[120]

Another reason we call Kant a retributivist is that he insists that a person may be punished only because he has committed a crime, and not for any other purpose:

The criminal must be found to be worthy of punishment [strafbar ], before it is to be thought that from his punishment some use for himself or his fellow citizens can be drawn.^[121]

Kant argues that we cannot punish someone merely to achieve some further good; he opposes, for example, a proposal to

[116] Kant, Lectures on Ethics , p. 214.

[117] Kant, Metaphysik der Sitten , A197–98, B227–28.

[118] Ibid., A199, B229.

[119] Ibid.

[120] Ibid., A198, B228.

[121] Ibid., A196–97; B226–27.

have a person on death row participate in dangerous experiments that could yield beneficial results.^[122] Underlying Kant's position is the view that human beings should not be treated only as a means to some end:

Now I say that man, and in general every rational being, exists as an end in himself, not merely as a means for arbitrary use by this or that will: he must in all his actions, whether they are directed to himself or to other rational beings, always be viewed at the same time as an end.^[123]

This aspect of Kant's theory has been emphasized by other retributivists who oppose what are called humanitarian theories of punishment. A humanitarian theory is a variation of utilitarian theory that advocates therapy and treatment, not punishment.^[124] Karl Menninger, whose views we considered in chapter two, gives such a theory. For Menninger, punishment is nothing but the infliction of pain and suffering; punishment is inhumane and does no good, and our resources would be better spent reforming the criminal who has done "so human a thing."^[125]

The retributivist objection to this seemingly benign position is best expressed by Hegel. In his early Propaedeutik Hegel opposes those who think the state should help people by treating or reforming them even though the person being helped

[122] Ibid. We might ask Kant why he thinks it is wrong to conduct medical experiments on a prisoner, so long as her imprisonment is deserved. In that case she is not being treated merely as a means.

[123] Immanuel Kant, Groundwork of the Metaphysic of Morals , trans. H. J. Paton (New York: Harper Torchbooks, 1964), p. 95; see in general pp. 95–103.

[124] MacKenzie, Plato on Punishment , prefers to distinguish humanitarian from utilitarian theories. For her, both are teleological theories, but for utilitarians the end for which punishment is the means is a general good, whereas for humanitarians it is an individual good.

[125] Another example of a humanitarian theory is that of Norval Morris and Donald Buckle, "The Humanitarian Theory of Punishment: A Reply to C. S. Lewis," in Grupp, ed., Theories of Punishment , pp. 309–16.

does not consent: "To help someone in need, that person must will that I help him, that I still will regard and treat him as equal."^[126] Why? Hegel, drawing on Kant's idea of human beings as ends in themselves, explains in his later Philosophy of Right :

Punishment is regarded as containing the criminal's right and hence by being punished he is honoured as a rational being. He does not receive this due of honour unless the concept and measure of his punishment are derived from his own act. Still less does he receive it if he is treated either as a harmful animal who has to be made harmless, or with a view to deterring and reforming him.^[127]

Kant's theory of legal punishment is not deontological, since for him a legal punishment is not justified that could not deter wrongs. But the theory is retributive, inasmuch as Kant holds to the negative retributive principle that we must punish for no other reason than that a wrong was committed. (However, we've seen that Bentham, a classic utilitarian, also holds to this principle.) Kant gives another hypothetical example, one that appeals to the stronger, positive retributive principle: on an island where all the people were to depart the next day, forever dissolving and dispersing the community, the last murderer in jail would have to have his execution carried out before the diaspora, because justice demands this. Kant is a retributivist because he holds that justice must prevail, "else a people is doomed."^[128] The utilitarians we have discussed would contend that in such cases, where punishment could not possibly be justified by the principle of utility, we should refrain from punishing. On this point Kant and the utilitarians disagree.

At this point we might be puzzled by Kant's views, for the diaspora example is difficult to reconcile with Kant's other

[126] Hegel, Philosophische Propaedeutik , part 2, par. 66.

[127] Hegel, Philosophy of Right , par. 100, Remark.

[128] Kant, Metaphysik der Sitten , A196–97, B226–27. Note how Kant's formulation is consequentialist but not utilitarian.

hypothetical example, that of the lifeboat, whose point is that legal punishment must deter. With the diaspora example Kant claims that desert is a necessary and sufficient condition for punishment. Kant chose the diaspora example precisely because it presents a case in which we must punish even though there is no conceivable basis in utility for doing so. How could Kant claim both that the person on the lifeboat who kills the other need not be punished, because doing so would not provide any deterrent benefit, and that a murderer who could pose no possible future threat to his society must nevertheless be punished? I think a plausible answer is that in the lifeboat example Kant is saying, not that we should not punish a wrong where doing so would be ineffective, but, rather, that the killing should not be regarded as a criminal action, or wrong. Kant can with consistency hold both that we should not call a crime any action that could not be deterred by the threat of legal punishment, and that we must punish crimes even when doing so would not augment social utility.

A deontological theory that insisted we punish even though such punishing would lead to society's destruction would not be very attractive. Retributivists are typically characterized, rather unflatteringly, as nonconsequentialists, as only backward-looking. It is true that for the retributivist the fact that a crime occurred in the past is a compelling reason for punishing. But, as our discussion in this chapter should make clear, one can be a retributivist and still take consequences into account. Both the revenge and the condemnation version of retribution make reference to a future good: the satisfaction of personal desires, or the vindication of right. Utilitarianism is not the only theory that is forward-looking in any degree.^[129]

[129] Justice Thurgood Marshall wrote in Gregg v. Georgia , 428 U.S. 153 (1976): "The … contentions that society's expression of moral outrage through the imposition of the death penalty pre-empts the citizens from taking the law into its own hands and reinforces moral values—are not retributive in the purest sense. They are essentially utilitarian in that they portray the death penalty as valuable because of its beneficial results." Marshall, apparently of the understanding that retributivism means being oblivious to consequences and that arguing for something on the basis of its good consequences means one is utilitarian, winds up labeling as utilitarian what clearly is a retributive position. Marshall could have avoided stretching utilitarianism so far had he seen that to be a utilitarian one must justify an action or practice by appealing to a calculation of net utilities. On the rejection of the simple dichotomy of "deontological" and "consequentialist" (that is, that one must be either one or the other), see Charles Larmore, Patterns of Moral Complexity (Cambridge: Cambridge University Press, 1987), esp. p. xi.

The retributivist position I shall defend is not strictly deontological; it recognizes the importance of justice and right, but is unwilling to insist on justice "though the world perish."

2.4 Just Deserts

A final version of retributivism follows in the footsteps of Kant, up to a point. In this version, desert is a necessary but not sufficient condition for legal punishment. A leading advocate of this view is Andrew von Hirsch, who chaired the Committee for the Study of Incarceration, which clarified the conception of punishment as just deserts.

Von Hirsch argues that utilitarianism alone cannot justify a person's punishment—desert is needed: "While deterrence accounts for why punishment is socially useful, desert is necessary to explain why that utility may justly be pursued at the offender's expense."^[130] In the just deserts theory, desert and deterrence are both essential features of punishment. That someone deserves punishment does not imply that we must punish him, only that we have reason to do so. Because there are moral reasons not to punish, deterrence is needed to "tip the scales back in favor of penal sanction."^[131] The argument, then, is that: (1) committing a criminal act deserves punishment, and desert is a prima facie justification; (2) there is a moral obligation not to add deliberately to the amount of human suffering, which punishment does, and this overrides the case for punishment in (1). Deterrence disposes of the countervailing

[130] Von Hirsch, Doing Justice , p. 51.

[131] Ibid., p. 54.

argument in (2), and so (1) stands.^[132] Since we may safely assume that in some situations (3) will not outweigh (2), von Hirsch's version of retributivism as just deserts holds to the weaker negative retributive principle that demands merely that we punish only for an offense, not to the positive retributive principle that we punish all offenses.

Von Hirsch overlooks an important point, one that is central to the retributive position I shall defend: we can be committed always to punishing for an offense, for retributive reasons, without being committed to inflicting punishment in a certain way. In his argument above, von Hirsch acknowledges a moral demand to express condemnation (to punish), but also recognizes a moral objection against the infliction of pain or suffering (punishment). But not all legal punishment inflicts pain or suffering. The retributivism I shall defend insists that we punish when punishment is deserved, but recognizes that how (or how much) we punish is a separate matter. This is the position Hegel takes: "[T]he only interest present is that something actually done [i.e., that we punish] … no matter how."^[133] Von Hirsch fails to take into account that we can with consistency insist always on punishing, for retributive reasons, and insist that the way we punish be justified on utilitarian grounds.

In the theory of retribution as just deserts the idea of desert is used to account for why it is morally permissible to punish certain individuals, but in itself it does not account sufficiently for why we are justified in having a practice that punishes those who deserve punishment.^[134] But not everyone accepts even

[132] Ibid.

[133] Hegel, Philosophy of Right , par. 214, Remark; cf. par. 214, Z.

[134] David Dolinko has recently distinguished the "moral justification" of punishment—an account of why it is morally legitimate to punish, or of by what right we may punish—from the "rational justification" of punishment—an account of for what reason we punish at all. Dolinko notes that we can agree that there are reasons for punishing without agreeing that it's morally legitimate to punish, and vice versa ("Some Thoughts About Retributivism," Ethics , vol. 101, no. 3 [April 1991], pp. 539–40).

the idea that because a person has committed a crime, it is, on the theory of just deserts, morally permissible to punish him. Consider the (true) story of Leroy Strachan. Mr. Strachan was recently arrested in New York City by two Miami police officers. He was wanted for the murder of a police officer in Miami forty-three years previously. In the intervening years Mr. Strachan, now a sixty-one-year-old Harlem resident, married, raised three children, and operated an elevator in SoHo for twenty-one years. The manager of the building where Mr. Strachan worked describes Mr. Strachan as "a very good worker who never had a problem." Since the incident forty-three years previously, Mr. Strachan had not been arrested for any offense. When the Miami police officers confronted him with the crime, Mr. Strachan confessed. If extradited to Florida and convicted, he could face the death penalty.^[135] According to the theory of just deserts, should Mr. Strachan be punished?

Some retributivists would insist, for various reasons, that Mr. Strachan, if convicted, does deserve punishment. Von Hirsch, however, on the basis of the theory he lays out, would probably be committed to opposing his punishment, since in this case punishment is unlikely to have much of either a general or a specific deterrent effect.^[136] So, too, would another theorist who offers a more sophisticated version of the theory of just deserts. In James Griffin's view, retribution as just desert is premised on the idea that desert is a reason for action, but not one that necessarily trumps other reasons. Griffin argues that merit (or desert) is not a moral reason for action or a criterion for moral right and wrong, but it is a reason for action: "The element of appropriateness itself constitutes a reason."^[137] Griffin means that it is appropriate to express admiration for one who merits it—to admire only on utilitarian

[135] New York Times , February 17, 1990.

[136] General deterrence refers to the effect on members of society in general; specific deterrence, to the effect punishment would have on Mr. Strachan.

[137] James Griffin, Well-Being: Its Meaning, Measurement, and Moral Importance (Oxford: Clarendon Press, 1986), p. 259.

grounds would not really be to admire; but your merit does not give you a right to receive, nor impose a duty on me to express, my admiration. Griffin then takes the case of punishment—a case of demerit rather than merit—and comes to somewhat different conclusions. Griffin argues that your demerit is a moral reason for punishing you, but only under what he calls the "repentance view."^[138] What you did was morally wrong, so I have a moral reason to punish you, but "my response to your wrongdoing is appropriate only when, and to the extent that, it contributes to your going through the … process: perception, guilt, and repentance."^[139]

Griffin distinguishes the repentance view from what he calls the atonement view, which he dismisses. In the atonement view, my act of demerit needs to be punished; but the problem with this view is that "if I change, I should want [people] to respond now to the person I am now."^[140] If I have changed, neither deterrence nor retaliation seems appropriate.^[141] Griffin identifies Kant with the atonement view and suggests that Kant, who does not want to deny to a person the dignity of being morally accountable, fails to see that

the atonement view is not the only way to show respect for persons…. If anything, it is the repentance view, in which a person is given more weight than an act, that shows respect for persons.^[142]

In Griffin's view, then, the only desert-based moral reason for punishing me is, not to atone for the wrong I've done, but to

[138] Ibid., p. 270.

[139] Ibid., p. 272. Griffin's argument, then, seems to be: if you commit a wrong, then you deserve to be punished—we have a moral reason to punish you. But it's appropriate to give you what you deserve only if doing so will get you to repent. In other words, the moral reason for punishment is weighed against the criterion of appropriateness. In this argument is the same sort of utilitarian weighing we find in von Hirsch's argument. The Kantian would question Griffin's invocation of the criterion of appropriateness.

[140] Ibid., p. 268.

[141] Ibid., p. 269.

[142] Ibid.

make me repent.^[143] And "if the wrongdoer is an exceptional moral agent who spontaneously repents and reforms, there is virtually no place for punishment."^[144] On the other hand,

if he is an adult who still needs to learn but resists the lesson, then maybe punishment would teach the seriousness of wrongdoing where words alone would fail. If he is someone who cannot learn no matter what the lesson, then punishment has no place as a response to the person that we are dealing with.^[145]

Griffin, then, suggests a reason not to punish Mr. Strachan.^[146]

Although Griffin speaks of desert as a moral reason for punishing, in his view it amounts only to a necessary but not sufficient condition for punishing. Griffin comes to a conclusion similar to von Hirsch's, although some would call his view humanitarian and von Hirsch's utilitarian, since for Griffin the sufficient condition for punishment is moral improvement of the individual, whereas for von Hirsch it is the improvement of society that results from deterrence.

The retributivists who focus on just deserts seem to give in to utilitarianism, and we might ask whether there is any essential difference between the two sorts of justifications. Retributivism in the version of just deserts holds only to the weaker, negative retributive principle, that we punish only those who are guilty of a crime. Neither von Hirsch nor Griffin expresses Kant's demand that we must punish when we have reason to punish. None of the utilitarians we discussed denies that desert is a necessary condition for punishment. In the theory of punishment as just desert, utilitarianism and retributivism blur.^[147]

[143] Ibid., pp. 270, 272.

[144] Ibid., pp. 270–71.

[145] Ibid., p. 271.

[146] It's unclear whether Griffin himself would apply his argument in this way. He clearly states that his argument applies to moral punishment. Other considerations may enter for him when thinking about legal punishment.

[147] Cf. Dolinko, "Some Thoughts About Retributivism," p. 543.

3. Deciding between the Utilitarian and Retributive Accounts

We have examined both utilitarian and retributive theories of why we legally punish at all. We have seen that some theorists who claim to be retributivists sound very much like utilitarians. Before deciding between the two accounts, we must first establish what version of retributivism we shall hold up to utilitarianism, and see whether we indeed need to choose between the two.

From our account of theorists commonly called retributivists it is clear that there is no distinct set of criteria the satisfaction of which is essential for meriting that label. Some retributivists (von Hirsch, Griffin) hold only to the weak, negative retributive principle that declares we may punish only for an offense. This principle is an offshoot of the Kantian demand that we treat each person as an end and not merely as a means—we may not punish for any reason other than desert (demerit). But we have seen that utilitarians also hold to the negative retributive principle (how they can do so consistently will be taken up in chapter 5, where we consider rule-utilitarianism). Other retributivists (Kant) hold to the positive retributive principle that we must punish all offenses, a demand the former group of retributivists reject. Some retributivists (Feinberg, Kant, Hegel) insist that we punish the act to vindicate right, and others (Griffin) insist that we punish the person and, therefore, that if the person changes, punishment is no longer justified on retributive grounds. Some retributivists (Henry Hart, Feinberg) insist that we legally punish only morally blameworthy conduct, and it is this demand that constitutes their retributivism. Kant rejects this demand.

The retributivist label, then, might not seem particularly useful, for the differences on particular issues among some retributivists may seem greater than the differences between some retributivists and some utilitarians. Still, there are features of some of the theories commonly called retributive that clearly distinguish them from utilitarian theories and that, I

believe, persuasively articulate an ideal immanent in the practice of legal punishment. I refer to the insistence that we punish to express condemnation of an act society regards as blameworthy, to mete out just deserts, and to vindicate right. These purposes are inextricably connected. We might say they amount to different articulations of one single purpose: that we punish for justice.

This version of retribution insists that we punish to mete out just deserts, but it does not commit to the nonconsequential Kantian version of (moral) retribution that insists we punish for the sake of justice "though the world perish." Yet, in holding that we have a compelling reason for punishing that is independent of what the consequences are either to the sum of individual utilities or to economic efficiency, it is opposed to the utilitarian accounts we considered. Although it draws on the idea von Hirsch emphasizes, that we punish to mete out just deserts, it insists, contrary to von Hirsch's compromise retributivism, that we must punish to vindicate right even if a utilitarian calculation would determine that we shouldn't punish. This retributive position is not oblivious to consequences. It need not insist that we send people to prison even if it were the case that prisons were "schools for crime" and doing so would lead to more crime. The theory insists only that we express our condemnation in some way , for if we do not declare that a criminal did something that was wrong, then "right" and "wrong" have no meaning.

The retributivist position that we punish to express righteous anger, mete out just deserts, and vindicate right points to an ideal of justice I believe is immanent in the practice of legal punishment, but an ideal which sometimes gets lost in the shadows of our institutions. It is the version of retributivism I shall defend. But how does one go about defending a justification for legal punishment? How does one justify the claim that the retributive and not the utilitarian theory is the best account of why we punish?

One approach is to appeal to what "punishment" means in ordinary language. Consider the following exchange:

Reb: "The essence of punishment is that it is deserved. Punishment is punitive. It is a response to a wrong, an expression of our condemnation. If it's not deserved, it's not punishment but something else."

Ute: "Sometimes we punish people not because it's deserved, but to deter future crime or to incapacitate the criminal. For example, we will inflict additional punishment on a repeat offender because we believe he poses a future threat to society. This is part of our practice, and it counts as punishment."

Reb: "The moment the prison sentence that is deserved expires, the repeat offender who remains in chains as a future threat is no longer being punished. If it's not for a past wrong then it's not punishment but some other practice, and we might even give it a different name, say, 'telishment.'"^[148]

Ute: "I bet this 'telishment' that's not really 'punishment' sure would feel like punishment to the poor man!"

Reb: "Well, telishment feels a lot like punishment, but sometimes so does illness. What makes something punishment is not that it hurts, but that it's deserved. Spanking a child is punishment, but not merely because it hurts; for we also call it punishment when a teacher sends a naughty pupil to the back of the room and makes her wear a duncecap. This is no doubt painful, but no physical pain has been inflicted. Sometimes punishment embarrasses, sometimes it physically hurts. Sometimes neither is the result. Being given extra homework or receiving a parking fine needn't embarrass or hurt; and the death penalty doesn't embarrass and can be quite painless. All of these things are punishment, because they are responses to an action regarded as wrong or blameworthy, as deserving of condemnation."

In this exchange Reb is pointing to the punitive character of punishment, a character implied by our use of the word in

[148] The idea of "telishment" comes from John Rawls, "Two Concepts of Rules," in Michael Bayles, Contemporary Utilitarianism (Gloucester, Mass.: Peter Smith, 1978), pp. 59–98.

ordinary language. The retributivist, like Reb, can appeal to the meaning of the word "punish" and the concept of punishment in arguing that we punish to express condemnation of those who deserve it. We shall see in chapter 5 that Reb's appeal to ordinary language is not entirely convincing, that there is no single essence of punishment: we do call it punishment when we inflict pain or hurt or stigmatization on those who, by Reb's standards, might not deserve it. Mistaken punishment of innocent persons, punishment of those who did not intend to commit a wrong, and extended punishment of repeat offenders are all still called "punishment"—and are punishment—even though we might not think the pain, hurt, or stigmatization is deserved. A defense of retributivism based on linguistic intuition won't be enough to satisfy everyone (or even myself).

The immanent critic insists that the ideals according to which she criticizes actual practice derive from the practice. Both the utilitarian and the retributive immanent critics claim they have adduced the principle immanent in the practice. Ultimately, to decide who is right, we have to look at the practice, see what is done and what principle best makes sense of what is done. We must ask whether the utilitarian or the retributivist ideal better fits the practice. In looking at actual practice, we shall see aspects that accord with both principles; the retributivist, for example, will find many instances where the ideal of condemning blameworthy actions and vindicating right is not what is being done by practitioners. In such cases, the retributivist might have to stick to his ideal in the face of contrary facts and argue that such instances, although part of actual practice, go against the best conception of what it means to punish. In deciding between retributivist and utilitarian accounts, then, we are deciding which ideal better accords, on the whole, with the purpose of the practice, with the concept of punishment, and even with the meaning of the word "punish."^[149] Ultimately, to decide between the two accounts

[149] We've seen that when justifying a practice as a whole, there are no clear standards for what counts as success (chapter 1, section 2). In chapter 5, section 2, we shall consider more explicitly how we decide among competing justifications of a practice. On my view that practices as a whole lack absolute grounds and on my commitment to nonfoundationalism, see also chapter 1, note 26, chapter 2, note 95, and chapter 5, note 2.

we must step inside the practice and see what is done. This we shall do in the following chapter.