RETRIBUTIVISM AND DESERT
189
RETRIBUTIVISM AND DESERT RUSS SHAFER-LANDAU Abstract: Retributivists claim that the point of legal punishment, and the standard that ought to govern the construction of penal institutions, practices and rules, is that the guilty must be treated in the way that they morally deserve to be. I construct an argument from elimination designed to show that there is no plausible index for measuring moral desert, and thus that a central plank of the retributivist platform is indefensible. I conclude by showing how deeply held intuitions that appear to support retributivism might be retained without embracing retributivism.
1. Retributivists claim that the point of legal punishment, and the standard that ought to govern the construction of penal institutions, practices and rules, is that the guilty be given their just deserts. Punishment must be commensurate with moral desert.1 This emphasis on moral desert straightforwardly implies the following view about the content of sentencing guidelines: such guidelines are morally justified if and only if, and because, they assign a punishment that is within the range of sanctions that a convicted offender morally deserves. Call this the commensurability thesis. Because the commensurability thesis is just a starchy way of expressing what so many take to be clearly true – that wrongdoers ought to be given their just deserts – it pays to briefly note the breadth of current practices that fall afoul of its requirements. If we must punish convicted offenders just as much as they morally deserve, then we would have to do away with most instances of suspended sentences, reduced sentences, and other applications of judicial mercy. We would also have to eliminate the possibility of pardons, amnesties and executive clemency. Statutes of limitations would have to be abandoned, since they specify conditions that disable the government from prosecuting, much less punishing, those Pacific Philosophical Quarterly 81 (2000) 189–214 0279–0750/00/0100–0000 © 2000 University of Southern California and Blackwell Publishers Ltd. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.
189 Publishers Ltd. © 2000 University of Southern California and Blackwell
190
PACIFIC PHILOSOPHICAL QUARTERLY
known or believed to have committed crimes. Gone too would be many evidentiary rules that force the dismissal of improperly acquired evidence, even though that evidence would be sufficient to convict a guilty person and so subject him to deserved punishment. Double jeopardy provisions would be expunged, since they prevent the state from prosecuting and so punishing some who are known to have broken criminal laws. The presence of these various legal rules reveals our attachment to values other than that of meting out just deserts. Of course, the retributivist could demand widespread changes that would eliminate these rules from our practice, or insist that these rules, contrary to appearances, can be accommodated within a system that is regulated by the commensurability thesis. Rather than focus on the plausibility of such replies, I am going to concentrate here on what seems to me more worrying – the assumption that there is some punishment, or range of punishments, that an offender morally deserves for his crime. I do not believe that we can make sense of commensurating punishment with moral desert. If we can’t, then the commensurability thesis is false. And if the commensurability thesis is false, so too is retributivism.
2. Sometimes when we want to know what a person deserves for his wrongdoing, our answer is properly given by identifying that punishment called for by the institutional rules of our criminal justice system. In such a case, all it means to say that a person deserves punishment P is that P has been authorized by a jurisdiction’s sentencing rules. But we may want to know whether these rules are themselves morally justified. The commensurability thesis supplies an answer: a sentencing rule is justified if and only if (and because) it recommends punishments that are morally deserved. Endorsement of the commensurability thesis is necessary, but not sufficient, to qualify as a retributivist. That’s because those who favor a paternalistic, or moral education, theory of punishment also insist on giving offenders what they morally deserve. The disagreement lies in whether wrongdoers deserve to suffer, or deserve to be benefitted in a quite specific way, viz., by being offered the opportunity to morally reform themselves.2 Because of this disagreement, I will take the specifically retributivist thesis about sentencing to be the endorsement of the commensurability thesis, combined with the view that a wrongdoer’s just deserts consist essentially in the imposition of hard treatment or suffering. Thus the distinctive retributivist thesis about sentencing guidelines is that such rules are morally justified if and only if, and because, they assign just the amount and kind of suffering that the wrongdoer morally deserves.3 © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
191
This understanding of retributivism opens up two avenues for possible criticism. The first challenges the idea that wrongdoers deserve to suffer. Wrongdoers might deserve something other than suffering, or they might deserve nothing at all. Retributivists need to provide some positive argument for the claim that wrongdoers deserve to suffer. I don’t believe that they have succeeded in doing this. But I will assume for present purposes that such an argument can be supplied. If we assume that criminals deserve to suffer, we will still want to know how much they deserve to suffer, and in what way. I don’t believe that we can ever know this. My suspicion is that we are in the dark on this matter because there is no determinate kind or amount of suffering that a criminal morally deserves. This proposed explanation of our ignorance is what we can call nihilism about moral desert: the view that there is no fact of the matter about what sanction(s) a wrongdoer morally deserves for his offense. If nihilism about moral desert is true, and if the commensurability thesis is true, then no legal punishments are morally justified. The commensurability thesis conditions the moral justification of punishment on its giving an offender just what he morally deserves. If there is no such thing, then punishment cannot be morally justified. That is a conclusion I’d like to come to only as a last resort. I plan instead to defend nihilism about moral desert, and use that as a means of applying pressure on the commensurability thesis.
3. If the commensurability thesis is true, and if legal punishment can be morally justified, then there must be some sanction or range of sanctions that a criminal morally deserves for his conduct. But consider some standard cases and see the difficulty for yourself: how much suffering is morally deserved for one who impersonates an officer, or counterfeits currency, or hijacks an airplane, or batters a child? Does the impersonator deserve a few weeks, a year or a couple of years in jail? It would be unfair to claim that the commensurability thesis is true only if there is, for some crimes, a precise number of days in jail that was the deserved legal sanction. There is nothing in principle problematic about allowing for a range of sanctions that “fit the crime.”4 But once we concede this, there still does not appear to be any way to know whether the impersonator morally deserves eighty, eight hundred or eight thousand days behind bars, or even whether some amount of jail time is the appropriate kind of punishment to impose in the first place. We can allow for some indeterminacy in the sentencing correlations, but at some point we must ask whether moral desert is giving us any guidance at all. Perhaps the greatest obstacle to developing this line of criticism is making sense of the broad consensus on the principle that we ought not © 2000 University of Southern California and Blackwell Publishers Ltd.
192
PACIFIC PHILOSOPHICAL QUARTERLY
punish people more than they deserve. This consensus isn’t grounded just on the thought that we ought not punish people more than they legally deserve. After all, authorized sentencing laws in certain jurisdictions allow for vicarious liability, or for twenty year prison terms for blasphemy. Individuals convicted under such statutes are being punished as the law allows, and so are receiving their legal deserts; yet the law is unjust, because it licenses treatment that (we think) is far harsher than what is morally deserved. Saying such a thing implies that there is something that a person morally deserves, and that the wrong consists precisely in exceeding the bounds of morally deserved treatment. And this implies that nihilism about moral desert is false. So we have a nice puzzle. On the one hand, there is the difficulty of showing how we can commensurate punishment with moral desert. On the other hand, many plausible judgments about cases seem to imply that there are morally deserved punishments. Since nihilists are skeptical about the prospect of commensurating punishment and moral desert, it seems that they may be forced to abandon deeply-held convictions about the injustice of (e.g.) vicarious penal liability. To get a sense of the difficulties here, imagine that we had to decide on how to reward do-gooders, rather than punish wrongdoers. We might well agree that the virtuous deserve to prosper, and the wicked deserve just the reverse. But setting out to give the good just what (no more or less than) they morally deserve – how could such a thing be done?5 Admittedly, there are many practices whose constitutive rules set criteria for generating determinate desert claims. Grading contexts, competitions and games are governed by rules that, among other things, supply precise guidance about who deserves what. The team that scored the most goals deserves to win; the pianist who won the competition deserves the announced $10,000 grand prize; the student who failed to turn in any assignments deserves an “F.” But in all such contexts, the rules are not justified on the basis of apportioning deserts that are specifiable prior to the existence of the rules. Instead, the rules themselves define what is to constitute just deserts. No one deserves $10,000 for being a fine piano player unless and until some competition rule announces eligibility conditions and the amount of the award. Desert is an artefact of the rules, rather than an antecedently existing relation that desert rules may succeed or fail to capture. This picture is incompatible with the traditional, but ordinarily unstated, retributivist assumption that sentencing guidelines are morally justified only in so far as they accurately capture moral desert claims that antedate the guidelines themselves. Retributivists, as we’ll see, may stray from this the traditional view, and so make the apportionment of just deserts fairly closely analogous to the distribution of competitive awards. In doing this they must frankly acknowledge that desert awards are by-products of © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
193
practical and institutional rules rather than the justificatory basis for them. This generates various unattractive implications, which I develop in the next section. Alternatively, retributivists may cleave to the time-honored path of justifying legal sanctions by demonstrating their commensurability with pre-existing moral deserts. On this line, retributivists must show how there are specific answers to questions of what a wrongdoer morally deserves, while abandoning the familiar models (of prizes, grades and competitive awards) that manage to supply these sorts of determinate answers. The argument that follows is my effort to show why this problem has no solution.
4. A. LEX TALIONIS
The natural way to attempt to commensurate punishment and moral desert is to adopt some form of lex talionis. The Mosaic version of lex (an eye for an eye, a tooth for a tooth) is best known, and stands for the view that wrongdoers morally deserve just the sort of treatment that they have imposed on their victims. But any such view encounters three familiar difficulties.6 First, lex is conceptually impossible to apply in many cases. Some crimes lack victims (e.g., failed criminal attempts, many instances of reckless endangerment), and others have only nominal victims (e.g., tax fraud, insider trading and various free-riding offenses, where the harm suffered by any given individual is either nonexistent or negligible). And many crimes have victims who have suffered determinate, non-negligible harm but whose harm cannot be requited by lex (e.g., that inflicted by a childless kidnapper, a propertyless arsonist, or the fabled blind man who blinds another). Further, even if we restrict ourselves to crimes that do impose determinate harms, the same harm can be inflicted by agents with very different mens rea. Differences in mens rea are directly relevant to culpability, and yet are not capturable by sentencing rules that determine punishment solely by the harms suffered by a victim. Finally, in many cases where lex does offer concrete advice, many of the recommendations are morally unsavory – raping a rapist, or torturing a torturer, for instance. Some may be prepared to bite the bullet, and insist that such treatment is what these criminals deserve. Even if it is, we surely do not want such desert verdicts used as a basis for structuring sentencing guidelines. Better to knowingly fail to mete out such deserts than to authorize a line in the budget for an official rapist and torturer. Efforts to craft a more acceptable principle by insisting that a criminal deserves only the same amount of suffering, rather than just the same kind of treatment, are bound to fail.7 An equal suffering principle is prey © 2000 University of Southern California and Blackwell Publishers Ltd.
194
PACIFIC PHILOSOPHICAL QUARTERLY
to both of the first two objections to the identical treatment version of lex, and perhaps the third as well. And it introduces the complexity, surely practical but also theoretical, of identifying a measure by which to make interpersonal comparisons of suffering. Robert Nozick has attempted to deal with two of the major obstacles to lex in his formulation of a desert measure.8 According to Nozick, one’s deserved punishment is a product of the wrongness of one’s (intended) act, multiplied by the responsibility one bears for the wrong. Nozick summarizes his view with the formula: r × H = deserved punishment (where “r” stands for responsibility and “H” for wrongness). Nozick’s understanding of responsibility is not causal, but moral, and this allows him to introduce considerations of mens rea into determinations of just deserts.9 And Nozick agrees that not every wrong need involve a harm to determinate individuals, thus avoiding an obstacle that confronts classic formulations of lex.10 Without a wrong, or without any mens rea at all, no punishment is deserved. And the amount of deserved punishment increases with the degree of responsibility and the degree of wrongness exemplified in one’s conduct. Both implications are attractive. But there are two problems at the heart of the Nozick’s scheme. The first is that, in most cases, there is no way to specify values for the relevant variables, r and H. The second is that, for the remaining cases in which values can be supplied, the advice given by Nozick’s formula is as unattractive as that offered by earlier verions of lex. For Nozick, r = 1 for all intentional crimes that lack excuses. Nonintentional (i.e., merely foreseen, reckless or negligent) offenses have an r-value of something less than 1. Nozick gives us no idea of how to quantify this diminished moral responsibility. We can plausibly say that intentional wrongs are worse than those that are merely foreseen, which in turn are worse than reckless and negligent wrongs, respectively. But until we can know just how much worse one is than another, we have no prospects for doing any moral arithmetic. The same kind of worry – that of knowing how to quantify values of the relevant variable – presents itself when it comes to wrongness (H). Nozick tells us that we may identify H with “the maximum amount of disutility the victim could reasonably have expected to undergo.”11 But this holds only for cases in which the wrong is centrally a harm to a determinate victim. It gives no advice for other wrongs. Even where wrongs are harms, Nozick doesn’t tell us how to measure disutility. Perhaps a natural first step is to envision disutility in terms of dollars. But this would be misguided, since presumably we don’t want to restrict punishments to monetary fines. Nozick himself, though a fan of the r × H formula for assigning compensation claims in civil cases, recognizes that the formula will have to be used differently for criminal matters.12 Since that is so, we should sometimes expect something other than dollar figures © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
195
to represent the measure of wrongness when punishment is called for. Nozick doesn’t tell us what this alternative measure might be. With the exception of one type of case, these problems prevent Nozick from offering determinate advice about how much or what kind of punishment an offender deserves to receive. The exception arises when the offender is fully responsible for his misdeed and the wrong is exhausted by a determinate harm to an individual. In this case, however, determinacy is purchased at the cost of plausibility. For in such a case, Nozick’s formula is extensionally equivalent to classic versions of lex, and so inherits their difficulties. When we intend to make another suffer, and do, then we are to be made to suffer just as much. When we don’t intend this, but do wrongly impose suffering anyway, we should be made to suffer less. How much less? This is determined by a discounting factor that directly tracks diminished mens rea. As we’ve seen, Nozick gives us no idea what such a factor would look like, and so no idea of what the relevant sorts of calculations would look like. And in cases where the offender is fully responsible for his wrongdoing, but the wrong is not a harm, or is a harm distributed broadly and negligibly across many individuals, Nozick’s formula again fails to provide determinate guidance for identifying a deserved punishment. In the worst case scenario, where a nonintentional offense involves something other than, or in addition to, harm to a determinate individual, Nozick’s formula collapses entirely: we have no way to assign numbers to either of the relevant variables, r or H, and so no idea, even in principle, of how to identify a deserved punishment. What so many have found so appealing about lex is the prominence of the victim’s harm in paradigmatic cases of criminal offense, coupled with the intuition that the harm itself must serve as the justificatory basis for structuring a penal response. Fans of lex must therefore identify some central class of harms whose commission should be “matched” by punishments that mirror their form. Jean Hampton takes this view when developing her retributive theory of punishment.13 Hampton embraces the standard retributivist view that only harms that are also moral wrongs can serve as the basis of deserved punishment.14 What distinguishes Hampton’s view is the way she understands such wrongs – as instances of conduct that demeans its victim, thereby expressing a message of the victim’s inferior status. She takes punishment to be justified as a way of nullifying this message of inferiority, of asserting the moral truth of the equality of worth of victim and offender.15 Thus when we ask how much and what kind of punishment is deserved, the answer is: just that amount and kind necessary to ensure that the moral message of equality is communicated. Hampton’s proposal shares with other versions of lex the difficulty of accounting for punishment for offenses that fail to have victims, or have them in only an attenuated sense (e.g., money laundering, defacing public monuments, illegally importing animals across state or federal lines). It © 2000 University of Southern California and Blackwell Publishers Ltd.
196
PACIFIC PHILOSOPHICAL QUARTERLY
assumes that crimes with victims are instances of demeaning conduct, and that the central wrong to be requited in such cases is that of demeaning someone. This may be true of some crimes – rape is perhaps the best example. But many kinds of action that are properly criminalizable, and which do have victims, are not instances of demeaning behavior. The thief, the embezzler, the criminally negligent purveyor of goods, may sometimes feel a kind of moral superiority to their victims, but this is not essential to the offenses they commit, and isn’t the central wrong that we should be trying to redress. Finally, and perhaps most importantly, it isn’t clear why punishment is necessary to effect the desired communication. Hampton offers the following explanation: Suppose we gave the victim a ticker-tape parade following the crime to express our commitment to his value. Still the fact that he had been mastered by the wrongdoer would stand. He would have lost to her, and no matter how much the community might contend that he was not her inferior, the loss counts as evidence that he is. Hence the victim wants the evidence nullified and punishment is the best way to do that. The evidence isn’t made to vanish by punishment, nor made never to have existed. Instead it is explained away.16 (Emphasis in the original.)
But some victims may not in fact want the evidence nullified (they may believe in their inferiority, they may be indifferent to others’ opinion of them, or they may be a victim of a crime in which the message of inferiority hasn’t been sent). Nor is it clear, even when they do want such a thing, why their wishes should be determinative. Criminal trials debar private citizens from being litigants, and the government prosecutors file charges in the name of the state (or country) rather than on behalf of private individuals. The wishes of private citizens are rarely determinative in criminal proceedings. Even if we should defer to the wants of the victim, it remains unclear why the deliberate infliction of hard treatment on an offender is necessary to explain away the relevant message of inferiority. Why couldn’t a very forcefully worded, publicly promulgated message from the bench be sufficient to do this? Hampton’s passage doesn’t help us here. What needs explaining is why a convict must be intentionally made to suffer. Hampton’s explanation must take the form of showing how such an imposition is necessary to send the relevant moral message. But she hasn’t shown this; she hasn’t shown how one elevates the ( perceived) status of the victim by diminishing the status of the convict. There is (at best) only a contingent relation between the avowed point of punishment – the erasure of a message of inferiority – and the intentional infliction of hard treatment on an offender. On Hampton’s view, the central justification of punishment is communicative. But we’ve yet to see a good argument for thinking that her © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
197
particular sort of communication is centrally relevant to justified punishment, and, even if it is, why punishment is necessary to achieve the intended communication.17 Lex’s last gasp may have been issued by Jeremy Waldron, who, in a recent article, has crafted a modified version that seeks to escape from traditional criticisms.18 According to Waldron, lex talionis is “the principle that the action visited as punishment upon a criminal should be the same as the action that constituted his offense.”19 A punishment is the same (type of ) act as the offense if the punishment exemplifies the same wrong-making features as the offense. (For morally justified punishments, of course, exemplifying such features will not lead to their being wrong.) Which features are the wrong-making features is decided by reference to supplemental normative theories – lex talionis is, by itself, neutral on that matter. Waldron accepts Rawls’ and Hart’s view that the questions of general justifying aim and the distribution of punishment are distinct.20 This enables Waldron to defend the initially surprising claim that lex, far from being a distinctively retributivist principle, is compatible with all major general justifying aims. Treating an offender in the same way he treated his victim may be a good deterrent, it may be a good way to engender moral reformation, and it may be a good way to ensure the kind of equality and fair treatment that Kant insisted upon.21 Because of his ecumenical approach, Waldron does not offer an argument specifically for the claim that being treated as lex demands is what an offender morally deserves. For all Waldron says, lex may be a justified sentencing principle even if it fails to specify just deserts. Yet even if it turns out that lex does specify just deserts, Waldron admits that, at its best, lex represents only “a rather good idea,” rather than a categorical imperative or complete principle for determining just sentences.22 In cases of rape, for instance, Waldron believes that the relevant wrong-making features will indeed make reference to the sexual nature of the violation (as opposed to a more general description of the wrong as a battery or an assault). Applying lex thus calls for rapists to be sexually violated. (Not a good idea, according to Waldron.) As Waldron sees it, lex will also sometimes recommend other forms of corporal punishment, though this would lend itself to the corruption of penal officials and so ought not to be administered. And the wrong-making feature of murder may well be that it ends another’s life, but capital punishment is, according to Waldron, nevertheless unjustified.23 In taking these stands, Waldron admits that other moral principles can be properly introduced to temper lex, and so admits that lex imposes only a prima facie moral requirement. If lex is the principle that specifies just deserts, the tempering of lex is a disavowal of the commensurability thesis. For if lex’s edicts are sometimes © 2000 University of Southern California and Blackwell Publishers Ltd.
198
PACIFIC PHILOSOPHICAL QUARTERLY
morally repugnant, then sentencing guidelines that enshrine its edicts are not necessarily morally justified. Waldron’s view thus implies that even if lex is our best desert principle, it isn’t always morally justified to mete out just deserts.24 Thus if lex does specify an offender’s just deserts, and Waldron’s moral misgivings about raping rapists, flogging batterers or executing murderers are correct, then the commensurability thesis is false. Unless one is prepared to argue for the acceptability of state-sponsored rapes, floggings, etc., one is faced with a choice of abandoning lex or abandoning the commensurability thesis. Before we do the latter, as I’ll ultimately recommend, we must do the former, and proceed to examine alternative formulae for assigning morally deserved punishments. B.
DEMOCRACY
What we might call the democratic proposal says that a sentencing guideline is morally justified if and only if, and because, it has been ratified by a majority of duly elected representatives of qualified voters, and survives the relevant battery of legislative defeaters (executive veto, judicial review, etc.). A sanction is morally deserved if and only if it is called for by a sentencing guideline enacted in this way. The democratic proposal will be relativistic, since majorities in different legislatures will enact different sentencing guidelines. Many will think of this as a plus, since many think that the punishments criminals deserve should be determined by such things as existing conventions and indigenous social expectations about what constitutes appropriate behavior and sanctions. Further, and perhaps more importantly, the democratic proposal may appeal because it abandons the view, embedded in lex, that moral desert can be determined prior to legislative enactments. A central intuition that supports lex – the idea that there is an intrinsically fitting relationship between certain wrongs and certain sanctions – is here abandoned, since the democratic view makes moral desert an artefact of legal desert. We don’t begin by identifying a wrongdoer’s moral deserts and then craft a law to match them. Instead, moral deserts are by-products of the law. Thus, democratic retributivists are barred from taking the familiar justificatory route that vindicates sentencing proposals by citing their accurate match with pre-existing moral deserts. According to the democratic proposal, there isn’t anything a wrongdoer morally deserves for his crime until a legislature specifies a sanction. And once a legislature does so, the wrongdoer legally (hence morally) deserves the specified punishment. In effect, the criminal is being treated as he morally deserves to be, and hence has no complaint, if he is treated as the (democratically enacted) law allows. The democratic proposal thus supplies a neat justification of the commensurability thesis. According to the proposal, justified sentencing rules © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
199
will invariably match a wrongdoer’s just deserts, because his just deserts are defined by the legal rules themselves. Given the difficulties faced by lex, abandoning one of its crucial supporting ideas may not be such a bad thing. But the basic intuition underlying the democratic proposal – the idea that a democratically enacted sentencing guideline is necessary and sufficient to identify what offenders morally deserve – fares no better. For even democratically enacted legislation can be wide of the moral mark. We rightly reject the idea that legislatures are the ultimate source of moral authority. We entrust certain decisions to legislators in part because of our confidence that they will be able to implement views which, prior to their being enacted, are morally right and seen to be such by the elected officials we endorse. There is no conceptual connection between a rule’s moral content and its having been democratically enacted. There are two kinds of response available to advocates of the democratic proposal. The first narrows the scope of the proposal so that democratic legislative enactments specify moral deserts for only a certain well-defined range of cases. The second imports supplemental constraints on democratic decision-making in order to increase a legislature’s chances of making morally accurate recommendations. Neither move succeeds. The first strategy proceeds by drawing attention to the distinction between offenses that involve actions wrong in themselves (mala in se), and those that are wrong because they are illegal (mala prohibita). The best that a legislature can do with regard to actions of the first type is to recognize their true nature, and to respond by prohibiting the worst of them. Their wrongness doesn’t depend on legislative recognition, and so presumably the morally fitting response to their commission is likewise independent of such recognition. But it seems appropriate to take a different tack when dealing with offenses mala prohibita. If certain kinds of conduct become wrong through legislative enactment, then it seems equally within the legislature’s moral purview to create a scheme for responding to such wrongs. The democratic proposal may work for mala prohibita offenses, even if it fails for those that are mala in se. Such a move would still leave us wondering how to account for just deserts for offenses mala in se, and wondering whether the emerging hybrid account would be internally consistent.25 Setting this worry aside, we can assess the democratic proposal for mala prohibita offenses only if we’re clear about what such offenses amount to. There are at least three relevant possibilities: Conduct is malum prohibitum = df. (i) it is illegal. (ii) it is illegal, and it prohibits intrinsically morally neutral activities. (iii) it is illegal, it prohibits intrinsically morally neutral activities, and does so in the service of some genuinely valuable end. © 2000 University of Southern California and Blackwell Publishers Ltd.
200
PACIFIC PHILOSOPHICAL QUARTERLY
Only the third definition will do; neither the first or second offer sufficient conditions of conduct that morally deserves to be punished. Some laws proscribe morally good behavior, and others require morally bad behavior. Since those who break such laws aren’t necessarily morally deserving of punishment, (i) fails. Nor is it always immoral to break laws prohibiting intrinsically morally neutral activities, since some such laws are drafted for bad moral reasons. Laws that require minorities to sit at the back of a bus, or forbid them from sitting at a lunch counter, are laws that prohibit activities that, in themselves, are relevantly neutral. But the rationale and the effect of such laws is immoral, and so again, presumably their offenders are not morally deserving of punishment. Thus (ii) also fails as a satisfactory definition. Option (iii) handles the problems that beset (i) and (ii), and should suffice for now. The democratic proposal under scrutiny claims that for mala prohibita offenses thus construed, a sanction is morally deserved just if the sanction is permitted by a democratically enacted sentencing guideline. But even here, democratic enactment seems insufficient, and perhaps unnecessary, for specifying moral deserts. It won’t be necessary if there are other, non-democratic forms of justified political organization, whose lawmakers specify sanctions for mala prohibita offenses. And it won’t be sufficient, because even if democratically elected legislatures are within their moral rights to prohibit some class of intrinsically neutral conduct, they may be morally mistaken about how to respond to such offenses. If the democratic proposal is true, then such mistakes are conceptually impossible – offenders morally deserve just what the legislature says they do; neither more nor less. Yet such legislative decisions may be made in haste, or in ignorance of relevant facts, or with the motive of assuaging extremists in the party whose support one needs for re-election, etc. All sorts of morally irrelevant or positively morally corrupting factors may play a role in determining the strategy of a legislator called to draft or vote on a sentencing guideline. It would be awfully surprising if, given this, a simple majority’s approval of a sentencing guideline were sufficient to fix moral deserts. The natural move to make on behalf of the democratic proposal is to elevate the procedural requirements that legislators have to satisfy before their laws fix moral deserts. And by insisting on enhanced procedural constraints, we can generate an integrated democratic defense, one that covers mala in se as well as mala prohibita offenses. Democratic enactment is not enough to fix moral deserts – perhaps (as a start) only rules that are ratified by a legislative majority that is sufficiently informed and nonpartisan are fit to do so. No matter the increased standards we select, there is the possibility that no denumerable set of procedures can guarantee morally accurate results. Enhanced procedural constraints might get us increasingly closer approximations of the moral © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
201
truth without guaranteeing that we hit the target every time. If that were so, then no version of the democratic proposal could succeed, no matter how elevated the constraints on majoritarian decision-making. Any link between democratically enacted legislation and genuine moral deserts would be contingent. I don’t want to settle this large issue here. Perhaps there are constraints on legislative majorities that can serve to ensure that their edicts are morally satisfactory. But insisting that such standards be met is tantamount to abandoning the claim that democratic passage is a necessary or sufficient condition of morally justified sentencing guidelines. Democratic passage would be, from a justificatory point of view, entirely irrelevant. The constraints that are to ensure a morally appropriate outcome would be supplying the moral justification, and it really wouldn’t matter whether a legislative majority enacted the outcome or not. For instance, what would confer moral justification on a sentencing guideline would be its being the object of (say) impartial, informed choice, rather than its being voted for by at least 51 percent of the legislature. If moral deserts are fixed only after those doing the fixing are fully informed and impartial, then why must it be legislators whose decisions do the fixing? Why not any actual or hypothetical figure(s) sufficiently well-informed and dispassionate? Once we begin importing elevated constraints on the determination of the choices that fix moral deserts, there appears to be no reason for insisting that the moral status of a sentencing guideline (or anything else) must be fixed by majoritarian legislative rule, rather than by reference to the choice or opinion of any individuals satisfying the selected constraints on choice. Thus the democratic element of the democratic proposal falls away. Democrats are faced with an unhappy choice. The original proposal fails to specify sufficient conditions for moral desert. Any amended democratic proposal stringent enough to supply sufficient conditions for moral desert will not supply necessary conditions. Thus even a radically revised democratic proposal will fail to supply a plausible index for what individuals morally deserve. C. PROPORTIONALITY
A different solution to the problem of specifying what wrongdoers morally deserve is given by a principle of proportionality. On this line what criminals deserve is just to be treated in such a way as to satisfy the proportionality thesis: graver crimes ought to be met with harsher sentences. Like the democratic proposal, the proportionality suggestion gives up the idea of a deserved sanction that antedates the creation of a system of sentencing guidelines. Proportionality theorists cannot justify a sentencing guideline by showing that it accurately gives an offender just what he morally deserves, where such deserts are thought to precede the creation of a law and to serve as its justificatory basis. © 2000 University of Southern California and Blackwell Publishers Ltd.
202
PACIFIC PHILOSOPHICAL QUARTERLY
This does not impugn the proportionality proposal, since we haven’t seen any reason to abide by the traditional understanding of retributivism that fixes moral deserts prior to legal ones. The real problem is that proportionality is conceptually distinct from commensurability, and thus it would be (at best) fortuitous if the demands of proportionality and commensurability lined up at all tidily. Commensurability may entail proportionality, but not vice versa. Proportionality is a fundamentally comparative notion – whether a punishment is proportional to a crime depends on how much other crimes are being punished. By contrast, commensurability is a fundamentally noncomparative notion – a punishment is commensurate with a crime if and only if it is fitting or appropriate, where this is a matter to be settled without regard to the sanctions meted out to other crimes. To see how these two considerations might pull apart, we can imagine a jurisdiction that has identified just twenty types of criminal conduct, and has ranged them from least to most serious, and has aligned them with twenty punishments of increasing severity. Proportionality is thus satisfied. The least worst crime in this jurisdiction is littering. The mildest punishment is forty years in jail. Or the harshest punishment is 20 days in jail, and the worst crime is mass murder. Readers are invited to supply their own variations on this familiar theme. The upshot is that it is implausible to suppose that a system of punishment satisfies the commensurability thesis just because it satisfies the proportionality thesis. Criminals may deserve to suffer punishments that are proportional in the relevant sense, but then again, as the examples showed, they may not – forty years in jail is too much for a litterer, and twenty days too little for a murderer, even though these sanctions may figure in a sentencing code that is perfectly proportional. Proportionality does not supply an appropriate index for commensurability. A variation on the proportionality theme is offered by John Kleinig, who recognizes the need to identify an anchor point that would fix a noncomparative desert assignment from which all other assignments might follow.26 What we need, having in hand an ordinal scale of offenses and a similar scale of punishments, is some way to correlate the two so as to avoid arbitrariness. Kleinig suggests we do this by assigning the least serious offense to the mildest morally justified punishment and the gravest offense to the harshest morally justified punishment, scaling up and down from these two fixed points. There is no pretense to absolute determinacy, but this does allow for a range of sanctions to fit the crime, and that’s all that is needed to avoid nihilism about moral desert. Kleinig’s suggestion faces three difficulties, two of which he recognizes but does not think damaging. The first two problems have to do with construction of the scales of offense and punishment. Kleinig recognizes that there will be difficulties in determining for every given offense whether © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
203
it really is morally worse than another, and whether particular punishments are harsher than others, but he chalks this up to practical difficulties, and sees no principled obstacle to the construction of the scales.27 Yet incommensurability arises even within the scales of offense and punishments, though perhaps more obviously in the former case. Joel Feinberg has wonderfully retailed the various factors that must be taken into account when making determinations of criminal culpability: A great number and variety of factors go into the determination, whether we are talking now of criminal sentencing guides or the moral judgments they partially incorporate. A sound if blurred insight is that the harm intended is much more important an indicator of an offender’s desert than the harm actually caused. Far more useful, however, than the concept of intentionality, are the four “culpability conditions” first proclaimed in the Model Penal Code – acting purposely, knowingly, recklessly, or negligently in regard to some harmful result. Then, of course, the concept of a motive, ruthlessly kept out of the original criminal trial, forces its way back at the sentencing stage, and contributes its flavor to the emerging blameworthiness stew. Did the offender act cruelly? spitefully? from mercenary motives? out of greed? in an emotional explosion provoked by sexual rejection? or by sexual jealousy? or through political conviction? out of mercy or compassion for another’s suffering? after forethought and deliberation? out of conscientious conviction or the determination to do at any cost to oneself what one sincerely believes to be one’s moral duty? out of sudden violent impulse as mysterious to the actor as to anyone else? And after all those questions have been answered, and provocation considered, and other types of mitigation, and diminished responsibility, and the questions they raise, now at last the more traditional justifications and excuses enter the arena with their talk of subtle coercion and mistakes, mistakes of fact and mistakes of law, defenses that undermine and defenses that affirm, defense of self and defense of others, with duress and necessity, involuntary and voluntary intoxication, insanity and short of insanity a host of neuroses and psychoses, compulsions and obsessions, and the great parade of syndromes, and on and on. There is surprisingly little disagreement about the factors that belong on this list, but much disagreement about the weight to be given different factors when they conflict.28
Exactly. Not only has Feinberg got the contents of the list correct, but his concluding remark is on target as well. He has identified a problem that no one has yet solved – how, even in principle, we are to balance the several culpability factors that to all appearances seem incommensurable with one another. True, this is a problem not only for retributivists, but for anyone who wants to ensure proportionality between punishment and moral culpability. But it is no less a problem for its being widely shared. There is no common currency among considerations of harm, motive, the culpability conditions, the several allowable defenses, excuses and justifications, that should give us confidence in there being, in principle, a way to determine their relative weight in conflicts and so order offenses in terms of their increasing moral seriousness. And so too with punishments, though here the problem appears less acute because we have fewer variables to deal with. There are only a few © 2000 University of Southern California and Blackwell Publishers Ltd.
204
PACIFIC PHILOSOPHICAL QUARTERLY
forms that punishment can take: fines, incarceration, community service, public shaming, corporal punishment, exile. The last three aren’t much in favor these days. So suppose we restrict ourselves just to fines, jail time and public service. We can pretty easily form a scale of severity for each punishment – we measure the harshness of incarceration by its duration, the severity of a fine by its amount, the harshness of community service by its duration (and perhaps its type, though here things get murkier). But how to align these scales? Not every jail stint is worse than a fine; not every fine is worse than community service. We need some way to determine whether a given fine is a harsher punishment than a particular amount of jail time or community service. It’s not at all clear how this is to be done. And, even if there is an answer here, there remains a further question: which punishment(s) should occupy the upper and lower bounds of the punishment scale? It is here, ultimately, that the most serious worry about Kleinig’s view presents itself. For even if the relevant scales are in principle determinable, Kleinig offers no advice about how to identify these boundaries. His suggestion is that we fix deserts by reference to two anchors, which correlate the mildest and gravest offense with the lightest and harshest morally justified punishment. For retributivists, however, the moral justification of a punishment depends crucially on whether it is deserved. According to Kleinig, we cannot know whether (e.g.) mass murder deserves the death penalty until we have determined that the death penalty is the harshest morally justified punishment. But retributivists make such a determination based on whether the death penalty is deserved for such offenses. There is thus a vicious circularity at the heart of Kleinig’s efforts to avoid arbitrariness in the correlation of punishment and offense scales. All punishments but the mildest and harshest are morally deserved only if these anchor punishments are morally deserved. But Kleinig gives no reason for thinking that the anchors we conventionally employ are themselves morally deserved, and offers us no basis for determining whether any other anchor point is deserved. Thus we can have no justified confidence that a schedule of punishments run along Kleinig’s lines would mete out just deserts. D.
EQUAL TREATMENT
Another strategy for commensurating crime and desert relies on the claim that an offender’s moral desert is fixed exclusively by considerations of fairness and equal treatment. On this line, criminals deserve nothing more (or less) than to be treated fairly and equally. So long as the liability and sentencing rules are publicly promulgated, are such as to be obeyable by the citizenry, and are applied regardless of social class, then one punished under such a regime is getting what one morally deserves. The problem, © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
205
of course, is that public, followable, equally applied rules may nevertheless be extremely unjust. It may be that everyone without exception is required to refrain from criticizing the head of state, and it may be that everyone can do this, knows that it is legally required to do this, knows that the punishment for failing to do this is death, and knows that this punishment is carried out on rich and poor, white and black, male and female alike. But the punishment isn’t any sort of just desert. In fact, fairness and equal treatment in the specification of what wrongdoers deserve is entirely epiphenomenal. Any plausible desert principle will specify like sanctions for like offenses. Straightforward application of any such principle will therefore ensure equal treatment for like offenders. But equal treatment is a by-product of correct application of a desert principle. It is not the principle itself. Considerations of fairness and equality alone are insufficient to fix what it is that a person morally deserves. E. UNFAIR ADVANTAGE
A currently popular approach to commensurating wrongdoing and desert begins by trying to accommodate the insights of the previous two principles without making either proportionality or equal treatment the sole criterion of deserved punishment. This approach claims that the point of punishment is to correct unfair advantages obtained by a criminal through his wrongdoing.29 In order to implement this strategy, we need first to identify the relevant sort of unfair advantage. Crucially, the benefit is not measured by a victim’s suffering or loss. Unfair advantage theorists clearly see that such a measure leads directly to lex talionis, and they are intent on developing a retributivism free of any association with lex. Rather, the unfair advantage is understood as that amount of extra benefit one enjoys vis-à-vis all law-abiding citizens who, in obeying the law broken by the criminal, have refrained from taking to themselves the advantage that the criminal has secured. There are two central difficulties with unfair advantage theories. The first is that it is a mistake to think that the fundamental wrong associated with most crimes against persons is a kind of free-riding offense. Rape isn’t wrong because of the advantage the rapist enjoys with regard to all other law-abiding citizens, but because of the harm imposed on the victim. And even if there is this peculiar kind of free-riding advantage involved in rape, as there allegedly is in all crime, this isn’t what we are trying to isolate and identify as the source of the grievous wrong that rapists do. Hence it cannot plausibly be put forth as the basis for identifying morally satisfying punishments for rapists. Unfair advantage theories understand what lex does not – that some properly criminalizable conduct does not involve significant harm to determinate victims. But in accommodating © 2000 University of Southern California and Blackwell Publishers Ltd.
206
PACIFIC PHILOSOPHICAL QUARTERLY
this fact, they lose sight of what makes lex so intuitively satisfying, at least for paradigmatic criminal cases – such cases do involve serious personal harm, and such harm is the basis for prohibiting the conduct in the first place. This is a preliminary difficulty. Suppose we set it aside and just concentrate on the positive proposals that unfair advantage theorists have offered as a means of commensurating wrongdoing and punishment. According to all such theories, the point of punishment is to nullify the wrongfully gotten gains of the criminal. These gains are said to consist in some added degree of liberty. But how do we measure liberty? Two prominent expositors offer a clue: graver immoralities necessarily signal greater appropriations of freedom. But even if this is so (and its defense amounts to a stipulation),30 we still have nothing more than a commitment, per definitione, to the proportionality principle as a basis for effecting the desired commensuration. As we’ve seen, that will not do the work that retributivists require of it. If the unfair advantage theory is going to work, then there must be some determinate measure of unfairly gained liberty, and we must be able to use such a measure as a basis for nullifying the wrongly gained extra liberty. But there is no way even to undertake the first step of this process. Just how much more liberty has a batterer enjoyed than a tax cheat or a blackmailer? There really is no plausible answer to this question. I propose we take this as an indication that this is the wrong question to be asking in the first place. Michael Davis is the only one among the unfair advantage theorists who has made a serious effort to supply a commensurating measure for punishment.31 Davis measures the extra liberty a criminal has obtained by asking how much an ordinary person would pay, in a relatively free market, to be exempted from punishment for its commission. The greater the sum, the greater the extra liberty. The greater the extra liberty, the more serious the offense. The more serious offense, the harsher the deserved punishment. The tidiness of the proposal masks two significant difficulties. The first is that increased payments for exculpation licenses will not always accurately track increased immorality. It may be, for instance, that being excused for a multimillion dollar robbery would net a higher bid than an excuse for a murder. This entails, on Davis’ view, that the robber is deserving of a greater punishment than the murderer. And here we face a dilemma: either this means that robbery is a fortiori morally worse than murder, or it means abandoning the proportionality requirement so attractive to most penal theorists. The first horn is true only if an action’s relative immorality is necessarily well-measured by reference to the exculpatory preferences of ordinary people. This doesn’t strike me as a promising route to take. But if we grasp the second horn, and so allow that lesser © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
207
immoralities may deserve harsher punishments, we abandon a fundamental plank of the retributivist platform. Suppose that this problem can be solved. Still, even if graver crimes always attracted pricier exculpation bids, the bids themselves do not offer a plausible basis for specifying just what it is that a person deserves for an instance of criminal wrongdoing. Suppose the ordinary reasonable person in a relatively free market would pay X amount of dollars for committing crime C. How, precisely, is this fact going to translate into a specification of a just desert for having committed C? The straightforward answer is that paying X just is the deserved punishment for having done C. But this will mean quite mild punishments in a society comprised mostly of miserly bidders, and far heavier punishments for those unfortunate enough to live among enthusiastic bidders. It will mean that the rich can discharge their punitive obligations without any suffering. It will mean that a wrongdoer’s just deserts are always exhausted by a financial payoff. (Presumably such payment goes to government coffers, since the fundamental wrong is not one done to any victim, but instead to all lawabiding citizens.) If this all sounds far-fetched, we need to find a different way to base the specification of a deserved punitive sanction on the average cash bid for an exculpation license. I don’t know what this alternative strategy would be. F. INSTRUMENTALISM
We might instead understand desert instrumentally. Instrumentalism about desert says that a punishment is morally deserved if and only if it is instrumentally efficacious in satisfying an important goal.32 Few important goals will be satisfied by punishment – punishing people won’t make the citizenry any more literate, or reduce infant mortality, or improve race relations. But it might deter crime, or assist in social rehabilitation or moral reformation. Rather than engage here in the intramural debate of selecting the relevant goal, let us suppose, just for illustrative purposes, that deterrence is the chosen goal. Then, on the instrumentalist view, if deterrence of (e.g.) arson required an average of seven years imprisonment, then seven years is what an arsonist would morally deserve. Instrumentalism about desert is not going to help the retributivist. Suppose that instrumentalism is true. Then we need to identify the goal whose satisfaction will fix desert claims. This is just the question of which among many competitors is the best general justifying aim of punishment. If we are retributivists about general justifying aim, then the goal of punishment is to do justice. But then, according to this version of instrumentalism about desert, deserved treatment is that treatment that will serve justice. This is pretty innocuous stuff, and certainly doesn’t by itself represent, or supply us with a basis for identifying, a concrete, © 2000 University of Southern California and Blackwell Publishers Ltd.
208
PACIFIC PHILOSOPHICAL QUARTERLY
informative measure of morally deserved punishment. We will get more assistance only if we abandon retributivism as a general justifying aim. If morally deserved punishment is that treatment that will deter, or rehabilitate, or reform, then we have at least in principle a mechanism for assigning specific desert claims. So we may either retain retributivism as the general justifying aim of punishment, but abandon hopes for justifying specific desert claims, or secure the possibility of such justification, but at the cost of abandoning retributivism about the general justifying aim. We can be merciful. We don’t have to force this choice on the retributivist, because the dilemma is predicated on the truth of instrumentalism about desert. This view seems to me to be false, though I can’t attempt to show it so here. To put it tendentiously, the instrumentalist move is a very close cousin of two quite suspicious philosophical maneuvers: Mill’s effort to show that justice is only a species of utility, and a hard determinist’s claim that criminals deserve to be punished because it would be efficacious to do so. To give what is deserved is to give what is meet, appropriate, fitting. If the only fitting response to persons is a response that maximizes nonmoral good, then instrumentalism is true. But it certainly seems that both conceptually and practically, what is deserved and what is useful can come apart. Exemplary punishment or vicarious liability may well be most useful at effecting some valuable social goal, yet they don’t seem deserved for all that.33 Desert awards or sanctions are essentially retrospective, while determinations of efficacy are prospective. And, crucially, desert is predicated on a kind of personal responsibility that may be irrelevant to the usefulness of a particular (kind of ) sanction. It is admittedly simpler to reduce desert to considerations of utility, but as in many instances, simplicity here means oversimplification.34 Certainly there is more that could be said about instrumentalism’s prospects, but rather than say it here, let us instead pose a dilemma. If instrumentalism about desert is false, then a fortiori it is useless as a means of fixing just deserts. But even if it is true, it fixes just deserts only if we abandon justice as punishment’s goal, and so retributivism as a general justifying aim of punishment. Either way, it seems that retributivists should look elsewhere for assistance in combatting nihilism about moral desert. Here, then, in summary fashion, are the various candidates for commensurating punishment and moral desert: A punishment P metes out the suffering that a criminal S morally deserves for crime C iff (i) (ii)
Lex1: P imposes on S just the same kind of treatment that S imposed on his or her victim. Lex2: P imposes on S just the same amount of suffering that S imposed on his or her victim.
© 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
209
(iii)
Lex3 (Nozick): P = S’s degree of responsibility (r) multiplied by S’s wrong (H). (iv) Lex4 (Hampton): P is necessary and sufficient for nullifying the message of inferiority sent by C. (v) Lex5 (Waldron): the suffering or hard treatment occasioned by P is similar to that occasioned by C. (vi) Democracy: P is authorized by a democratically enacted sentencing rule. (vii) Proportionality: P is a member of a set of proportional punishments. (viii) Equal Treatment: all who commit C (without excuse or justification) receive P. (ix) Unfair Advantage: P deprives S of that amount of liberty that S has unfairly gained in doing C. (x) Instrumentalism: P is efficacious in achieving some valuable goal.
5. This concludes the argument on behalf of nihilism about moral desert. As an argument from elimination, it is sound only if all of the available alternatives have been canvassed and properly disposed of. So this reveals two avenues of possible criticism – an important possibility for commensurating punishment and moral desert may have been omitted, or I may have dismissed too quickly an option that is in fact viable. Either route (or the both of them) may prove to be correct, but I will not attempt here to try to anticipate objections and make replies. Instead, let us direct our attention to two apparent vulnerabilities that emerge from an endorsement of nihilism about moral desert. The first stems from the intuitive appeal of the view that wrongdoers deserve to suffer. This view, shared by all retributivists, does need defense, but I am prepared to grant its plausibility in the ensuing discussion. Having done that, however, it seems just a short step to a refutation of nihilism about moral desert. Consider: 1. Wrongdoers morally deserve to suffer. 2. If wrongdoers morally deserve to suffer, then they morally deserve to suffer in some particular way. 3. Therefore wrongdoers morally deserve to suffer in some particular way. 4. If wrongdoers morally deserve to suffer in some particular way, then nihilism about moral desert is false. 5. Therefore nihilism about moral desert is false. © 2000 University of Southern California and Blackwell Publishers Ltd.
210
PACIFIC PHILOSOPHICAL QUARTERLY
Since the argument is logically valid and the conclusion one I believe to be false, I must reject one of the premises. Premise (4) is an analytic truth. This forces a choice between premise (1) and (2). The former seems the natural target, and the nihilist may in the end have to reject it. Still, my aim here is to show how we might accommodate the view expressed in (1) while eschewing any retributivist associations. Thus my task becomes that of showing how it might be true that wrongdoers morally deserve to suffer, even though they don’t morally deserve to suffer in any particular way. I want to undertake this task in the context of a further, familiar problem that besets nihilism about moral desert. This sort of nihilism seems to conflict quite directly with all of the plausible, deeply-held convictions we have about those particular cases described earlier. If there is no determinate sanction that a blasphemer morally deserves, then handing him a twenty year sentence isn’t punishing him more than he deserves. If there is no penalty that a murderer morally deserves, then what could possibly be wrong with freeing him after just twenty days? Surely there is something wrong with such treatment. The readiest explanation of this is that these offenders are being punished either more or less than they morally deserve to be. Nihilism about moral desert is incompatible with this explanation. That may mean that those who endorse such nihilism are forced to abandon these plausible convictions about cases. What follows? Not all, and perhaps none, of our deeply-felt moral convictions are self-certifying. Our convictions about these examples may represent large-scale error on our part. Perhaps we should stick with this brand of nihilism and reconsider our intuitions about the cases. But a better strategy would be to save the appearances, if we could. I believe we can. Consider this argument: 1. Punishment is morally justified only if it is morally deserved. 2. One is morally deserving of some sanction only if it is morally appropriate that one be made to suffer because of some past misdeed for which one is responsible. 3. Therefore, punishment is morally justified only as a response to a past morally wrong action for which one is responsible. Neither premise is beyond question. Hard determinists may reject the second premise, and some nihilists about moral desert will reject the first. But I am going to suppose without argument that both premises are true, and so I will grant the truth of the conclusion. Moral desert is a necessary condition for justified punishment. This explains why punishment for “violation” of vicarious and strict liability criminal statutes is unjustified. Particular punishments can be morally © 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
211
undeserved because (and only because) those being punished don’t deserve to be punished at all: either they haven’t committed any legal wrong, or have, but possess some valid excuse or justification. So moral desert can explain why some punishments are unjustified. But even if it is morally appropriate to make wrongdoers suffer just because such people deserve to suffer, moral desert fails to supply an index that can be used to fix the kind or amount of morally appropriate punishments. It does not explain why forty years in jail for littering is too much punishment, and it does not explain why twenty days in jail for murder is too little. So what of our intuitions? I believe that we can retain them, so long as we see them focused not on the issue of a criminal’s getting more or less than what he deserves, but instead on his getting too much or too little punishment. This is the same thing only if one endorses the commensurability thesis – only if one thinks that a punishment’s justification hinges entirely on whether it gives an offender what he or she morally deserves. If one rejects the commensurability thesis, the burden becomes that of explaining how a forty year sentence for littering is too much punishment, and a twenty day jail stint for murder is too little, without invoking the notion of moral desert. I am sure I cannot discharge this burden here. Let me conclude by offering something in the way of an explanation for my inability. Suppose that nihilism about moral desert is true. If the commensurability thesis is also true, then no legal punishments are morally justified, because no punishments are such that they represent just what (no more or less than what) a criminal morally deserves. It follows that the punishment imposed on the litterer and blasphemer are morally unjustified. Good. But we get this attractive result only by abandoning the moral justification of every punishment. That’s not so good. If legal punishment is sometimes morally justified, and if nihilism about moral desert is true, then the commensurability thesis must be false. Taken together, this entails that punishment’s general justifying aim must be something other than meting out moral deserts. My simple thought is that if punishment can be justified, then whatever it is that serves as its ultimate justifying aim can itself serve as, or be the basis for determining, an index for fixing what is to count as too much or too little punishment. Some punishments will count as morally unjustified, precisely because they impose too much or too little hard treatment. The key point is that the proper index of such treatment will be something other than moral desert. The identity of the correct measure depends on the nature of punishment’s general justifying aim. I don’t intend here to choose among eligible candidates. Let me say just this – either some nonretributivist general justifying aim can be morally justified or it can’t. If it can’t, and if my arguments thus far are sound, then legal punishment cannot be morally justified. If such punishment © 2000 University of Southern California and Blackwell Publishers Ltd.
212
PACIFIC PHILOSOPHICAL QUARTERLY
can be justified, then we may be able to retain our intuitions about particular cases. These intuitions will be justified not by considerations of moral desert, but by some other goal. The justification of punishment depends crucially on our ability to identify it.35 Department of Philosophy University of Kansas NOTES 1 See, e.g., Jean Hampton, “A New Theory of Retribution,” in C. Morris and R. Frey, eds., Liability and Responsibility (Cambridge: Cambridge University Press, 1991), pp. 377– 414; Immanuel Kant, Metaphysical First Principles of the Doctrine of Right, sections 331–7; John Kleinig, Punishment and Desert (The Hague: Martinus Nijhoff, 1973); G. W. F Hegel, The Philosophy of Right, section 100ff; H. J. McCloskey, “A Non-Utilitarian Approach to Punishment,” Inquiry 8 (1965), pp. 239–55; Michael Moore, “The Moral Worth of Retribution,” in F. Schoeman, ed., Responsibility, Character and the Emotions (Cambridge: Cambridge University Press, 1987), pp. 179–219; Robert Nozick, Philosophical Explanations (Cambridge, MA: Harvard University Press, 1981), pp. 363–97; Igor Primoratz, Justifying Legal Punishment (London: Humanities Press, 1989); A. von Hirsch, Doing Justice (New York: Hill and Wang, 1976). 2 See Herbert Morris, “A Paternalistic Theory of Punishment,” American Philosophical Quarterly 18 (1981), pp. 263–71; Jean Hampton, “The Moral Education Theory of Punishment,” Philosophy and Public Affairs 13 (1985), pp. 208–38; and R. A. Duff, Trials and Punishments (Cambridge: Cambridge University Press, 1986). These authors believe that making a criminal suffer is necessary to effect the desired moral reformation. So both moral reform and retributivist theories will say that justified punishment requires the imposition of suffering. However, only the retributivist will insist that suffering is what is morally deserved. 3 The distinctive retributivist thesis about judicial obligation is that judges always ought to impart sentences that comply with the commensurability thesis. This view obviously fails if the commensurability thesis turns out to be false. But the view about judicial obligation may be false even if the commensurability thesis is true. For instance, it might be that a judge’s paramount obligation is to enforce the sentencing guidelines on the books, even when such guidelines impose undeserved punishments. In jurisdictions whose sentencing guidelines fail to assign morally deserved punishments, judges are faced with a choice between fidelity to law and fidelity to the commensurability thesis. The commensurability thesis does not dictate a choice, since it concerns itself with the content of sentencing guidelines rather than a judge’s moral obligation vis-à-vis those guidelines. Thus retributivists about judicial obligation must offer arguments in addition to those in defense of the commensurability thesis if they are to secure their conclusion. 4 Thus my argument here is not that the commensurability thesis is false because there is no principled way to determine whether an offender deserves (say) three hundred and sixtyfour days rather than three hundred and sixty-five days in jail. (cf. G. W. F. Hegel, Philosophy of Right, section 215, for the opposite worry.) My argument, in other words, does not rely on an appeal to the vagueness of the notion of desert, where vagueness is understood to characterize properties that are embeddable in sorites paradoxes. Were considerations of moral desert capable of specifying a determinate range of sanctions (e.g., one to three years
© 2000 University of Southern California and Blackwell Publishers Ltd.
RETRIBUTIVISM AND DESERT
213
in jail), that would be sufficient to fend off my concerns. My view, however, is that moral desert fails to do even this. 5 Of course, if one is inclined to say that virtue is its own reward, and do-gooders don’t deserve any further benefits, then the burden is to explain why the correlative point doesn’t hold for wrong-doers; why, that is, the vice that they have exemplified is not punishment enough for the wrongs they have done. 6 The flaws of lex have been widely chronicled. Perhaps the best treatment is given by Michael Davis, “Harm and Retribution,” Philosophy and Public Affairs 15 (1986), pp. 236– 66. For fine discussion on how the biblical injunctions were originally interpreted and put into practice, see David Daube, Studies in Biblical Law (Cambridge: Cambridge University Press, 1947), pp. 102–53. 7 See J. Reiman, “Justice, Civilization and the Death Penalty: Answering van den Haag,” Philosophy and Public Affairs 14 (1985), pp. 115–48. At p. 125 he writes: “The equality and rationality of persons implies that an offender deserves and his victim has the right to impose suffering on the offender equal to that which he imposed on the victim.” 8 Nozick, Philosophical Explanations (Cambridge, MA: Harvard University Press, 1981), pp. 363–97, esp. pp. 363–5, 388–90. 9 That the relevant responsibility is not causal is clearly implied in the third paragraph on ibid., p. 363. 10 Ibid., p. 388. 11 Ibid., p. 365. 12 Ibid., p. 364. 13 See Hampton, “A New Theory of Retribution,” in C. Morris and R. Frey, eds., Liability and Responsibility (Cambridge: Cambridge University Press, 1991), pp. 377–414, esp. at 400–1. 14 A notable exception to the retributivist consensus on this point is J. D. Mabbott, in “Punishment,” Mind 48 (1939), pp. 150–67. 15 Op. cit., p. 398. 16 Ibid., pp. 401–2. 17 For complementary criticisms of Hampton’s project, see David Dolinko, “Some Thoughts about Retributivism,” Ethics 101 (1991), pp. 535–57, at 549–54. He is there criticizing a version of Hampton’s views as given in her contribution to Forgivenness and Mercy, by Hampton and Jeffrie Murphy (Cambridge: Cambridge University Press, 1988), ch. 4. The article I am relying on for Hampton’s views is based on chapters 2 and 4 of her co-written work with Murphy. 18 Waldon, “Lex Talionis,” Arizona Law Review 34 (1992), pp. 25–51. 19 Ibid., p. 32. 20 See Rawls, “Two Concepts of Rules,” Philosophical Review 44 (1955), pp. 3–13; H. L. A. Hart, “Prolegomenon to the Principles of Punishment,” in Punishment and Responsibility (Oxford: Oxford University Press, 1968), pp. 1–27. 21 Ibid., pp. 27–31. 22 Ibid., p. 38. On the all-things-considered status of lex, see Waldron, pp. 29 n. 7, 31–2, 48. 23 Discussion of these examples occurs at pp. 36–41. 24 For sustained defense of the claim that it is sometimes morally unjustified to treat others as they deserve, see Joel Feinberg, “Justice and Personal Desert,” in Doing and Deserving (Princeton: Princeton University Press, 1970), and Douglas Husak, “Why Punish the Deserving?” Nous 26 (1993), pp. 447–64. 25 We could, of course, present the democratic proposal as a unified theory of moral desert by rejecting the idea of intrinsic wrongness and endorsing the view that all moral © 2000 University of Southern California and Blackwell Publishers Ltd.
214
PACIFIC PHILOSOPHICAL QUARTERLY
wrongs are mala prohibita. Such a move is perilous, and gains whatever superficial plausibility it may have through an ambiguity attaching to the notion of intrinsic wrongness. On the one hand, it may refer simply to pre-legislative wrongness. On the other, stronger understanding, something is intrinsically wrong just if it is wrong in its very nature, and so in every possible context, and is wrong prior to and independently of being thought so by any actual or hypothetical agent. The contrast between actions mala in se and mala prohibita requires only the weaker understanding of intrinsic wrongness; to reject the existence of this sort of wrongness is to commit oneself to the view that acts are morally wrong if, only if, and because they are deemed to be such by the legislature. I take it that no one holds such a view. 26 See Kleinig, Punishment and Desert (The Hague: Martinus Nijhoff, 1973), pp. 124ff. 27 Ibid., pp. 128–9. 28 Feinberg, “Equal Punishment for Failed Atempts: Some Bad but Instructive Arguments Against It,” Arizona Law Review 37 (1995), pp. 117–33, at 132. 29 This approach has been championed by Herbert Morris, “Persons and Punishment,” Monist 52 (1968), pp. 475–501, whose views have importantly influenced all of the following unfair advantage theorists: J. Finnis, “The Restoration of Retribution,” Analysis 32 (1972), pp. 131–5; J. G. Murphy, Retribution, Justice and Therapy (Dordrecht: Reidel, 1979), pp. 77–115, 223–49; W. Sadurski, Giving Desert its Due (Dordrecht: Reidel, 1985), pp. 221–60; G. Sher, Desert (Princeton: Princeton University Press, 1987), pp. 69–90; M. Davis, “How to Make the Punishment Fit the Crime,” Ethics 93 (1983), pp. 726–52, and “Criminal Desert and Unfair Advantage: What’s the Connection?” Law and Philosophy 12 (1993), pp. 133–56. 30 See Sadurski, op. cit., p. 229, and Sher, op. cit., p. 82. 31 See Davis, op. cit., n. 29. 32 Ernest van den Haag endorses such a view in “Refuting Reiman and Nathanson,” Philosophy and Public Affairs 14 (1985), pp. 165–76. Throughout his writings van den Haag freely pulls from both deterrence and retributivist traditions to defend his specific views about penal institutions and practices. At p. 167, van den Haag claims that “punishment must, whenever possible, impose pain believed to exceed the pain suffered by the individual victim of crime. No less is deserved.” He says this because he believes that what is deserved is determined by what will deter and what will vindicate the disrupted social order, and he believes that a proportionately harsher punishment than that called for by lex is what is necessary to secure these goals. 33 See J. Kleinig, Punishment and Desert (The Hague: Martinus Nijhoff, 1973), pp. 56ff., for a more detailed discussion of this point. 34 Nothing I’ve said here is meant to be critical of utilitarianism, since utilitarians may reject instrumentalism about desert. If they do, they must allow that (at least in principle) considerations of utility and desert can conflict. The utilitarian task then becomes that of showing how considerations of utility morally trump those of desert in cases of conflict. 35 Many thanks to Sandy Kadish for his comments on an earlier draft of this paper. My work on this topic has been aided by receipt of a General Research Fund grant from the University of Kansas.
© 2000 University of Southern California and Blackwell Publishers Ltd.
