As in the discussion of offenses, I begin by observing some of the gaps left open by a welfare-oriented, utilitarian account of punishment, and then suggest how a dignity-based Kantian account can fill those gaps. Philosophical discussions of punishment commonly address the question of justification: what makes punishment legitimate? Dealing with this question assumes that punishment requires some form of justification. Why? Bentham’s answer point to the similarity between punishment and crime. What makes punishment a presumptive evil is that like crime it diminishes utility by inflicting pain. This also points to the form a justification of punishment must supposedly take: punishment is justified inasmuch as it generates greater aggregate welfare by preventing crime, mostly through deterrence. A well-known difficulty with this justification is that it does not convincingly limit punishment to those guilty of crime: we can imagine extraordinary situations in which punishment would serve this purpose when imposed on someone who is innocent of any wrongdoing, as long as people believe in the defendant’s guilt. Yet “punishing” the innocent even in those situations would be a moral outrage. A utilitarian may bite this bullet by denying the outrage, but others will find the bullet too hard to bite.
what they are: as departures. Neither the imposter nor the actual but murderous surgeon exemplifies in these situations medical practice. Seen in light of the nefarious scheme, the usual terms used to describe medical practice assume an ominous, borrowed quality; they appear, as it were, in scare quotes.
Even if a utilitarian account were successful in limiting punishment to the guilty in principle, other problems arise. Every system of punishment is fallible, and so involves instances of punishing the innocent. We take such lapses to be particularly alarming, and so subject the imposition of punishment to an unusually stringent system of substantive, procedural, and evidentiary restrictions. But what’s so alarming about occasionally punishing the innocent? After all, government routinely disadvantages people for the sake of the greater good through taxation, civil damages, military conscription, fiscal policy, and the like. And yet none of these practices is as morally fraught as the practice of punishment, nor hedged by as strict a system of restrictions.
It is natural to focus in this connection on the severity of the deprivations punishment involves and view this as its distinguishing mark. But although punishment is often quite harsh, at least when understood along straightforward utilitarian lines, such a simple account of the distinctness of punishment does not hold up. First, the perception of the harshness of punishment does not depend on its aggregate but rather on its distributive effects. For example, the fact that the number of people who actually die by execution is, statistically speaking, quite small does not diminish our concern with capital punishment, since we deem it the most severe deprivation imposed on a particular individual. This stands in sharp contrast to the way other forms of governmental deprivation are evaluated— for example, the increase in road fatalities when, say, the speed limit is raised. Second, even in the individual case, the sheer level of hardship suffered through punishment does not quite make it stand out. For example, the hardships and dangers a conscript faces during training, let alone at combat, may far exceed what inmates face in jail. And yet we don’t think of military service as raising the same moral issues as punishment or of the barracks as equivalent to jail. Finally, while going to extraordinary lengths to ensure that the defendant is indeed guilty of the crime charged, the law accepts with near equanimity the severe impact punishment often has on innocent parties such as the defendant’s family.
Our practices and attitudes regarding punishment suggest that we view it as having moral significance that goes beyond the setback to welfare it involves. The similarity between punishment and crime that Bentham highlights is significant and morally consequential, but the preceding comments suggest that it has a more complicated shape. Like crime, punishment raises moral concerns, the intensity and shape of which are not adequately explained by the setbacks to welfare it involves. We need to adapt Bentham’s insight to our earlier depiction of criminal offenses as strikes against human dignity. As we have seen in the previous section, when a criminal offense involves the infliction of harm, its moral badness is not exhausted by the badness of the harm. Rather, the evil of the crime is a matter of the significance with regard to the victim’s moral worth of inflicting harm, when certain additional conditions obtain. As an example, consider first-degree murder. Its heinousness is based in the first place on the judgment that it involves the infliction of a most grievous deprivation. This judgment looks exclusively at the effects on the victim. In placing murder above, say, robbery, and at the top of the severity list, we do not investigate which crime is more socially harmful in the aggregate; we compare only the two felonies’ distributive effects. Furthermore, it is significant that the murderer targets an individual victim, whose death must be the offender’s conscious objective rather than just a foreseeable side effect of her action; and the action must be the product of “premeditation and deliberation,” which marks it as more reprehensible than, say, reckless killing. This combination of factors—call it mistreatment—amounts to an especially egregious expression of disrespect: by deliberately inflicting a severe deprivation on a particular individual one enacts a conception of the victim as a mere means, someone whose own rights and interests can be trampled at will.10
It is easy to see that unlike other deprivations wrought by government action, criminal punishment is similar in these respects to crime, and so, unless justified, would present an unambiguous case of mistreatment. The deleterious effects of a fiscal policy on employment, for example, can be accidental and unforeseen, whereas the deprivation involved in punishment is always deliberate. Unlike taxation, which usually affects anonymous individuals in an impersonal way, punishment focuses on a particular defendant. Unlike quarantine, which may in principle be conducted in a five-star hotel, some form of deprivation is the acknowledged purpose of punishment, not just a regrettable side effect. Finally, criminal punishment involves harsh deprivations, with the harshness measured in terms of effects on the particular defendant rather than in aggregate societal terms. Being a form of mistreatment, criminal sanctions pose a greater threat to human dignity than other deprivations, and the person whose dignity is at stake is the defendant on whom the sanction is visited, rather than any others, like the family, who may be indirectly affected. The battery of restrictions created by criminal law is designed to mitigate this threat to the defendant’s dignity by ensuring, as much as possible, that the defendant is justly treated rather than being the victim of the equivalent of a crime.11
This conclusion of course assumes that punishing the guilty, unlike inflicting sanctions on the innocent, is indeed just. This assumption can be interpreted in two ways, corresponding, respectively, to weak and strong retributivism: that punishing the guilty is licensed by justice or that it is required by it. On the former view, unlike inflicting sanctions on the innocent, punishing the guilty is respectful of their autonomy, and so is morally acceptable. But on this view a complete justification of punishment must include some affirmative reason, a reason that is commonly understood in crime-prevention, consequentialist terms.12 Since we can imagine cases in which punishing the guilty would not advance such con- sequentialist concerns, the result is that at least in principle punishing some offenders would not be justified. Kant famously holds the stronger view. He speaks of the duty to punish a convicted murderer even as the final act taken by an island community that is about to disband. Though this scenario blocks an appeal to any future-oriented justifying aims of punishment, Kant maintains that the offender must be punished,
so that everyone will duly receive what his actions are worth and so that the bloodguilt thereof will not be fixed on the people because they failed to insist on carrying out the punishment; for if they fail to do so, they may be regarded as accomplices in this public violation of legal justice.13
What do these dark words mean?
The example and the retributive view it conveys have spawned a broad range of interpretations, in part because what Kant says is unclear. So, for example, in a rightly celebrated paper, Herbert Morris argues along Kantian lines for a seemingly paradoxical right to be punished, seen as an aspect of an inalienable right to be treated as a person.14 Morris draws on the connections between punishment and responsibility, between responsibility and autonomy, and between autonomy and respect. The failure to punish is a failure to hold the perpetrator responsible, which impugns her autonomy, and so treats her with disrespect. I will not comment on this and related lines of thought other than to note that they propose an offender-oriented theory of retribution. My suggestion, though, is that close attention to the quoted paragraph reveals an additional strand in Kant’s view that amounts to a victim-oriented retributive approach, maintaining that in punishing the offender, or withholding punishment, the victim’s dignity is at stake.
The first reason Kant gives for his retributive view does indeed focus on the offender (“so that everyone will duly receive what his actions are worth”). But then he makes two puzzling claims. The first is that failure to punish will result in the “bloodguilt” of the killing attaching to the community itself. But how does the killer’s bloodguilt spread to the community, which after all had not shed the blood, just by virtue of its failure to punish? The second is the vague talk of the people becoming accomplices. Accomplices in what? The “public violation of legal justice” of which they supposedly stand accused just consists in the failure to punish, which is entirely the people’s own failure; whose accomplices are they in this regard? We can make some headway if we see these two claims as connected. By failing to punish, the people share the killer’s guilt because in this way they are implicated in the crime, and so become accomplices to it. But how can an ex post failure to punish implicate the community in a crime that had been already committed? And in what sense does punishing the offender take the community off the moral hook?
The answer I propose connects to my argument that criminal law’s primary concern is not the victim’s life, but the fact that the life was taken by another person, and that this, in turn, exhibits a concern for the victim’s dignity. For if the life of the victim were all that mattered, it would make no sense to saddle the community with responsibility for the death when it fails to punish the perpetrator: ex post punishment would not have prevented the victim’s death or restored him back to life. But if the victim’s dignity is the central concern, the situation changes dramatically. Although the victim cannot be revived, his dignity can still be vindicated. So if the moral significance of homicide lies in trampling the victim’s moral worth, there is still something posthumous the community can do about it: through punishment, that moral worth is reasserted. Hence, failure on the part of the community to punish the perpetrator is a failure to rectify, at least partially, the derogation of the victim’s dignity, and so amounts to participating in the evil of the initial act.
But why is punishment the appropriate medium for vindicating the victim’s dignity? Can’t the community rectify the offense to dignity in some other, gentler way? A radically different and gentler world is imaginable, and we should not write it off. But the burden of the argument is not to establish that punishment is the only appropriate reaction to crime in all possible worlds. A sensible objective is to examine our actual world, in which punishment exists and has roughly the shape and significance it does, and try to assess it in light of some broader moral considerations. Let me return in this vein to the initial claim that the mission of criminal law is to uphold dignity. In further elaborating this mission, we can distinguish between two ways of upholding dignity: by defending it and by vindicating it. Now, broadly speaking, a system of punishment is required in order to defend dignity by helping prevent dignity-offending criminal acts. Punishment does not always advance this goal, however; in particular, it does not advance it in Kant’s imagined community that is about to disband. So the following question arises: why is the same practice that ordinarily serves to defend dignity appropriate for its vindication as well? The answer echoes the point I made in discussing criminal offenses. It concerns the meaning that accrues to a practice and characterizes instances falling under it even when the underlying reasons for that meaning do not apply in the specific case. Once punishment becomes the dominant medium for responding to offenses against human dignity for preventive reasons, it acquires a meaning that persists in cases in which prevention is not served. Failure to punish under these circumstances (unless some special considerations apply) has the offensive meaning of a failure to uphold, by vindicating, the victim’s moral worth.
-  This victim-oriented ground for punishment is cumulative with the offender-oriented considerations that either license punishment (on the weak retributive view) or mandate it (onthe strong view).