Gertrude Ezorsky “Punishment,” writes McTaggart, “is pain and to inflict pain on any person obviously needs justification.” But if the need to justify punishment is obvious, the manner of so doing is not. Philosophers have advanced an array of diverse and conflicting arguments to justify punitive institutions. I shall sort their claims into three varieties: teleological, retributivist, and teleological retributivist.
What are the distinctive claims of our three kinds of philosophers?
Teleologists believe that punishment should yield, in fact, some further effect, which is desirable. Thus Bentham, a utilitarian, held that while the suffering of punishment is itself evil, nevertheless the threat of punishment, strengthened by enforcement, may serve a good purpose, e.g., deterrence of aspiring criminals and a consequent reduction in the misery wrought by crime. Notice that this sort of view may be empirically confirmed, or refuted, by a factual investigation, for punishment is conceived as a causal means which, given our laws of nature, will yield the effect of crime prevention.
Retributivists, however, take a different view of the matter. They claim that necessarily the distribution of deserved suffering for wrong doing is either just or intrinsically valuable, irrespective of any further good consequence, e.g., crime prevention. Some philosophers might put the matter in this fashion: punishment for immorality would exemplify justice or have worth not merely in our familiar world, but in any possible world.
Teleological retributivists pay their respects to a plurality of principles. Thus, they share with utilitarians the notion that penal laws should yield some demonstrable beneficial consequences. Justice is not served by the infliction of deserved suffering for its own sake. But they derive the following view from retributivism: justice is served if teleological aims are held in check by principles of justice, e.g., that the suffering of punishment should not exceed the offender’s desert.
Let us consider the merits-and the demerits-of these three perspectives on punishment.
Teleologists view punishment as desirable either primarily for the guilty man, i.e., making him a better person, or primarily for the world, e.g., by isolating and reforming criminals or deterring potential offenders, punishment makes the world a better place.
Early better man thinkers, like Plato and Aristotle, conceived crime as a spiritual disease, curable by the bitter medicine of punishment. For Hegel, (as interpreted by McTaggart) the pain of punishment yields repentance, whereby the criminal recognizes his sin. He does not merely change his ways. Fear of future punishment might yield this superficial reform. He really becomes a better man; thus, Hegel declares, realizes his true nature.
It is tempting to challenge curative theories by pointing to the sparsity of supporting evidence. (Do hardened criminals really mend their ways, either superficially or in depth, when punished?) But there are more fundamental objections at hand. Let us assume that in some cases punishment really does produce the effects claimed by these philosophers. It would not follow then that punishment is justified. Suppose the social costs of producing the punitive cure required were very burdensome? Or suppose that symptoms of a propensity to commit crime appeared before crimes were committed? Should preventive punishment-if effective-be imposed as we might impose preventive medicine against communicable disease? Imagine that a very severe punishment cured someone guilty of a petty offense. Would not such a punishment be undeserved, hence unjust?
The same sort of problems arise for those who see treatment, e.g., psychoanalytic therapy or drugs, as an alternative to punishment. Should preventive treatment be imposed on persons who will most likely commit crimes? Should the bill for slightly successful, but very costly, treatment be imposed on society? Remember, too, that cures can be more painful than punishments. According to a 1966 experiment performed in a Canadian mental hospital, treatment can, to some extent, reduce the amount of liquor ingested by alcoholics. The treatment? Intravenous injection of 20 mgs. of succinylcholine chloride which induces paralysis and suppression of respiration, that is, a drowning to death experience. Surely no reasonable person would suggest treating alcoholic petty offenders in this fashion. The treatment would, of course, not be imposed as a punishment. Yet who would hesitate to call it undeserved? The moral problems raised by curative theories of punishment (or treatment) require, not merely an account of beneficial effects on the criminal, but a comprehensive moral perspective. Utilitarians claim to have it.
Utilitarians are better world teleologists. They evaluate punishment as follows. A moral agent ought to choose that act which of all feasible alternatives, has maximum utility. (The utility of an act is measured by its efficacy in producing happiness or reducing suffering, for everyone.) How does the suffering of punishment fit into this scheme? Bentham puts the matter in this fashion:
The general object which all laws have, or ought to have in common, is to augment the total happiness of the community: and therefore, in the first place, to exclude, as far as may be, every thing that tends to subtract from that happiness; in other words, to exclude mischief ... But all punishment is mischief; all punishment is in itself evil. Upon the principle of utility, if it ought at all to be admitted, it ought only to be admitted in as far as it promises, to exclude some greater evil. (Principles of Morals and Legislation, Ch. XII).
Punishment serves to “exclude some greater evil” when by the workings of isolation, reform, and deterrence, the misery and insecurity created by crime is reduced.
It may be objected that the reform and deterrence effects of punishment have been exaggerated. Nevertheless, it is reasonable to suppose that some criminals, when punished, do not repeat an offense and that the threat of punishment stays the hands of some persons tempted to crime. In such cases, if punishment, as compared to other alternatives, e.g., psychiatric treatment, has maximum utility, the utilitarian is obligated by his views to endorse punishment.
Critics of utilitarianism claim that punishment which passes utilitarian standards may be undeserved, hence unjust. Indeed, as they see the matter, utilitarians are committed to undeserved punishment of two sorts, legal and illegal.
Let us consider the legal variety first. A legal punishment is undeserved when the offender is not morally responsible for the offense or if his prescribed punishment is excessive. An offender is not morally responsible when the following conditions obtain: he is punished either retroactively (by ex post facto law), or vicariously (for the act of another), or he has a valid excuse for committing the offense, i.e., insanity, ignorance or mistake of fact, necessity, incompetence, or automatism. When such excuses are not accepted, the offender is being held strictly liable.
How could retroactive sanctions, vicarious punishment, or strict liability serve utilitarian purposes?
Consider the following sort of situation. Suppose death and injury due to shooting were very much on the upgrade because too many people had guns. In that case there may be a utilitarian rationale for the following laws prescribing undeserved punishment.
Retail gun sellers are held strictly liable if any of their guns are found on a person without a permit. Sellers of guns would become extra careful in scrutinizing permits and protecting their stocks from theft. The iron clad rule, No excuses, toughens the deterrent threat of punishment.
Vicarious punishment is prescribed by law for parents if any of their children under twenty-one is discovered in possession of a gun. Parents, consequently, would make a greater effort to ensure that their children have no guns.
An ex post facto law prohibits the use of guns by individuals against unarmed persons trespassing on their property. As a consequence of this law, shootings by persons not covered by the law, e.g., armed guards, decline because they fear future ex post facto laws under which they would be retroactively punished.
It is perfectly possible that laws prescribing these three sorts of undeserved punishment would maximize utility by reducing deaths and injuries due to shootings. But consider the injustices that might be perpetrated under such laws.
Suppose gun sellers were held strictly liable if one of their guns were found in the possession of an individual without a permit. F, a scrupulously honest retailer, sells a gun to G, who has a permit. G plants the gun on H, who has no permit. Discovery of the gun on H sends faultless F to prison.
Imagine that parents are vicariously liable if a gun is discovered on the person of any of their children. An eight-year-old girl on a treasure hurt with her classmates finds a package which she brings back to school as her “treasure.” The child’s parents, ardent pacifists, are convicted and serve a prison sentence.
Suppose an ex post facto law makes it a criminal offense to shoot unarmed trespassers on one’s property for any reason whatsoever. Before enactment of this law, the following incident occurred. Gangsters threatened to kidnap Jones’ children. Jones saw three of them on his property making off with his child. He fired at one who is unarmed. Note that at the time Jones’s act was not only morally defensible, it was perfectly legal. But when the ex post facto law is enacted, Jones is convicted and punished, retroactively.
Laws prescribing excessive punishment, i.e., undeserved in light of the offense, may also satisfy utilitarian standards. Imagine a community where loitering is so widespread as to be a public nuisance. A law is enacted prescribing one year in prison for loitering. It might very well be the case that only one person would be punished under this law. The threat of such severe punishment, reinforced by infliction on one offender, is sufficient to deter all other potential offenders. The good utilitarian effects of the law may very well outweigh the bad ones. But a one year jail sentence for loitering is certainly undeserved, hence unjust.
Finally, we may note that the guilt of an offender is usually held to decrease with extenuating circumstances, e.g., that the crime was committed in a state of passion. But, as Bentham noted, where the temptation to commit an offense is greater, the threat of more severe punishment is required to overcome the temptation. Thus to secure effective deterrence, crimes committed in extenuating circumstances should be punished more severely. However, these are precisely the circumstances where the offender, being less responsible, deserves a lighter penalty.
In all cases just described, punishment while either wholly or partially undeserved would, by hypothesis, be perfectly legal. But utilitarians, it may be argued, are also committed to undeserved punishment which is illegal.
Let us suppose there has been a wave of vicious crimes, and the police are unable to find the culprit. Since no one is punished for these offenses, the deterrent threat of punishment becomes increasingly ineffective, and more persons are tempted into committing the crime. To frame and punish an innocent for these offenses may reinforce the deterrent threat of punishment. A few such scapegoat punishments might avert great harm and be worth while utility-wise.
It may be objected that those who hold a revised form of utilitarianism, namely, rule utilitarianism, are not committed to such illegal punishment of the innocent. A rule utilitarian believes that our conduct should be guided by estimating the utility not of any single, particular action, but of the general practice of that kind of act. Rule utilitarians may argue that they are not committed to what Professor Rawls calls “telishment,” the illegal punishment of innocent scapegoats. Such punishment might, in a particular case, maximize utility, but it is doubtful that the general practice of “telishment” would meet utilitarian standards. Consider that the deterrence effect of the general practice of “telishment” depends on the success of systematic deception. Unless the general public believes the innocent scapegoat is guilty, the desired deterrence effect would not be achieved. But it is extremely difficult to imagine that such systematic deception could be maintained successfully. Hence the practice of “telishment” cannot be justified on rule utilitarian grounds.
Notice, however, that the rule utilitarian argument against illegal punishment has no application to legal undeserved sanctions. Excessive penalties, strict liability, retroactive and vicarious punishment are perfectly compatible with rule utilitarian principles. No deception is required to sustain the practice of these undeserved punishments. Provision for such punishments can and has been incorporated into public law.
Moreover, it seems false that the rule utilitarian is never committed to “telishment.”
Suppose that D, the sovereign of a powerful nation, has a grievance against Smith, a citizen of a small nation. D demands that E, the sovereign of Smith’s country, ensure that Smith be convicted and imprisoned for some particularly disgraceful offense of which Smith is innocent, e.g., sexual assault on a child. Unless innocent Smith is convicted and punished, D will launch an attack on E’s country and massacre the whole population. E can with the aid of a few trusted security police manufacture evidence which would convict Smith in a court of law. The massacre of a whole nation would surely have more disutility than the harmful effect of Smith’s illegal punishment. Thus it seems that for rule utilitarianism, in this rare kind of case where averting a catastrophe depends on an innocent’s being “telished,” the rule should be, “telish” him.2
It may be objected that in this kind of case the rule utilitarian’s violation of justice is warranted. We might reluctantly agree to the punishment of one innocent in order to prevent the massacre of a whole nation. But as Hart notes “we should do so with the sense of sacrificing an important principle.” However, the rule utilitarian is not sacrificing any of his principles. On the contrary, punishing innocent Smith satisfies his principles.
Pure retributivists claim that deserved punishment is necessarily either just or of some moral worth. Undeserved punishment is always either unjust or of negative moral worth. As Kant sees the matter, to perpetrate undeserved punishment for any purpose whatsoever is to use a person as a means only, rather than as an end in himself.
For some retributivists punishment for wrong doing is just because it restores the moral balance disturbed by crime. What is a state of moral balance? One version derives from Kant. A moral balance is exemplified by “a proportion between welfare and well-doing.” The good are happy and the wicked suffer for their misdeeds. Thus an imbalance obtains when a criminal fails to suffer for his crime. Punishing him sets matters right, with respect to the criminal. Full moral balance is achieved, however, when the criminal is punished and the victim is compensated. Then a proper proportion between welfare and well-doing is achieved.
The trouble with this moral balance view is that if it justifies punishment it can also justify crime. Suppose a moral disproportion obtains in M’s case. He has been made to suffer excessively for his crimes. However, a moral disproportion also obtains in N’s case. Given N’s conduct his happiness is excessive. (N is good, but not that good.) Suppose M commits a crime against N. It is possible that M’s crime against N may achieve a state of moral balance for both M and N. Thus if obtaining a moral balance justifies punishment, it can do the same service for crime.
For all retributivists punishment has moral worth independently of any further desirable effects. Ceteris paribus, the world is better, morally speaking, when the vicious suffer. Thus it is not surprising that retributivism is sometimes characterized as the vindictive theory of punishment.
We may test this position by imagining a world in which punishing criminals has no further effects worth achieving. Thus the criminal, punished, is perfectly ready to go out and commit his crime all over again. Ordinary men are not deterred in the slightest from crime by the threat of punishment. Victims of crime have no desire for retaliation, and the pleasure of vengeance is unknown. According to retributivists, individuals in this world are still obliged to bear the burden of maintaining institutions of punishment. Indeed, they have an obligation to sacrifice so that punishment is kept going. But what sort of sacrifice is anyone obligated to make for the sake of utterly useless punishment? I suggest that in this case there is no obligation whatsoever to punish wrongdoing.
Retributivists believe, presumably, that their views should find expression in actual systems of law. Is this possible? Remember these philosophers emphasize that only the guilty may be punished. But no infallible method for determining guilt has ever been devised. Indeed, it is a virtual certainty that honest, reasonable jurors have convicted defendants who appeared guilty but were, in fact, innocent. Thus, as it turns out, the price of a system which punishes the guilty is sacrifice of some innocents. Unless retributivists avoid punishing the guilty, they will be unable to avoid punishing the innocent.
Moreover, it is possible that more innocents would be punished in a society governed by retributivist, rather than utilitarian, principles. Here is why: for utilitarians, punishment is justified only if, by comparison with other alternatives, e.g., treatment, punishment would maximize utility. Suppose a painless but expensive pill were devised which cured any propensity to crime. The pill’s utility surpassed that of other alternatives to punishment and slightly exceeded that of punishment itself. In that case utilitarians should endorse adoption of the crime cure pill and abolition of punishment. But retributivists believe that punishment of the guilty is either necessarily just or has some intrinsic value. In that case retributivists might refuse to substitute the crime cure pill for punishment. But juries would most likely remain fallible. Thus some innocents would be punished along with the guilty. Hence if retributivist principles held sway, some innocents would, as a consequence, be punished. While if utilitarians had their way, not a single innocent would suffer punishment.
Teleological retributivists, whom I shall call TR philosophers, are pluralists. They mediate between a teleological principle, i.e., utilitarianism, and principles of justice held by retributivists. Let us contrast utilitarianism and retributivism with one plausible version of TR pluralism. I shall refer to a philosopher who holds this version as TR, to the retributivist as R, and to the utilitarian as U.
(A1) If X deserves to suffer, then the amount of suffering in the world ought to be increased by X’s suffering, as much as he deserves.
(A1), a claim about the amount of suffering in the world, is an aggregative desert principle. R believes (A1). However, both TR and U deny that (A1) is true. TR and U believe a different aggregative principle, (A2).
(A2) The total amount of suffering in the world ought not be increased by anyone’s suffering.
Consider now (D).
(D) If either X or Y is to suffer, and X but not Y deserves to suffer, then X not Y ought to suffer (but not more than he deserves).
(D) is a distributive principle, since (D) is a claim about how suffering ought to be distributed. Both TR and R believe (D). However, U denies that (D) is true. Only the amount of suffering matters to the utilitarian. Who does the suffering is morally indifferent. Thus TR shares a distributive desert principle, (D), with R and an aggregative utilitarian principle, (A2), with U.
The different views of TR, R, and U may be illustrated in the following cases.
Suppose three men know they will die within 24 hours. The first, a philosopher, is not in pain, but the other two—a sadistic ex-SS criminal and his former victim—are suffering and to an equal degree. The philosopher has two pain killing capsules, each totally effective for 24 hours. If the philosopher were TR or U, he would, following (A2), give each sufferer a tablet. However, if he were R, he would, following (A1), refuse to give the Nazi criminal a tablet.
Imagine now that the philosopher had only one such indivisible pill. If he were R or TR, then he would, in accord with (D), give the pill to the innocent man, not the criminal. However, if the philosopher were U, he would have no reason to alleviate the innocent’s pain rather than the criminal’s. After all (A2) would be satisfied by either choice.
How would our TR philosopher justify punishment?
Consider the case of C who has committed a crime. If he were punished, as he deserves, then, and only then, would he be deterred from committing a further crime against some innocent person. Hence, punishing C, a criminal, would prevent the suffering of an innocent. Our TR philosopher, following (D), would claim that C ought to be punished.
Notice, however, that (D) is an affirmative principle indicating when the guilty should be punished. But TR philosophers, like retributivists, also hold some negative principle of justice, e.g., Kant’s principle of humanity, which implies that undeserved punishment is wrong.
Kant, we recall, urged that a person ought not be treated as a means only, but always as an end-in-himself. Suppose that an undeserved punishment would serve utility. Then, as Raphael, a TR philosopher, puts the matter:
The claim of social utility is opposed by the claims of the individual to be treated as an “end-in-himself” and not merely as a means to the ends of society.
(HP) is an application of Kant’s principle of punishment.
(HP) Never treat a person as a means by punishing him undeservedly in order to benefit others.
Both TR and R would endorse (HP). Thus TR is committed to three principles: (HP), (D), and (A2). [R adheres to (HP), (D), and (A1), while U is committed to (A2)].
Are TR’s three principles really compatible? Remember innocent Smith. Unless Smith’s sovereign punishes him for a disgraceful offense of which Smith is innocent, the head of a powerful nation will have the entire population of Smith’s country massacred. Suppose Smith’s sovereign were TR. He could not adhere to both (HP) and (A2). By punishing Smith he violates (HP). But by not punishing Smith he violates (A2). [If Smith’s sovereign were R or U, he would have no problem. R is committed to (HP) but not to (A2). Hence R would refuse to punish Smith and allow the nation to be massacred. U is committed to (A2) but not to (HP). Hence U would punish Smith to prevent the massacre.]
Some TR philosophers, e.g., Ross, hold that an innocent should be punished “that the whole nation perish not.”
If our TR sovereign took this view he would sacrifice’ a principle of justice, (HP), in favor of a utilitarian principle, (A2). But TR’s view of undeserved punishment would still be quite different from the utilitarian’s. Remember U is not committed to (HP). Hence U would endorse undeserved punishment to achieve a slight gain in utility. But TR would abandon (HP) and opt for undeserved punishment only to avert a catastrophe. In that case TR would avowedly be sacrificing one of his principles, (HP), in favor of another, (A2). But in punishing innocent Smith, U sacrifices no principle whatsoever. TR’s position is, I believe, closer to our common morality.
A complete TR view of punishment would incorporate two types of principles. The first may be dubbed a recognized first order principle, e.g., (A2), (1)), or (HP). However, such principles may, as we have seen, conflict. Hence, some absolute, second order principle is required to referee the outcome. Thus our TR sovereign would, by punishing Smith, be following some absolute, second order principle which implies that in catastrophe cases, (A2) takes priority over (HP).
R believes that deserved punishment, in toto, should be inflicted on the criminal. While for TR the deserved penalty merely sets the upper limit of permissible punishment. But the notion of a deserved penalty is not easy to explicate.
Note first that the deserved penalty may differ from the penalty a state is legally entitled to impose. A government is legally entitled to inflict any punishment whatsoever-no matter how undeserved-if law permits. Thus, should cutting off an offender’s hand be the legal penalty for petty thievery, then a judge is legally entitled to impose such sentence. But plainly the penalty is undeserved. Legal entitlement is not a moral concept. We ascertain legal entitlement by consulting a legal code. But a legal code cannot suffice to determine what a criminal deserves to suffer as a penalty. The misery he deserves depends on the moral wrong he has committed. Thus desert is a moral, not a legal notion. Suppose that torture of children were perfectly legal and shoplifting were illegal. It would be true, nevertheless, that those who torture children deserve to suffer more than shoplifters.
To determine the deserved penalty, one must decide how much misery an offender deserves for his wrongdoing. In such assessment, two kinds of questions may arise, one of degree and one of scope.
Let us look at the degree problem first. Consider the following kinds of offenses.
Type A offenses are usually more grave, morally speaking, than type B. Hence it is usually reasonable to claim that a type B offender deserves a lesser degree of suffering as punishment. But such a comparative claim is crucially incomplete. We should like to know just how much misery a blackmailer or petty thief deserves.
Suppose we ranked offenses in an ascending order of moral gravity and penalties in a parallel order of severity. Could we then match offense to deserved punishment? What would constitute deserved punishment for the most serious moral wrongs?
Hobbes describes the punishment for high treason in 17th century England as follows:
To be drawn upon a Hurdle from the Prison to the Gallows, and there be hanged by the Neck, and laid upon the ground alive, and have his Bowels taken out, and burnt, whilst he is yet living; to have his Head cut off, his Body to be divided into four parts, and his Head, and Quarters to be placed as the King shall assign.
By comparison a one year prison sentence seems mild indeed. But would we not consider a one year sentence too severe for a mild offense (e.g., loitering)?
Note again that plausible comparative rankings of offenses by their degree of moral seriousness and penalties by their degree of severity seem possible. But such rankings cannot determine the noncomparative degree of suffering deserved for a specific offense.
The difficulty may arise in part because desert is a moral, not a legal, notion. Yet our intuitions of criminal desert are moored to our legal code, i.e., what seems mild or severe by that code.
Some philosophers suggest that a punishment is deserved if it fits the crime. But a criterion of fittingness is notoriously difficult to explicate. Consider one plausible formulation of such a criterion:
A punishment is fitting if and only if the degree of suffering inflicted on the criminal equals the degree of suffering imposed by the criminal on his victim.
Is it? Compare a criminal who assaults a helpless invalid with one who assaults a healthy adult. Suppose both victims endured an equal degree of suffering. The first criminal would still deserve a more severe penalty. His offense is, after all, much worse morally speaking.
A fitting penalty should match the moral evil of the offense. But how does one know when the match is made? For example, how many years in prison does the first criminal deserve for his crime?
I turn now to the problem of scope.
Suppose one could match deserved misery to moral evil. Let us assume then that () is true.
() A penalty of Z’ and no more than Z’ is deserved for the offense of Z, of Y’ and no more than Y’ for the offense of Y, of X’ and no more than X’ for the offense of X....
Consider Jones, who committed the offense of X, a performance which ended at time t1. Since () is true, so is (I).
(I) Jones deserves to suffer X’ and no more than X’ for committing X.
But can we assume that (II) is also true?
(II) Jones deserves to suffer X’ and no more than X’ at t1.
If Jones had committed other offenses before t1, (II) might be false. Let us assume however that, in fact, X is the only offense Jones committed in his whole life. Can we now assume that (II) is true?
Suppose that before committing X at t1, Jones had suffered X’ at time t0 in a natural disaster, e.g., a flood. Then, one might claim—and not incredibly—that Jones had paid his desert debt in advance. His preoffense ordeal of X’ at t0, nullifies his post-X desert debt of X’ at t1. Hence while he deserves to suffer X’ for committing X, he does not deserve to suffer X’ at t1. Although (I) is true, (II) is false.
Let us call this perspective on criminal desert, the whole life view. It may be objected that this view seems plainly misguided. Jones’ flood ordeal was both prior to his crime, and nonpenal, i.e., came about through natural causes. Hence that ordeal has no bearing on the penalty Jones deserves at t1, after he committed his crime. If this objection is sound, then criminal desert at a given time should be unaffected by two items:
(1) suffering prior to an offense
(2) nonpenal suffering
Let us consider the first item. Suppose that an immoral government penalized Jones by X’ at t0 for an offense he did not commit. Jones served an undeserved prison sentence of one year. A new and idealistic regime is installed and Jones proceeds to commit X at t1. Jones might inform these idealists that he deserves compensation for the undeserved penalty he suffered. But the standard to which compensation should, where possible, conform is restitution, i.e., restoring the equivalent of what was taken. At t0 Jones was undeservedly deprived of his freedom from incarceration for one year. Deserved restitution at t1 can only be made by giving Jones one year of freedom from incarceration, i.e., suspending his deserved one year prison sentence for committing X. Hence what Jones deserves at t1 is that he not endure his deserved X’ penalty for committing X. Thus he does not deserve X’ at t1.
To so compensate Jones would of course be moral madness. As a consequence any person punished undeservedly would earn the right to commit a crime. But remember that what reason forbids is treating Jones as he deserves at t1.
I conclude that criminal desert, at a given time, may be altered by item (1), i.e., suffering prior to the offense. But can such desert be affected by item (2), nonpenal suffering, e.g., an ordeal caused by some natural disaster.
Imagine a world W, quite different from our own. Only one sort of evil obtains in W. A class of persons, dubbed Fists, occasionally feels sadistic impulses which they vent on others. How? Fists have very powerful right fists which they use when so inclined to pummel others. Moreover, Fists are so strong and so swift that they can neither be controlled nor deterred in the slightest. A potential victim can only try to stay away from these Fists. But once assaulted the victim feels no desire for vengeance. Indeed, Fists aside, inhabitants of W are (by our standards) remarkably pacific and reasonable. They know that even if punishment of Fists were possible, no good would come of it. A Fist would not be affected in the slightest by punishment.
However when a Fist attacks, then as a causal consequence, he feels severe pain—just the degree he deserves—in his right hand. Thus in W, there are only two sorts of misery, undeserved pain inflicted by Fists and deserved pain which these Fists endure as a consequence. Since Fists receive their full measure of deserved pain through natural causes they deserve to suffer no more. Hence, both R and TR would agree that there are no occasions in W for penal intervention. (A1) as well as (D) are superfluous principles in W.
Reflection on W should serve to remind us that, from the perspective of criminal desert, punishment intervenes to accomplish what, in fact, could come about through perfectly natural causes. Hence, it is false that such nonpenal suffering cannot affect criminal desert.
If the whole life view of criminal desert is correct, then R and TR are in serious difficulty. Assessment of a criminal’s desert after an offense would require that one balance all of his moral wrongs against the suffering of his entire life. But such reckoning is usually beyond ordinary mortals. An omniscient deity could of course know all the wrongs a person has committed as well as the tribulations he has endured. However those who determine legal penalties are not omniscient. To rely on them for such life-spanning estimates of desert is plainly out of the question. In that case how can we be certain that when an offender is penalized, he does in fact deserve to be so treated?
Moreover, we may, I believe, safely assume the following: first, a large number of legal offenders have led miserable, deprived lives. Secondly, at least some, and possibly, a great many, such offenders, do not, when punished, deserve their ordeal. If we persist in legal punishment, for society’s benefit, we violate (HP), a principle of justice to which both R and TR are committed.
Let us now make a rash assumption. Only a penal system can prevent crime. No alternative can do the job. Should we then secure protection against crime by legal punishment, at the expense of (HP)?
How would our three principled philosophers view the matter?
Only U could accept the situation with equanimity. He, after all, had no commitment to (HP) in the first place.
R’s dilemma, however, would be most extreme. R, we recall, feels a double obligation: to punish deservedly (A1) and not to punish undeservedly (HP). But now it seems that both sorts of punishment arrive in one package, tagged, a human penal system. Moreover R did not expect his principles to conflict. He, unlike TR, is not the man for compromise.
I imagine that R would give up and wash his hands entirely of our penal system, i.e., one not directed by an omniscient deity. Perhaps he will spend his time dreaming of a possible world where an all knowing and righteous god rains deserved suffering-just the right amount-on those who fall from virtue.
And TR? He of course is not averse to compromise. For TR, the sacrifice of (HP), to avoid a massacre, was not out of the question. But would TR endorse a penal system where injustice may be an everyday matter? Or would he, like R, turn away from a reality so resistant to his moral principles.
I say TR would face up to moral fact and strain his principles to the utmost. Perhaps he would pose the issue in the following fashion. Many offenders, punished, would not deserve to be so treated. But many who would suffer at the hands of criminals, undeterred, would not deserve to be so victimized. In that case, TR may opt for legal punishment. But he can take little pleasure in his choice.
TR aspired to that best of all plausible moral worlds, one where both justice and utility receive their due. Only in some exceptional circumstance would justice be denied. But now TR knows that in a human penal system justice may be set aside, not on some rare and grim occasion, but with dreadful regularity. In that case he could only regard the practice of legal punishment as a most unhappy compromise.
1 This text appears originally as the preface of “Philosophical Perspectives on Punishment” 1972, SUNY press, pages xi-xxvii. This digital version is intended solely for the use of students in PHL 205. Please do not redistribute.
2 This sort of case is suggested by Alan Donagan. See "Is There a Credible Form of Utilitarianism? in Contemporary Utilitarianism, ed. M. Bayles, Anchor Books, 1968.