Part I
Retributions
[1]
VARIETIES OF RETRIBUTION
BY JOHN COTTINGHAM
One of the functions of philosophy is to combat ambiguity and muddle. Yet in discussions of the justification of punishment, philosophers persist in talking of āretributionā and āretributive theoryā as if these labels stood for something relatively simple and straightforward. The fact is that the term āretributiveā as used in philosophy has become so imprecise and multivocal that it is doubtful whether it any longer serves a useful purpose. In this paper I shall attempt to separate out no less than nine distinct approaches to punishmentāI shall call them ātheoriesā, but the term will be qualified laterāwhich are or have been labelled retributive. The first object of the exercise will be clarification for its own sake; but along the way I shall allow myself some brief observations on the logical adequacy of various strategies for justifying punishment.
Most dictionaries give the first meaning of āretributionā as repayment. The etymology of the term is clear: re+tribuo, Latin, to pay back. I am inclined to suggest that this notion encapsulates the basic or fundamental sense of āretributionā. Etymology, of course, is no arbiter in philosophy. But there is a perfectly ordinary (pre-philosophical) way of talking about punishment which exactly fits this original sense of āretributionā. Newspapers and public speakers often hold forth about the need for the criminal to repay his debt to society. And the argot of criminals (or at least the criminal of fiction) is full of expressions like āIāve paid the priceā:
Iām going straight, I am, straight as an arrer,
Iāve paid the price and done me time;
Iām going straight, and I donāt mean the straight and narrer,
āCos Iām going straight back to crime!
On our standard theory of retribution, then, punishment is inflicted in order to make the offender pay for his offence. Exactly how or why suffering something unpleasant (e.g., ādoing timeā for six months) should count as payment for an offence is left unexplained; but it has to be admitted that the notion is both ancient and widely held. In the Pauline doctrine of the atonement, the passion (suffering) of Christ is supposed to āpay forā the offences of mankind. (Note: this is merely an example of the belief that suffering pays for wrong; I do not mean to suggest that the repayment version of retributivism is really an atonement theory. There could not be an atonement theory of punishment; atonement is something voluntarily undertaken, punishment something exacted.)
A recent textbook defines āretributive punishmentā as āpunishment meted out because it is deservedā.1 Similarly, Ted Honderich in his book Punishment asserts quite bluntly that to give as a reason for the rightness of punishment that it is deserved by offenders is retributivism.2 Unlike our first theory, which is grounded in the original meaning of āretributionā, these accounts are initially puzzling. No doubt it may be true that many who call themselves retributivists have held that desert is a sufficient condition for just punishment; but to say, tout court, that this is what makes them retributivists leaves the rationale for the label compressed to the point of unintelligibility. Suppose I say: āIām a retributivist: I believe that where punishment is deserved this is sufficient to justify itā. I think the initial reaction of that ubiquitous figure the āintelligent laymanā would be: āWell, go on, explain! Where does the retribution come in?ā. To put the point more precisely, if someone claims that a sufficient condition for the justification of university scholarships is that they are deserved, does this mean that he has a retributive theory of scholarship-awarding? Perhaps it does; but only in a rather old-fashioned sense of the term āretributionā, meaning any kind of reward or recompense (ānever did a charitable act go away without the retribution of a blessingā3). If we are allowed to unpack and amplify the desert theory in this way, the claim becomes that punishment is justified because it is a deserved requital or reward for wrong-doing. Viewed as an exercise in justification, this account shares the curiously jejune quality of theory (1); and its detractors have made the complaintānot without forceāthat it reduces to the bald assertion that it is simply just that the offender should be punished.
Is there a logical connection between the desert account (2), and our basic account (1)? Thesis (2) is certainly compatible withāperhaps even entailed byāthesis (1). For it may be that to claim that punishment is a just retribution (repayment) implies that the offender deserves to pay. The converse entailment, however, does not seem to hold. For it seems that one could be a āretributivistā in the unexpanded Honderichian sense (punishment is justified because deserved) without subscribing to the idea that punishment is some kind of repayment of a debt. This alone is enough to warrant keeping theories (1) and (2) distinct.
Kant, universally regarded as a retributivist, asserts that the reason why punishment is imposed must always be because the individual on whom it is inflicted has committed a crime.4 A possible, though unlikely, interpretation of this remark is that the breaking of a law is, in itself, a sufficient condition for just punishment. This latter proposition was put forward, half a century ago, in a highly influential paper by John Mabbott. In connection with his own experiences in dealing with defaulters who had broken college rules, Mabbott wrote: āthey had broken a rule and they knew it and I knew it; nothing more was necessary to make punishment properā.5 Mabbottās theory seems to construe punishment as a kind of automatic penalty, whose appropriateness is ensured simply by the knowing commission of an offence.
It is not clear that this theory, as it stands, will serve as an account of the justification of punishment (and the author later made substantial modifications in his position to deal with this point). However, Mabbottās original unmodified theory deserves inclusion in our list for historical reasons. Many of those influenced by Mabbott took him to be putting forward not a moral thesis but a logical one. They took him to be claiming, as a necessary truth, that there is an essential āretrospectiveā feature to punishment; punishment, that is, must logically look backward to the past commission of an offence.6 Because Mabbott adopted the label āretributiveā to describe his theory (though he warned that he was not an āorthodox retributivistā), subsequent writers who wished to support this logical thesis tended to characterize their position asāat least partlyāretributivist. To be aware of this background is a vital prerequisite for following the argument of section (4) below.
Whatever the truth of the purely logical claim that proper punishment must necessarily involve reference to a past offence, it seems both arbitrary and inappropriate to signal this claim by using the label āretributiveā. There are many terms in the language of obligation whose correct understanding involves a necessary reference to a past act. āPromiseā is one of them. To understand why promises should be kept, it is, it seems to me, crucial to understand thatāwhatever the future benefits of keeping promises may beāthe obligatoriness relates to a past act: the giving of a promise. But in saying this have I advanced a āretributive theory of promisingā? Hardly! A retrospective or backward-looking element in a theory is never normally characterized as āretributiveā; it is only in the literature on punishment that these two notions are muddled (so that, in some discussions, āretributiveā becomes virtually an antonym of āteleologicalā).
āMost contemporary retributivistsā, asserts Martin Golding in his book Philosophy of Law, āmaintain a minimalist position. This holds that no one should be punished unless he is guilty of a crime and culpableā.7 Similarly, Anthony Quinton claims āthe fundamental thesis [of retributivists] must be that only the guilty are to be punishedā.8 Thesis (4), then, asserts that guilt is a necessary condition of just punishment.
As with our definition (2), some initial puzzlement is in order here. What on earth is supposed to be distinctly retributivist about thesis (4)? The most likely explanation for the classification of (4) as a kind of retributivism seems to be the historical one sketched in the previous section: even though Mabbottās thesis was principally about the logically sufficient, as opposed to necessary, conditions for just punishment, his adoption of the label āretributivistā paved the way for any theory asserting a logical connection between proper punishment and past offence to be classified as retributivist.
But whatever its historical genesis, the labelling of thesis (4) as a kind of retributive theory has sowed nothing but confusion. For it is clearly possible to hold thesis (4) as a principle of natural justice, or as a utilitarian secondary rule, without having any truck with the notion of retribution in the standard senses outlined under (1) and (2) above. A good example of the awkwardness that can arise in this connection is an otherwise exemplarily clear paper by H. L. A. Hart. Hart mounts a staunch and well constructed defence of thesis (4), but then proceeds to insist on labelling it the principle of āretribution-in-distributionā. This in spite of the fact that the route taken by Hart to defend (4) has, as he himself admits, nothing to do with retribution in its ordinary sense, but is concerned instead with principles of simple justice, and in particular the need to guarantee āa method of social control which maximizes individual freedomā.9
A further unfortunate consequence of calling thesis (4) āretributivistā, or mislabelling it the principle of āretribution-in-distributionā, has been the widespread view (hinted at by Hart) that some kind of compromise is possible between utilitarian and retributive accounts of punishment. The supposed compromise runs: utilitarian considerations provide the answer to the general question āwhy punish at all?ā, while āretributionā provides the answer to the more specific question āwhom shall we punish?ā (answer: āonly the guiltyā).10 To put the matter this way can be highly misleading, since, once again, if thesis (4) is held as a utilitarian secondary rule or as a liberty-maximizing principle, then of course no concession at all has necessarily been made to the strict notion of retributionāi.e., to the notion that punishment is due as repayment for a crime.
āA man is rightly punished because his punishment brings satisfaction to others.ā This thesis, which I label the satisfaction theory, has been described as a view which āhas given force to retributivismā.11 The idea behind the claim seems to be that there should be some kind of reciprocity between the sense of grievance felt by the victim of an offence and the satisfaction he gets from the suffering of the offender. What connection, if any, does this theory have with our ābasicā sense of retribution? The answer, I think, hinges on which of two possible interpretations is given to the theory.
(5a) On the first interpretation, the claim put forward is that it is intrinsically desirable or appropriate that grievances of victims should be matched by suffering of offenders. There is a close link here with retribution as repayment. If child A hits child B causing him pain and a sense of grievance, child B will frequently be heard to say, āIāll make you pay for that!ā. The payment is felt to have been exacted once B has inflicted a similar hurt on A. (Such beliefs are by no means confined to children, but children tend to make them more explicit.) Unfortunately for this version of the satisfaction theory, it is far from clear how the indubitable psychological facts just cited are capable of providing a satisfactory moral justification for the practice of punishment (unless we fall into the error commonly attributed to Mill and argue that what is desired is therefore desirable).
(5b) A second, and more sophisticated, version of the satisfaction theory is put forward by Justice Steven: āthe criminal law regulates sanctions and provides a satisfaction for the passion of revengeā.12 If the underlying idea here is that the penal system provides a substitute for private revenge, then it turns out that the focus of justification does not centre on the notion of retribution at all. Rather, we seem to be dealing with a utilitarian approach, where the penal system is justified as a mechanism for the prevention of vendettas, which furthers the goal of social stabilityāmaking society better ordered and more secure.
āFailure to punish is unfair to those who practise self-restraint and respect the rights of others.ā13 ā[Retributivism proposes that] a manās punishment is justified or obligatory because ā¦ unlike non-offenders he has gained satisfactions attendant on the commission of an offence.ā14 These quotations are strongly reminiscent of John Rawlsā ājustice-as-fairnessā defence of political obligation. The citizen, according to Rawls, has a duty to obey laws of which he disapproves because it is unfair to his fellow citizens if he voluntarily accepts the benefits of the social system without being prepared to shoulder his share of its burdens.15 The parallel suggestion, in the sphere of punishment, seems to be that the state is justified in punishing the offender on behalf of the good because otherwise he would be gaining an unfair advantage: he would be profiting from his own selfish refusal to play fair and respect the rights of his fellow citizens.
Is it correct to describe the fair-play theory as āretributiveā in our basic sense? Only, I think, indirectly. The object of punishment, on theory (6), is the preservation of justice and the maintenance of fair play. The immediate focus of justification thus centres not on the offender but on the law-abiding citizen and the duties owed to him. It is not that punishing the offender is intrinsically appropriate, b...