A Dialogue on the Death Penalty

Gerald Dworkin and Justin E. H. Smith

FrenchGuillotine Jerry and I began this dialogue after he, in the process of preparing an ethics course on the topic of capital punishment, happened upon some pieces I wrote a few years ago at various activist venues (they are archived here, here, here, here, and here). The articles are polemical rather than scholarly, and I never expected the issue of capital punishment would someday get any attention from me qua philosopher (as opposed to qua polemicist). But Jerry found some of the issues I raised in them worthy of attention, and in turn has raised for me a number of issues that I never really worked through before in my very visceral opposition to the death penalty. I'm grateful for this, and I think what has resulted is a discussion that should be of interest to reformers and philosophers alike (as well as to those who belong in both of these camps.) –JEHS.


Gerald Dworkin: Justin and I agree that capital punishment as currently administered in the United States, and in the absence of convincing evidence that it deters more than a sentence of life imprisonment without possibility of parole (LIWP) for any crime, should be abolished. Where we may disagree –I put it this way because I am not sure what view I will emerge with at the end of this discussion– is whether there is an argument for abolition that does not depend on contingent facts, such as that it does not deter, or that as currently administered the selection of who gets executed is both arbitrary (chance and luck play an enormous role) and unjust (the poor and racial minorities are executed at higher rates that those with money and those blacks who kill whites). For, as Justin points out elsewhere, believing that the facts are as they are is compatible with believing that in a world where deterrence is established and fairness reigns CP is justifiable. This position could be true even if one believed that as a matter of contingent fact our system will never be sufficiently just, and the evidence for deterrence will never be sufficiently strong so as to warrant CP. The first thing I want to do is see if Justin and I agree on a more rigorous definition of the issue. For it is, as I shall argue later, very important exactly how the problem is framed.

I propose that what the defender of what I shall call the non-contingent wrongness of CP must show is the following:

• There is no successful argument for the legitimacy of CP which survives empirical and normative scrutiny.

• In addition to this via negativa there is a valid positive argument whose conclusion is that CP is illegitimate which does not rely in an essential way on contingent matters of fact.

This last claim is difficult to formulate accurately. Does an argument which involves the premise “death is different” (whatever that may mean) rely on the fact that once we die we don’t come back to life? That, presumably, is a contingent fact. Does an argument which says that CP is “cruel” rely on some characteristic of CP that makes it cruel? But maybe that characteristic is not an empirical one. If one defines cruelty as acting in a way which is intended to cause great suffering to another person for its own sake then, although it is a contingent fact, whether or not CP is designed or justified by reference to this feature, if it is then it is not a contingent fact that it is cruel.

Justin Smith: Jerry's initial characterization of my position is right on. I do not yet know whether there is an argument against CP that is not based on contingent facts, though I am quite certain that it is not a contingent fact that CP is cruel. I believe that CP is indissociably rooted in a social practice that until very recently was explicitly cruel: one of its principal reasons for being was to set an example of the infinite power of the state over the lives of its subjects. It is thus not surprising that in most of the Western world, the practice of CP died away along with the shift to democracy, and even more definitively with the somewhat later shift to a conception of the ultimate end of the penal system as a corrective one.

It is interesting that in many countries, such as Great Britain and France, the last vestiges of it survived as the form of punishment reserved for treason alone: this seems to me to have to do with the fact that under earlier, absolutist systems, CP, even though it was employed for all sorts of crimes, including murder, was in essence political, again in the sense that it demonstrated the power of the sovereign over the subjects. Long after murder and like crimes came to be seen as merely a criminal matter in Western Europe, and not a matter of state, direct crimes against the state could still be punished by death. The rest of the Western world eradicated CP by the late 20th century, even for treason, while somehow it managed to survive in the United States. Indeed it underwent a revival in the 1970s after some decades of desuetude, and in spite of a parallel commitment to individual rights at the political level, and to an understanding (which since the 1970s we may fear is only vestigial) of centers of punishment as 'correctional' institutions.

I apologize for starting off with genealogy, when the topic is whether there are any a priori arguments available against CP. I'm inclined to think that there is such an abundance of a posteriori problems with CP as actually practiced, and that these problems are so unlikely to be resolved, that the importance for any critic of CP of finding an a priori argument is not great. I do think however that there is at least one, so to speak conditional a priori argument against CP: if punishment is supposed to be correctional, then the death penalty is at odds with the purposes of punishment. Now a defender of CP might argue that there is a 'correction' of sorts that happens in the application of the death penalty: the wrongdoing is being corrected, or made right or absolved, by the death of the criminal. But of course this is not what our system of punishment claims to be in the business of correcting, nor is this kind of correction at all compatible with the rest of the legal and penal philosophy that our society claims to uphold.

Very broadly, it seems to me that the survival of CP in the United States leaves us with the impossible task of making two different conceptions of justice fit together: one that is based on the absolution of cruelty through an equal measure of cruelty (this is an ancient conception, and it implies a cosmology few people would explicitly assent to if pushed), and one that rejects cruel and unusual punishment out of hand in view of the fact that it is not conducive to the improvement of the individual criminal. This, then, is my first stab at meeting criterion (2) above. Death is different because it requires the destruction of the target of correction (it also requires the destruction of evidence, which is strictly prohibited in all other cases, a fact I might have occasion to return to later). But our system of punishment is a correctional system, not a restorer of cosmic balance through ritual sacrifice. Trying to be both at the same time leads to absurd results, as illustrated most vividly, I think, in the practice of swabbing the prisoner's arm with rubbing alcohol in order to sterilize it before lethal injection. As for criterion (1), perhaps Jerry can say a bit more about what would have to be shown in order to meet it.

G.D. I certainly agree that there are sufficient practical difficulties with the existing system of CP that we do not need an apriori argument to justify immediate abolition. Why then should I care if there is an apriori (I prefer non-contingent) argument as well? Qua reformer I do not. Qua philosopher that’s the business I am in. I want to know whether there is something about the very nature of CP which justifies its abolition. By analogy, it may very well be that eating animals is bad for us, both in terms of health and in terms of its ecological effects. But, while I am not (yet) a vegetarian, were I one I would want to say that there is something wrong with killing animals, even if this practice in fact made us slightly healthier and was ecologically sound. I am looking for an argument which says that even if we had reliable evidence that for each murderer executed we would prevent eight murders, and even if the process could be administered in a fair and equitable fashion, it would nevertheless be required that we abolish CP.

Justin presents a conditional argument: if the justification for CP is some kind of corrective notion, then CP cannot be corrective. True enough. Others have argued that even if the conditional is, “If the justification for CP is retributive, i.e. its function is see that people get what they deserve, then CP cannot be retributive.” Although this sounds somewhat paradoxical I accept this claim as well. For it requires showing that what murderers deserve is to be killed, as opposed to LIWP at the lesser-punishment end or torture followed by execution at the greater-punishment end. Although I don’t have the space to argue it here I believe that there are a number of reasons for thinking this cannot be established. But what I am looking for is an argument that there can’t be a successful argument of the form, “If the justification of CP is X, then for any x, the conclusion cannot be that CP can be justifiable on the basis of x.” I want to know why for any reasonable theory of punishment –and unless we are abolitionists about punishment itself there must be one– it turns out that LIWP is permissible (or some lesser sentence if you prefer) but death is nor permissible. Now there are some punishments that do fit this model. To take a widely disparate set of cases: punishing the spouse of the murderer; cutting off the murderer’s arms and legs; daily torture instead of execution; killing the murderer in order to salvage his organs for people who would otherwise die.

One needs in each of these cases an explanation of why the punishment is impermissible, and I believe I can give (different) explanations. My problem is that none of these explanations seems to rule out execution.

J.S. Interesting that Jerry should mention vegetarianism, because I too think this is a very revealing parallel case. And here, as with CP, I think the most one can hope to find is a conditional a priori argument in favor of it: if you are, say, a hunter-gatherer, I am not prepared to say that your killing of animals is wrong; if you are, like me, a member of a consumer society with fully nourishing plant-based foods available to you that allow you to avoid complicity in the gruesome system of factory farming (gruesome in a way that traditional spear- or bow-hunting is not), then you should be a vegetarian. I share the philosophers' desire to find non-contingent reasons for avoiding certain practices, but I'm fairly convinced that as concerns both meat-eating and capital punishment, one searches for them in vain. The kind of contingent circumstances I'm focusing on are however relatively wider-scoped than the ones that are usually adduced against CP (fails to deter, is administered unfairly, and so on). I am saying that it fails to cohere with other values to which we are supposedly committed, and that even if it did deter, could be administered fairly, could result in the prevention of eight other murders, and so on, it would still be wrong for the reason that it conflicts with these other values.

What are these other values? One, which I've already mentioned, is a commitment to the penal philosophy of correction. This only became orthodoxy in the 19th century, and has been rapidly eroding since the 1970s. Another (in the United States, anyway) is the commitment to upholding the US Constitution, which prohibits cruel and unusual punishment. There has of course been a great deal of controversy over what this phrase means, and whether any instance of CP is by definition a violation of this prohibition. In actual fact, the US justice system has consistently taken seriously the argument of the defense that a given method of execution qualifies as cruel and unusual. This is in part what is responsible for the continual migration from one method to another, even though each (starting with the guillotine) is introduced as the new and improved humane way of carrying the punishment out. What this shows is that there seems to be general agreement that there is a difference between the taking away of a criminal's life, on the one hand, and the killing of the criminal on the other. These are conceptually distinct, even if the one always implies the other. The general line of thinking in the US has been that the first of these is not in principle cruel and unusual, yet it has nonetheless proven impossible so far to find a way of carrying out the second that is not cruel and unusual. I am strongly inclined to believe, however, that taking away a life is cruel, if not unusual, whatever the method of killing, and whatever the crime for which it is punishment (mutatis mutandis, I similarly do not believe that free-range cultivation of beef or lamb takes away all of the moral concerns about carnivorism). Why do I think it is cruel? Because it is motivated by a desire to see the criminal suffer in the way that his victim has suffered (the fact that we can't really flagrantly make him suffer by, say, torturing him to death, results from the conflict between the two conceptions of justice of which I've already spoken), and I just don't know what cruelty could be other than the desire to see suffer.

No other currently legal form of punishment is motivated by this desire. Amputation or daily torture, as Jerry mentions, are not permissible, and Jerry already believes there are good reasons for this. I believe that the reasons for keeping torture illegal are not much more compelling than the reasons to ban CP: both are motivated by the desire to see suffer, to which our society, in its constitutional ban on cruelty, has a contingent –they did not have to bring this up in the Constitution at all– but pretty deep –this part of the Constitution stems from an earnest attempt to reflect the principles of natural justice, so far as we can tell what those are– opposition.

LIWP, in turn, is the default punishment not because it is commensurate with the crime –no non-cruel punishment could possibly be that– but because it prevents the criminal from committing more murders and, one hopes, gives him the rest of his natural life to seek absolution by changing his heart and his deeds. This means that, in effect, commensurateness to the crime is not a viable principle of punishment in a system that forbids cruel punishment. Criminal deeds, after all, are cruel. But cruel punishment is forbidden. Therefore punishment cannot be commensurate to the crime. This leaves us with LIWP, which might not be emotionally satisfying to many affected by the crime, but at least is not in conflict with other of our society's basic commitments. We could of course just scrap those commitments, and reverse the ban on cruelty. What we can't do, coherentl y, is to try to uphold the two together.

In sum, so far: no non-contingent arguments against capital punishment found, but some –I hope– philosophically interesting discussion of the sources of the contingency involved.

G.D. First, I like the idea of wide-scoped vs. narrow scoped contingent circumstances. But I am not sure that Justin’s argument above is not a non-contingent argument. For if CP is, in its nature, cruel, and if cruelty is always ruled out normatively, then it looks as if this is the argument we are looking for. So what we need is a conception of cruelty which 1) is inherent in CP and 2) is ruled out on normative grounds. With respect to the second point we need not say that it is always ruled out. Philosophers are good at coming up with weird cases which show that there are exceptions to everything. But at least the connection should be almost universal.

The analysis of cruelty Justin gives is the following: “I just don't know what cruelty could be other than the desire to see suffer.” I have a number of problems with this. I think it is important to keep separate (sometimes) questions of motivation from questions of justification. I may be motivated to save the drowning child because his father is rich and I hope for a large reward. I am not a very nice person. But, surely, my saving the drowning child was justified.

So the issue is really: must a justification of CP have as an essential component a reference to the suffering that death causes? Well, undoubtedly some do. Any retributive justification does make reference to seeing to it that the offender gets what he deserves. And in this case death must be what he deserves. But there is a sense in which any theory of punishment assumes that what happens to the convicted is something that they do not want to happen, that they find unpleasant. For a deterrence theorist that feature is required to explain why the threat of punishment deters. But the point or purpose of the institution is not to make the guilty suffer. It is to deter. So at most suffering enters as a means to the end of deterrence. If that relation to suffering makes CP cruel it makes any punishment cruel. That is a reductio ad absurdum.

The phrase 'desires to see suffer' is misleading because it suggests that suffering is desired for its own sake. And that is not the case for the deterrence theorist. If, for some strange reason, sending convicted felons to Hawaii to lie on the beach were to deter future potential murderers from killing a deterrence theorist would be delighted. So I suggest that what Justin really means is that the desire for people to suffer for its own sake is cruel. And this seems to me a plausible idea. But on this conception CP is not inherently cruel on any justification. Only on very special ones, namely, very strong versions of retributivism.

I am however intrigued by another point Justin makes. He thinks our prohibition against torture is on the grounds of cruelty. And that seems right. I said that my reason for prohibiting torture did not extend to CP. But if I believe that then I must have some other reason for the prohibition of torture. Or think that there is a conception of torture on which it comes out as cruel but CP does not.

J.S. Now what I'm about to say might sound pretty close to what has been described as 'punishment abolitionism', but if that is the case, so be it. I do not think that a theory of punishment that holds that its unpleasantness is an essential part of it is a good theory of punishment, or at least I do not think that the unpleasantness of the punishment should be part of the reason of the punishment, even if it always accompanies the punishment. The reason for the punishment is correction. By comparison, I think that unpleasantness might be an ineliminable part of coronary bypass surgery, but the surgery does not exist for the sake of its unpleasantness. So on my account the retributivist approach is the wrong one with respect to any punishment. This is as true for, say, white-collar crimes as it is for murder. Here my point is not that I think retribution is wrong, but that it conflicts with the fundamental commitments of our society's explicit penal philosophy (which I endorse, but even if I didn't endorse it this would not change the fact that retribution is incompatible with it). So that is how we can deal with the retribution theorists. As for the deterrence theorists, I had thought that what we were looking for here was a non-contingent argument against CP, which is to say in part an argument that would stand up whether CP in fact deters or not. Now, whether it does or not is a contingent question, but while we are on the subject of contingencies it is worth pointing out that CP does not, in fact, deter. States in the US in which CP is legal have higher murder rates, on average, than states in which it is not legal. (Then again, other countries in which CP is legal, such as Saudi Arabia and North Korea, have relatively low murder rates, but it is almost certainly not the case that this is because of the deterrent effect of CP.) In any case, if we are speaking to the deterrence theorist on his own terms, it seems to me that we can only speak in terms of contingencies, and as it happens in the US context deterrence-based arguments are ungrounded.

So retributivism and deterrence theory are dispatched, and the remaining questions I need to deal with, I think, concern, first, my claim that the desire to see someone suffer is inherently cruel, and, second, whether CP is itself inherently cruel even if some people defend it because of a cruel desire to see people suffer. I said in my last volley that 'I just don't know what else cruelty could be than the desire to see suffer'. I was hoping, by putting it like this, to let it pass as a not-further-arguable intuition. If an argument is required, it will be one that gets us back to genealogy, to a consideration of the kinds of practices from which I take CP to descend. My thinking here is greatly indebted to Bernard Williams's Shame and Necessity (and, though I hate his positive conclusions, to Nietzsche's Genealogy of Morality), and perhaps in a subsequent round I'll have occasion to say more about this, but for now I think it takes us pretty far afield from our narrower task. Now, supposing that one of the motivations for defending CP is the desire to see suffer, and supposing this is in fact cruel, does it follow that CP itself is cruel? I think we are dealing with a case here that is essentially very different than that of the man who saves the child because of the child's rich father, in which we can clearly separate the character of the act from the motivation for the act. I do not think that you can find, in the current American context, a non-deterrence-based argument in favor of CP that does not base itself largely on the fact that it provides 'closure' to the victim's family. And I don't know what this closure could be other than the (partial) satisfaction of a desire for revenge, which might also be accompanied by some folk-cosmological assumptions about the way the second death 'balances out' the first one. So here the desire for infliction of suffering is not at all incidental, in the way that the desire for a reward is incidental to the saving of the rich man's child.

G.D. Let me summarize our agreements and disagreements at this point. We agree (1) that there is no evidence CP deters with respect to LIWP, and hence no deterrence argument can work; (2) that it is worthwhile to look for a non-contingent argument; (3) that there is an argument from the cruelty of CP where cruelty is defined as the desire to see another suffer; (4) that cruelty is necessarily wrong. We disagree (1) on the definition of cruelty. According to me, it has two interpretations; on one, I do not accept that CP is cruel; on the other I accept that CP is cruel but disagree that cruelty is necessarily wrong. We also disagree (2) on the question whether any deterrence theory must have as part of its reason for CP that it makes people suffer. I do not believe that it must.

Since we are looking for a non-contingent argument against CP I am going to assume, for the sake of argument, that CP does deter differentially. In particular, since assuming is cheap, I am going to assume that for each murderer executed eight potential murder victims are spared. (The number is not competely arbitrary. The recent study that Cass Sunstein –a distinguished legal theorist– thinks is fairly plausible claims this figure. I take no view as to whether he is right about its plausibility).

Justin's argument is that, even so, CP must be abolished because it is cruel. But if cruelty means that the deterrence theorist must desire for its own sake that murderers suffer then I disagree. Classic deterence theorists from Bentham on have always insisted that the suffering of the murderer is a bad thing, that it must be minimized, and that if we could deter without making offenders suffer that would be the best state of affairs.

Deterrence theorists accept that murderers must be made to suffer in the same way that –in your example– the heart surgeon accepts that his patient will suffer in performing a coronory bypass. It is an (unfortunately) necessary means to a good end. Just as you do not think the surgeon is cruel, or the operation is cruel, so I think that CP is not cruel if one's justification is deterrent.

Do you, Justin, agree with the above? If you do, is there another notion of cruelty?

J.S. Now I think our agreements and disagreements are starting to come clear. I'm afraid the list of disagreements is a little bit longer than the one Jerry gives above. First, I don't think cruelty is necessarily wrong, or at least I don't have any argument that could prove that it is. What I do think is that if a society is committed to not being cruel, then it cannot be a society that employs the death penalty. I have already said that I do not know what cruelty could be, other than the desire to see suffer. This was deemed insufficient, so I will add that, coming at the question from a different direction, a punishment is cruel that rules out the possibility of a criminal's rehabilitation over the course of his natural life (even if this rehabilitation were not to result in exemption from LIWHP, there are still many ways in which prisoners serving life sentences have been able to 'make something of themselves' within the very limited confines of prison life). Why? Because it lacks mercy. I might be accused of trying to import a religious virtue into a discussion in which this sort of consideration can have no place. But I think as an un-argued-for assumption, 'mercy is good' is at least as sturdy as 'cruelty is bad'.

It dawns on me that if we want something more than this sort of smuggled assumption, the thing to do would be to look at the history of legal reasoning about cruel and unusual punishment. Beyond our intuitions about what is cruel –and my intuition has proven less than universal– I don't think conceptual analysis alone can bring us to any shared conclusions about what cruelty is. The boundaries of the predicate '…is cruel' as actually used seem to me both vague and flexible, and the closest thing we can find to a definitive answer as to the range of this predicate's application would be to consult the tradition on which current jurisprudence is based.

One passing observation about the Sunstein-sanctioned study showing that eight murders are deterred for every one execution. I'm sure this is dealt with somewhere in the literature, but I'm wondering how one might respond to the concern that deterrence-based reasoning seems to open up the possibility of, or at least go one step further down the path towards, preemptive punishment of the sort envisioned in Philip K. Dick's excellent short story, The Minority Report. What if executing one man who, say, shows signs in adolescence of incipient violent psychopathy could, per implausibile, save eight hundred lives? It seems that on a deterrence-based justification of CP, the punishment is no longer really a punishment for the murder carried out by the particular criminal, but rather a punishment carried out in advance of any potential future murders. But here any justification of deterrence-based punishment would seem to hold up even when the person punished is not in fact guilty of a crime. (I suspect, by the way, that this is not so far from how CP actually works, and explains why it is applied so grossly disproportionately to a certain group in the United States whose members, historically, have been 'kept in their place' by violent means, whether the individual members of the group are in fact guilty of any crime or not.)

G.D. I think that the new premise that Justin introduces, the claims that no form of punishment which precludes the possibility of an offender's rehabilitation is justifiable, is an interesting one to explore further. My only objection is that labeling it 'cruel' may simply be a way of saying that it is wrong rather than bringing it under one particular way of being wrong with its own special features.

The other point I want to make is that any plausible deterrence-based justification of punishment will have components which are constraints on the ways in which the goal of deterrence will be pursued. This part of the theory will address issues such as the one Justin raises about whether preventive punishment is legitimate. Notice that, although it is not punishment and that is important, we do have preventive detention as part of our civil law, e.g., quarantine of infectious people and civil commitment of dangerous people.

Let me try another, non-contingent, argument against CP. It is an objection mainly to general deterrence, i.e. that we are entitled to kill those convicted of murder to deter others from committing future murders. Special deterrence, i.e. killing murderers to prevent them from killing again does not fall prey to this attack, but special deterrence has always seemed a rather weak rationale even on deterrent grounds. Someone serving LIWP is not going to kill anyone outside of prison –barring prison breaks– and while it might seem obvious that a prisoner serving a life sentence has nothing to fear from killing a guard or a fellow prisoner since there is no harsher sentence in the absence of CP, in fact there are many ways of making a life sentence much more unpleasant –complete lockdown, loss of all privileges, etc.– so that it is not surprising to find the empirical evidence comparing states which have CP with those which do not does not support a higher rate of assaults or killings within prison in the latter.

The argument against general deterrence is that our justification for killing convicted murderers is that we need to do so to achieve a desirable end –reducing the number of total murders– and that this is simply to treat each murderer as a means to promoting good consequences. His death is justified only by its being part of a causal process which deters others from killing. Treating people simply as a means is the well-known Kantian objection to all consequentialist views which justify failing to respect the rational agency of one person in order to achieve some –otherwise valuable– end.

Suppose that the murderer about to die asks, “What justifies you in taking my life?” The answer cannot simply be, “You have murdered someone,” as it would be on a retributive view. It has to be , “You have murdered someone and our threats to punish murderers cannot be credible unless we carry out the sentence. We need your death as part of a credible system of threats which we believe deters other murderers.” Now the right type of argument against this response –right in the sense that it is a non-contingent argument, which applies even if CP is an effective deterrent and even if the process of administration is fair– claims that there is an intrinsic feature of CP, use of persons as simply a means, which, if a certain normative theory is correct (Kant), forbids CP.

I have two objections to this theory. First, it is not true that CP uses convicted murderers simply as a means. Second, if the theory were true, it would apply as much to LIWP as CP. With respect to the first point I want to set out why a certain kind of deterrence view need not fall victim to the Kantian objection. This theory is an amalgam of various proposals that exist in the philosophical literature. In particular Phillip Montague and Daniel Farrell have put forward key elements.

Start out from the idea that (almost) all abolitionists accept that it is legitimate to use deadly force if that is the only way of stopping an aggressor from killing oneself or others. One plausible way of explaining this is the following principle: When X has unjustly created a situation such that either X or Y will die (X is about to kill Y) then it is fair that the death falls on X rather than Y. Call this the 'Shifting of Harm' principle.

Now turn to the institution of punishment. Assume that we live in a world in which some class of people will unjustly kill others. Assume that by setting up a system of threats –murderers will be executed– we can in fact deter some people who would otherwise murder. Note that there may well be normative restrictions on the kinds of threats that are permissible. For example, we are not allowed to threaten the children of offenders. We are not allowed to threaten people who are doing perfectly acceptable things, like wearing green on St. Patrick’s Day, with punishment if they continue to do so.

We announce this threat in advance. All that citizens must do to avoid the punishment is to not unjustly kill other people. In effect, we have adopted the Shifting of Harm principle to a set of (unknown) people who are not imminent threats. We say to them that the members of their group –the potential murderers– have created a situation in which either some of them will die as a result of our carrying out our threat, or some of us will die by failing to carry out the threats, and it is fair that the harm should fall on them.

Notice that this argument does, in effect, extend the liability to harm for the individual (who is always liable for his own harm) to the class of all potential murderers. It says that each offender is paying a price not only for his crimes but for the crimes of others. But this is not simply treating the executed murderer as a means to the good end of murder reduction. If he has had a fair opportunity to avoid being part of the group creating the risk of death to others, if he has been warned that this is what will happen, if the harm that is produced by the punishment –death– is really required to achieve the protection of others, then he has no complaint. Just as the aggressor about to stick his knife into you has no complaint that he is being treated simply as a means when you use deadly force against him.

As for the second objection above –that the argument, if valid, applies to any prison term as much as to CP– simply note that the objection is not specific to any particular penalty. It says that if the purpose of any punishment is general deterrence, i.e. preventing future crimes, then we are using the convicted to prevent the crimes of others and that is wrong. It would be just as wrong to lock burglars up to deter others from burglary. In short, this argument does not work but, if it did, would prove too much.

Let me summarize my position. I believe we ought to abolish the death penalty because we have no reason to think it deters and it is administered in an arbitrary and unjust manner. I am open to the possibility that there is a non-contingent argument against CP which would justify abolition even if it did deter and was administered fairly. I have not yet been presented with such an argument that I regard as plausible.

J.S. Well it looks like Jerry saved his lethal blows for the very end. He considers a fairly strong non-contingent argument against CP (that is, not a cherry-picked one), and proceeds to knock it down. The argument takes up Kant's principle that a human being should always to be treated as an end and not as a means, and says that any deterrence-centered defence of CP necessarily violates this principle. Jerry offers two major objections to this argument: first, that deterrence-based CP might be seen as a legitimate application of the Shifting of Harm principle rather than as a violation of the Kantian ban on using human beings as ends; second, he notes that, if the argument were to stand against CP, an analogous argument against any punishment for any crime would be just as strong.

As for the second line of attack, I say so much the worse for deterrence-based application of punishment (whatever the crime). If using human beings as ends is wrong, and punishment for the sake of deterrence uses human beings as ends, then deterrence-driven imprisonment for burglary is just as wrong as CP for murder. But does punishment for the sake of deterrence necessarily violate the Kantian principle? I think Jerry's invocation of the Shifting of Harm principle here is fairly sound, in that I definitely think it's legitimate to suspend the usual rule about not treating a human being as a means to an end if that human being is coming at me with a knife, and it could be the case that the state's threat to potential future murderers is just a somewhat more complicated application of this same principle. But I wonder whether there isn't something about the immediacy of the threat that gives the Shifting of Harm principle its legitimacy. After all, when a dude comes at me with a knife, I don't run through the arguments of the Critique of Practical Reason in order to decide what I should do. So my suspension of the no-men-are-means principle isn't really a suspension of a principle at all, so much as a failure (whether laudable or blameworthy) to uphold it in certain circumstances. But to extend the Shifting of Harm principle to potential future murderers is in effect to say: we intend to fail to live up to something to which we are otherwise committed.

Mutatis mutandis, my uneasiness with this reminds me of one of the objections, from whom I do not recall, to Alan Dershowitz's horrible proposal that the government start issuing 'torture warrants' to federal agents who found themselves in situations in which they could, by getting cruelly and unusually rough, extract information that might save the lives of hundreds or thousands. The objection went as follows: it is perhaps not that agents will never find themselves in such a situation, and perhaps if they do they should just go ahead and start torturing. But what we don't want is to enshrine into law the possibility of suspending what are otherwise our deep moral commitments.

This last point leads me to a concluding thought I have about CP. I think some people almost certainly do deserve to die, but for better or for worse there simply is no person or body that can be entrusted with the grave responsibility of killing them. For me, one of the strongest arguments against CP has not to do with what it does to the criminal who is punished, but what it does to those involved in the application of the punishment. It makes it possible for killing to be the normal carrying out of a bureaucratic procedure, rather than a transgression or a suspension of our ordinary commitments. That to me is more terrifying than the murder to which the punishment is a response: the murder was plainly a transgression, whereas the compensatory execution is allowed for in our books of law. This means that to uphold CP is to make killing normal, something that it is not even for the great majority of murderers.

To sum up: I agree with Jerry that no compelling non-contingent arguments against CP are to be found. This negative conclusion, I think, is of less significance to me than it is to Jerry. I, unlike Jerry, am not an ethicist, and my interest in CP is, as my earlier pieces on it reveal, far from academic. As a non-ethicist, it never crossed my mind before to look for a non-contingent argument against CP. As a contextualist historian of philosophy, throughout this dialogue on the possibility of a non-contingent argument against CP, I've found myself returning again and again to the questions: what could even count as a non-contingent argument? The closest thing we've got to such an argument here is one that relies on a certain normative theory (Kant's), but that seems very much to me like something that one could either accept or reject, and accepting it seems to have a lot to do with what kind of society one lives in, and when. So I just don't know where any deeper non-contingency is supposed to come from, though I recognize that my perplexity here could very well arise from the fact that specialists in different domains of philosophy are trained to look for different things. Perhaps whether there really is such a thing as a completely non-contingent argument as to what it is human beings ought to be doing could serve as the starting point for another dialogue…