In his latest book, The Ends of Harm (OUP 2011), our own Victor Tadros advances a fascinating theory of punishment, which is articulated in Chapter 12, made available by OUP here. Discussion of the chapter kicks off today with a critical precis by another of our own, Cecile Fabre, a professor at Oxford University. Professor Fabre's precis begins below the fold.
Theories of punishment fall into three broad categories. Retributivists believe, roughly, either that wrongdoers deserve to be harmed, and that harming them by punishing them is therefore good, or that punishment is a fitting response to offenders’ denial of their victims’ rights. Communicativists, for their part, hold that punishment is a means by which we, the community, on the one hand communicate to wrongdoers that they have done wrong and on the other hand get them to acknowledge their wrongdoing. Finally, deterrence theories of punishment hold, again roughly, that punishing offenders is justified as a means to deter convicted wrongdoers themselves from reoffending, and putative wrongdoers from committing their first offence.
The difficulty with deterrence, it is usually claimed, is that it cannot really explain the intuition that we ought not to punish the innocent even if such punishment has strong deterrent effects – an intuition grounded in the Kantian view that we are not permitted to use individuals as means only to our ends – in this case, the end of reducing the incidence of offending.
In his wonderfully rich book The Ends of Harm (OUP 2011), Victor Tadros seeks to reconcile deterrence with a properly modified Kantian view. If successful, his non-consequentialist deterrence justification for punishment would accommodate deterrence theories’ powerful insight that the consequences of punishment matter for justifying the infliction of this particular harm, but would not commit itself to the punishment of the innocent.
Tadros’ reconciliation strategy, which is at the heart of Chapter 12, comprises two key claims. First, the revised Kantian view (which he calls the means principle) holds that it is sometimes permissible to harm a culpable wrongdoer W as a means to avert threats to third parties from future wrongful harms at the hands of other wrongdoers. Second, punishment is one such permissible harm.
In defending the means principle (MP), Tadros helps himself to the ethics of defensive killing. His starting point is the following claim:
(MP1) It is sometimes permissible to harm W as a means to avert a threat which W culpably poses to his victim V1.
MP1 is uncontroversial. Suppose that I hire a hit man to kill you, and that the only way you can save your life is by using me as a shield against him. Surely you may do so: I am after all morally culpable for putting your life under threat, and so I may treat you as a means to my end. (p. 187.)
Note however that this important revision to the means principle can help justify punishment on grounds of general deterrence only if it is permissible, not just to treat W as a means to avert the threat he poses to V1, but also as a means to avert other threats. In other words, Tadros must show the following:
(MP2) It is sometimes permissible to harm W as a means to avert a harm which someone other than W himself culpably poses to V1.
(MP3) It is sometimes permissible to harm W as a means to avert a harm which someone other than W himself poses to another victim, V2.
The main reason why we may punish W to protect his victim from other wrongdoers is that W, by culpably harming V1, incurs a duty to compensate her for the harm she has incurred at his hands. Consider the following hypothetical scenario, which Tadros uses in ch. 8 to show that causing a threat is not necessary to liability to defensive harm, and which he revisits in ch. 12:
Double Hit Man 2. Evelyn hires a hit man to kill Wayne. Fred has also hired a hit man to kill Wayne. Both hit men arrive at the same time. Because of where they are standing, Wayne can only use Fred as a shield against Evelyn's hit man and Evelyn as a shield against Fred's hit man. He manages to do that, resulting in the deaths of Evelyn and Fred (p. 192, 274).
Tadros believes that Wayne may so act, on the following grounds. Fred and Evelyn are each under a duty to sacrifice their life to protect Wayne from the threat which they each have initiated: Evelyn ought to stand in between her hit man and Wayne, and Fred ought to do the same. Unfortunately, they cannot do that. So they ought to do the next best thing — namely get the other to take their place. As Tadros puts it, ‘in that case, Fred and Evelyn ought to form and execute an agreement to protect Wayne by fulfilling each other's duties to avert the threat that they are responsible for creating. In that way, Evelyn and Fred could avoid harming Wayne. If they cannot form and execute such an agreement, however, or they refuse to do so, Wayne can permissible use them as a shield to avert the threats that he faces. For were he to do so he would merely force them to fulfil their duty; to make themselves available together to avert the threats to Wayne.’ (pp. 274-275).
This is puzzling. The argument relies on the claim that there is no particular reason as to why Fred should perform his own duty rather than Evelyn’s, and Evelyn her own duty rather than Fred’s. After all, they each already are under a duty to give up their life to protect Wayne’s from their own hit man, so it seems appropriate to hold them under a duty to give up their life to protect him simpliciter, and in this instance from one another’s hit man. But one might think that there is a very strong reason to suppose that Evelyn is under a duty to give up her life only as a means to rescue Wayne from the lethal threat she poses, and that Fred is only under a similar duty — namely that it is a necessary condition for a wrongdoer to be liable to defensive force that he should be causally responsible for subjecting another person to a wrongful threat. Given that Tadros’ stated aim with his example is to show that causation is not necessary for liability, he cannot use as a premise the undefended claim that Fred and Evelyn stand in no special relationship to their own wrongdoing, on pain of begging the question.
Furthermore, Tadros construes Fred’s and Evelyn’s remedial duty, after they have initiated their respective threats, as a duty to agree that they will together protect Wayne by fulfilling each other’s duties. This is odd: in fact, it would seem that Fred’s remedial duty is, more straightforwardly, a duty to interpose himself between Wayne and Evelyn’s hit man, and vice versa, and that Wayne is permitted to enforce that particular duty (by using both of them as shields against the other’s hit man) if they are derelict. To be sure, one might think that it makes little difference whether the remedial duty is a duty to form and execute a joint protection agreement, or whether it is a duty to throw oneself in front of the other wrongdoer’s hit man. But I think it does. For a start, by conferring on those wrongdoers a remedial duty to agree with one another that they will each block the other’s threat, Tadros bolsters his case against the relevance of causation for liability since under the terms of this agreement each wrongdoer takes responsibility for the other’s threat. Remove the agreement from the picture, however, and what is left is the stark claim that, just because Fred culpably contributed to subjecting Wayne to a lethal threat, he is under a duty to sacrifice his life so as to ensure that Wayne survives another lethal threat which Evelyn contributes to posing him independently of Fred. I find that stark claim particularly counter-intuitive. Unless one is already and for independent grounds willing to reject the claim that causation is necessary for liability, then it is unclear why one should accept Tadros’ argument.
Suppose that Tadros does in fact succeed in showing that it is sometimes permissible to harm a wrongdoer W as a means to protecting V1 from future wrongful harms at the hands of other wrongdoers. Remember that, in addition, Tadros must show that MP3 is correct:
(MP3) It is sometimes permissible to harm W as a means to avert a harm which someone other than W himself poses to another victim, V2.
The whole point of deterrence-based punishment is precisely to punish wrongdoers as a means to avert future harm to future people by other future wrongdoers. Unless Tadros can provide support for MP3, together with support for the claim that punishment is one such harm, the project collapses. Let me start with the latter. Tadros acknowledges that punishment is one such harm only if one generally may punish wrongdoers as a means to ensure that putative future wrongdoers are less likely to form intentions to commit offenses (281-282). Tadros claims that the institution of punishment does in fact have that effect. Unfortunately, however, he does not provide empirical evidence for it. Given that worries about whether punishment works is another central plank of the case against deterrence, such evidence would have strengthened his case considerably.
In any event, and more importantly, Tadros must also be able to convince his critic of the soundness of MP3. The following case highlights the difficulty. Suppose that on Monday at 10am, Fred subjects Wayne to a lethal threat, which Wayne successfully and durably parries. At 5pm on the Friday, Evelyn, who is not aware of Fred’s attempt on Wayne’s life, poses a lethal threat to Robert. Robert can save his life only by using Fred as a shield. On Tadros’ account, it seems that Robert may so act, since Fred would be under a enforceable remedial duty to offer himself up as a shield against Evelyn. But this is tantamount to holding Fred under a duty to use himself as a means, and as a means only, to Robert survival just in virtue of the fact that he posed a threat to Wayne on Monday, and even though he no longer does so. Ex hypothesi, however, Fred is both causally and morally innocent of the lethal threat which Evelyn poses to Robert on the Friday. True, he is not innocent simpliciter. We might say, thus, that he is narrowly innocent (vis-à-vis Robert) but widely guilty (on account of his previous threat to Wayne.) It seems to me that to insist that he may thereby be harmed if this would thwart Evelyn’s threat to Robert is to deprive the Kantian prohibition on harming the innocent of much of its bite.
Tadros might perhaps concede that one ought not to kill the narrowly innocent in punishment, for that clearly is to use them solely as means, but that we may nevertheless punish them to a lesser degree if they are widely guilty of previous wrongdoings. More precisely, he might say that we may not use the narrowly innocent as means only, but that we may use them as means (to protect other future wrongdoers’ future victims) provided we use them as ends as well. His principled objections to capital punishment in the book’s last chapter (e.g., 307ff) combined with an account of the kinds of harm imposition which are compatible with respecting people as ends would seem to support this interpretative move. However, the main and intuitively acceptable form of punishment for serious wrongdoings, namely freedom deprivation for a number of years, does belong to the category of ‘using as means only’ when carried out for instrumental reasons: if I kidnap you and lock you up in my house for five years on the grounds that I need you for some admittedly important purpose, we would consider this as unacceptable, on precisely those Kantian grounds. Likewise with ‘standard’ imprisonment.
In short, either it is permissible on deterrence grounds to punish the narrowly innocent to the degree that many of us think is acceptable, in which case one must give up on the Kantian view as a price for endorsing deterrence, or it is not permissible to punish the narrowly innocent, however widely guilty they might be, in which case one must give up on deterrence. Either way, Tadros’ reconciliation strategy fails.
Great chapter, Victor. I want to focus on a key bit in which you say (283) “We have very strong reasons to prefer using wrongdoers as a means to deter others than we have to use them as a means to avert imminent threats. We can predict that were it not for the deterrence effects of the criminal law, there would be very many wrongdoers, and their lives, as well as the lives of those they harm, would go significantly worse as a result.”
I want to ask about the ideas in this section from two angles. First, you discuss the worry that general deterrence doesn’t avert imminent threats, and what you say makes sense: harming the criminal now, rather than waiting until an actual threat is posed by another, allows us to harm the criminal in a more controlled fashion, benefiting him; and it also benefits people generally by ensuring that they are “less likely” to be morally corrupt. But it seems to me that the imminence is less a worry than the inevitability is. That is to say, what matters is that the threat being averted in punishing the wrongdoer would indeed have happened otherwise. But I don’t yet see that this is the case. In particular, I don’t see why we should think punishing a specific person for a crime eliminates any other specific threat of harm that would definitely have occurred otherwise. And if it wouldn’t have, how could punishment of that particular person be justified on the duty-based model? (In a way, this is akin to worries about rule worshiping for rule utilitarians.)
Second, if what you say is correct, wouldn’t we have even stronger reasons to prefer using wrongdoers as a means to deter others in particularly dramatic ways, e.g., torture?
Thanks very much for the excellent questions David. Your first worry might be one of a number of worries. One worry is that punishing one less wrongdoer will make no difference to general deterrence, and we have good reason to let some wrongdoer go unpunished. Suppose that the deterrence effect of punishing 999 murderers to degree n is exactly the same as the deterrence effect of punishing 1000 murderers to degree n. If n is the only available magnitude of punishment, I doubt that we should punish all 1000. We may have some reason to punish the extra murderer – we might do so for communicative reasons, for example – to attempt to ensure that the extra murderer recognizes that what he has done is wrong. Is your worry that it is implausible that we should not punish the 1000th murderer in this case?
On your second worry, whilst torturing is almost certainly always wrong in the real world, I wonder whether you think that there are fundamental moral constraints against torturing. Consider:
On a three person island, D rapes V. V manages to imprison D. V now faces a threat of being raped again by X. V threatens X that if he rapes her, she will later torture him. She can make her threat credible, deterring X from raping her, only by torturing D to some degree (say by breaking one of his fingers). Do you think it clearly wrong for V to break D’s finger?
Why do I think torturing prisoners clearly to be wrong in the real world? Simply because the evidence strongly suggests that humane rehabilitative forms of criminal justice are better than barbaric ones in almost every respect.
Thanks, Victor. Yes, I had a few worries packed into the first concern, one of which was along the lines you’ve articulated. On the rule worshiping objection to RU, the worry is that, if the justification for the rules was utility maximization, but adherence to a rule in some particular circumstance failed to maximize utility, then adhering to the rule constituted a form of rule worship and was without justification (of the right kind). Similarly, if the motivation for punishment is deterrence, but punishment in any specific case would fail to deter, then we’re left with deterrence worship without justification. What you seem to suggest is that we shouldn’t adhere to the “rule” in this case, that while there may be some other reason (communicative) to punish, there’s indeed no deterrence-based reason to do so. This does indeed seem worrisome to me, as I would think the state ought to be justified in punishing every criminal violator on precisely the same grounds as every other violator of that kind. Doing otherwise undermines equality before the law.
But then the other worry in my remarks is relevant: on what basis should we think that there is any connection between the punishment of any single murderer and any other single person (or persons) deterred from forming the intention to harm, say? In other words, I don’t doubt that a policy of general deterrence may prevent people from forming criminal intentions, but it seems that the details of your argument require there to be direct connections between specific individuals punished and other specific individuals who are deterred. Perhaps I have misread you on this point, though.
Regarding the torture scenario on the island, ugh. I’m not sure what to think. Ratchet up what’s required for the credibility of the threat, however (e.g., bamboo shoots in the nails every hour, punchings to the brink of death, then resuscitation), and I really start to feel that it would be clearly wrong. But the credibility for threatening some would-be criminals could be that sickeningly high.
Thanks Dave
On the equality point, consider this:
D and D2 assault V. V faces a threat of a further assault from X. The only punishment available imposes some cost, n, on the person punished. The assaults of D and D2 are sufficiently grave to render them liable to be punished to degree n. V could punish D, D2 or both to degree n. X will be deterred if either D or D2 is punished. Punishing both also deters X, but has no extra instrumental benefit. There are no further instrumental or communicative reasons for punishing D or D2 (there is no one else to deter, and each is already repentant, or each is intransigent so will not recognise what they have done is wrong even if punished)
V may clearly punish D or D2. The equality argument seems to imply that if he punishes D he must also punish D2. I doubt that this is true. Suppose that V picks a fair mechanism for selecting between D and D2 – he flips a coin, and D is punished. We now need a reason for V to punish D2 even though doing so protects no one. Perhaps it might be argued that it is obnoxious that D and D2 are left unequally placed. But punishing D2 would level down for the sake of equality. I doubt that there is any reason to level down, but even if there is, I think it is too weak to justify punishing D2. Furthermore, D is also unequally placed with innocent people, but we lack a reason to punish them.
If we are to justify punishing D2, then, I think that it is not equality that we should appeal to, it is desert. Retributivists can find a reason to punish D2. I don’t think retributivism is true though (but that’s another chapter).
You appealed to equality before the law as the relevant principle. I don’t think, though, that it makes a difference if the island has a law prohibiting assault, with an authority, A, who will do the punishing. The law will already have treated D and D2 if it selects them for punishment in a fair way.
Reasons of equality do seem very important in the real world, though, because in the real world we normally face a choice either of punishing D but not D2 to degree n or punishing both D and D2 to some degree less than n. Reasons of equality militate powerfully in favour of the second option. We reduce the punishment level of all offenders rather than imposing severe punishments on selective offenders (those who know their Foucault might find this historically interesting…)
On torture – there is a liability threshold on how much torture can be imposed on the offender even in the idealised (sic) scenario I outlined. That depends, I think, in part (but only in part) on the ex ante cost that it would have been permissible to impose on D as a means to avert the threat that he poses. Suppose that D hires X to rape V. X will be deterred from carrying this rape out if V can make her threat of punishment to X credible. She can do this only by punching D to the brink of death. In this case, D would be used as a means to avert a threat that he is responsible for. We might also balk at V punching D to the brink of death.
Suppose that we don’t. It still doesn’t follow that D can punch V to the brink of death in the earlier scenario. There are differences between the scenarios that are morally salient. The fact that D is not responsible for X’s threat in the earlier example may reduce, to some degree, his liability to be harmed.
I am very grateful to you Cecile for your excellent responses to ch.12 of The Ends of Harm. There is a great deal in these comments to address, and I want to keep my posts reasonably short. Let me begin by responding to your analysis of the compensation argument and your response to the analysis that I offer of Double Hit Man 2 and offer a defence of MP2.
You present the compensation argument and the Double Hit Man 2 argument as more closely related than I intended them to be. They are intended to provide independent ways of showing that a familiar principle, what we might call the responsibility principle, is false.
The responsibility principle holds that it is permissible to harm a person as a means to avert a threat only if that person is responsible (in some sense) for creating that threat. This principle is familiar, for example, from Jeff McMahan’s work on self-defence.
The compensation argument challenges the responsibility principle in the following way. Suppose that D wrongfully harms V to degree n. Typically, and to simplify a bit, this will be true if D renders V worse off than D would have been had D not wronged V. Where this occurs, it is uncontroversial that D has a duty to compensate V; something like:
The Duty of Compensation: D has a duty to fully compensate V. Again, speaking crudely, D fully compensates V if he renders V as well off as he would have been had D not wrongfully harmed V.
Now, suppose that V faces a further threat of harm from X to degree n, a threat that V would have faced whether or not D had wrongfully harmed V. Suppose also that there is no other way for D to render V as well off as he would have been had D not wrongfully harmed him. In that case, D has a duty to avert X’s threat. For if D averts X’s threat, D renders V as well off as he would have been had D not wrongfully harmed V.
If it is true that this duty is enforceable (and duties of compensation are surely often enforceable), it is a fairly smooth ride to Cecile’s MP2.
The analysis of Double Hit Man 2 probably needs more work than I gave it in Ends. The idea, very briefly stated, was something like this. Fred has a very stringent duty to ensure that his hit man was diverted. Evelyn has an equivalent duty. If each could ensure that the threat that their hit man poses would be diverted only by forming and executing an agreement that Evelyn would divert Fred’s hit man and Fred would divert Evelyn’s hit man, they would have a duty to do so. A person’s complaint about being used as a means to an end is significantly eroded if that person has a duty to serve the relevant end. As Evelyn and Fred together have a duty to ensure that their hit men are diverted, they can be used as a means to divert the two hit men.
Now, Cecile objects to this analysis on the grounds that I have not shown that Fred and Evelyn lack a special relationship to their own wrongdoing. But, on the contrary, it is precisely in virtue of the fact that Evelyn has a very powerful duty to ensure that the hit man that she has hired does not kill Wayne that she would have a duty to form and execute an agreement with Fred of the kind outlined above. The argument thus relies on, rather than denying, the special relationship that a person has to ensure that her duty is done.
It is worth noticing that, unlike the compensation argument, the permissibility of using Evelyn as a shield in this case depends on the fact that Fred will be used to divert Evelyn’s hit man. This argument would not show that the fact that Fred poses a lethal threat to Wayne would provide sufficient grounds to use Fred as a shield to divert Evelyn’s hit man were Evelyn not used in the same way. The stark claim that Cecile finds counterintuitive is thus not implied by this argument. However, the compensation argument above, does imply that a person who causes another harm does have a duty to avert other threats that the victim faces.
Cecile is right that chapter 12 offers little defence of MP3. I aim to defend MP3 in chapter 13. I don’t intend to imply, though, that it would be permissible for Robert to use Fred as a shield in Cecile’s last case. It is not obvious why Fred, in this case, has any remedial duty at all as his threat no longer exists and has not been realized. He is only an attempter. It therefore does not follow from the arguments offered that Fred can be used as a shield to protect Robert.
There is a difficult question about whether attempters incur duties to protect as a result of attempting. Some doubt that the duty view has plausible implications for criminal attempts. There is a great deal to say about this. But I am yet to be persuaded that this thorny issue provides sufficient grounds to give up on the attempt to reconcile the means principle with deterrent punishment.
Victor: I may indeed be thinking that there are retributive reasons for punishing both D and D2 in your example, in which case I suppose I need to be disabused of that notion by reading the relevant earlier passages in the book. But at any rate, what matters now, since you’re not a “deterrence worshiper,” is my second stated worry in this area, namely, that there has to be a direct causal connection between individual punishments and some specific, deterred individual(s), but what reason do we have for believing this will be the case? In other words, you may get your needed elimination-of-inclinations-to-threaten amongst the populace from general deterrence, where the punishment is equally applied to every equal offender, but that doesn’t seem to be the move you’re now making, which seems to be drawing a necessary connection between individual acts of punishment and specific individual(s) deterred thereby. I hope this is making sense.
Your remarks about the torture case bring to mind another question I had about your treatment of the Ferrell argument, which relies on punished agents being responsible for the threat they are being punished to prevent. Here you say, “The fact that D is not responsible for X’s threat in the earlier example may reduce, to some degree, his liability to be harmed.” But what do you mean by “responsibility” here? What if he’d been merely causally responsible? Would that be sufficient to reduce his liability to harm? (Your answers here may produce other questions about responsibility generally. I’ll wait and see what you have in mind.)
Hi Victor,
A quick question about the final two paragraphs of your post on the 28th in response to Cécile. You wonder why, in Cécile’s case, Fred has any remedial duty when his threat has not and will not be realized. He is, as you say, only an attempter. But suppose in your Double Hit Man 2, Wayne first uses Evelyn as a shield against Fred’s hit man. Fred’s hit man no longer poses a threat (presumably he is now out of bullets). Wayne now faces a threat from Evelyn’s hit man that he can avert only by using Fred as a shield. You think it’s permissible for Wayne to use Fred manipulatively in this way, as Fred has an enforceable duty to protect Wayne from the threat posed by Evelyn’s hit man. But Fred’s threat has not and will not be realized. He is now a mere attempter. He seems, in other words, relevantly like the other Fred in Cécile’s case of Fred and Robert. So how can it be permissible for Wayne to use him manipulatively?
You always have a clever reply to any objection I make to your arguments and I’m eager to know what it will be in this case.
Hi Dave
Good question! It is true that the argument moves a little bit quickly between justifying individual acts of deterrence and justifying deterrence in an institutional context. I’m not sure that problems arise through uncertainty, though, where we have reasonable expectations about the effectiveness of punishment.
For example, suppose that we had robust empirical evidence that punishing 10 offenders of a certain kind would prevent, on average, 10 crimes to 10 victims, punishing 9 of the crimes to 9 victims, and so on. We would not be able to show to any offender that his punishment protected a specific victim from a specific potential offender. But I don’t think that the offender could complain about being punished on these grounds. This would be the best that we could do to ensure that the protection that he has a duty to provide is provided.
Now, Cecile, in her comments, rightly complains that I have not provided the empirical evidence that criminal justice systems are effective in deterring. She is quite right, of course, and I’m no expert on this. I doubt that anyone is an expert on the appropriate question, though. There is a great deal of empirical evidence, which is a bit contested as I understand it, about whether increasing punishment rates reduces crime. That evidence may provide reasons to reduce the severity of punishment (which, of course, would be especially good in the US context, which is a world leader in preposterously harsh punishment). But what I am interested in is whether punitive systems can be justified at all, even with relatively mild punishment. That is very hard to test empirically – what would be the effects of complete abolition on the crime rate in large liberal societies? I don’t know of any work that examines this question in a serious way empirically, and I doubt that it could be done, as no large liberal society has ever operated without a criminal justice system. I am not completely confident, but I can’t say that, were I able to, I would abolish the criminal justice system altogether. I am just not confident that alternatives would be sufficiently effective in keeping the crime rate down. If that is right, I don’t think that offenders can complain about being harmed to contribute to the threat reduction given the protective duties that they incur through offending.
Hello Jeff
Great question. I’m not sure I have anything clever, let alone right, but let me try.
When Wayne uses Evelyn as a threat to avert Fred’s hit man, Evelyn is used as a means to ensure that Fred’s duty not to kill Wayne is not violated. Think of this as a kind of benefit not only to Wayne, but also to Fred – Fred is now not a murderer in virtue of the fact that Evelyn has been used. How could we justify using Evelyn as a means to ‘benefit’ Fred in this way? We might think: only on condition that she receives a reciprocal ‘benefit’ from Fred – that he will be used as a means to ‘benefit’ her in the same way by ensuring that she is not a murderer. Harming Fred as a means, we might say, is justified because harming him to avert Evelyn’s threat is a necessary condition for the permissibility of harming Evelyn as a means to avert Fred’s threat. And Fred has a very powerful reason to want Evelyn to be harmed in this way – so that he does not violate his duty not to kill Wayne. (This is like a restatement of the ‘agreement’ argument without appeal to hypothetical contracts – I am grateful to Massimo Renzo for pointing out to me that the agreement might turn out to be redundant).
The difference in Cecile’s case, is that no one else was harmed to avert Fred’s original threat. So Fred doesn’t ‘owe’ anyone. Compare your own attempter case where X is about to kill me, Y is trying to kill me but his gun is jammed, and I can avert the threat from X only by using Y as a means. The kind of argument offered above can’t be offered to Y. For whilst X needs Y to be used in order that X does not violate his duty not to kill me, Y does not need X to be used in return. That is why it is more difficult to justify harming Y in this case.
As thought-provoking as ever, Victor! Here’s one of the questions that struck me when reading your chapter, namely “Does practising deterrence punishment involve welcoming the harm (up to some threshold) that it imposes on wrongdoers; and if so, is this phenomenon relevant when deciding whether wrongdoers’ remedial duties explain their liability to be harmed?”
The question arose because on my own (perhaps overly dark) understanding, using punishment as a deterrent involves welcoming wrongdoers’ suffering, or deprivation, because it is an effective instrument to change the attitudes and behaviour of potential wrongdoers. If so, the kind of harm involved in deterrence punishment seems quite different from the kind you often discuss. In many of your examples, wrongdoers are morally required and liable to be forced to provide protective services for others not because of but despite the harm they are likely to suffer in the process. (Even if we can sometimes complain about criminals being treated too leniently to provide enough deterrence we surely can’t complain about Evelyn and Fred wearing body armour in an attempt to minimize the level of harm they suffer when shielding Wayne.)
Assuming support for deterrence punishment does involve welcoming the harm it imposes on wrongdoers, I worry whether the kinds of remedial duty you need to rely upon in order to justify the institution also require that wrongdoers themselves welcome their own suffering. If so, this type of duty seems significantly more difficult to defend than the remedial duties you more frequently mention because of the type of self-destructive attitude its endorsement involves. It’s one thing to expect Evelyn and Fred to make (heroic) amends for their wrongdoing by trying to shield Wayne even at the cost of their own lives. It’s another (for me, less credible) thing to expect people like them to welcome their own suffering providing that others are likely learn from it. If the appeal to remedial duties relies on this type of very demanding duty then we might need to look elsewhere to explain wrongdoers’ liabilities.
It would be an advantage, then, if you could provide a less dark understanding of deterrence than the one I took for granted, or show that your justification of punishment relies on remedial duties that don’t demand wrongdoers take quite so hostile an attitude to themselves. (Sorry if I have overlooked a solution you supply elsewhere.)
Andrew
this is an excellent question, and I don’t address it in Ends. I hadn’t focused very sharply on the difference that you mention.
I take it that you don’t think that it is always wrong for a person to welcome her own suffering if others will learn from it. Consider:
ACTOR. D hires a hit man to kill V. D is now realises that it was wrong for him to hire the hit man. The hit man will not go through with the hit if he thinks that V is very powerful. There is no other way for D to prevent the hit going through than to create the impression that V is very powerful. He can do this only by breaking his own thumb and pretending that V did this.
Do you think that D lacks a duty to break his own thumb in virtue of the fact that in doing so he welcomes his own suffering for instrumental reasons? I don’t find that very plausible. After all, the alternative is that D fails in his duty not to kill V.
I agree, though, that there is something dark about having a duty to welcome one’s own suffering (retributivist and communicative views of punishment also have this feature I think).
Justifying deterrent punishment to offenders on the duty view does rely on this idea. The above case suggests that this feature is not always a decisive reason against deterrent punishment. It may affect the magnitude of punishment that may be imposed on offenders – perhaps the magnitude of harm that D is required to impose on himself in ACTOR is less than the magnitude of harm that a person is required to impose on himself in the standard hit man cases I consider in Ends.
I suspect that we can help further to justify punishing offenders in this way if we can demonstrate that the harm that we impose on them is aimed not only at the good of potential future victims but also at their own good. They would then have a self-regarding reason to welcome their suffering. It would improve our justification of punishment if we aim at deterring offending on condition that we also aim to rehabilitate offenders. Do you think that this helps ‘lighten’ the dark side of deterrence? In Ends, I suggest that we have good reason to pick rehabilitative methods of punishment when they don’t set back our deterrent aims very significantly. This would be another reason to endorse that view.
Hello, I very much enjoyed both the chapter and the exchange so far. I have a question about the compensation argument.
Victor proceeds in a number of steps, but here I intend to consider only two of them. First, Victor argues that if William wrongfully harms David (say, by assaulting him), William acquires a duty to compensate David. A plausible way to characterize this duty, according to Victor, is as a duty to protect David from similar harm he might suffer in the future (both at the hands of William and at the hands of other possible wrongdoers). Notice that at this stage of the argument, William’s duty only correlates to David’s right to be protected. The second step of the argument that I’m interested in here is the one in which David’s right is extended to cover other people that David cares about, as well as David’s fellow citizens. [Incidentally, is there a reason to stop there? Could we further extend the right to cover non-citizens as well? Would that justify international punishment according to the Duty View?]. Victor argues that the state is justified in assisting David in harming William as a means to avert future threats that William, as well as other potential wrongdoers, might pose to: 1) David himself, 2) his loved ones, 3) his fellow citizens.
This suggests that David is the primary right-holder of the right to protection, in the sense that William’s duty is owed primarily to him, and only derivatively to his loved ones and to his fellow citizens. But if so, there seem to be cases in which the Duty View is not capable of justifying the right of the state to punish William. For we can think of cases in which the best way for William to discharge his duty toward David is not to be punished for the sake of general deterrence. Perhaps David leaves in a dodgy neighbourhood, and he’s more likely to be protected (and to have his family protected) effectively from future threats if he has William acting as a security guard for his family. If so, by punishing William the state would not be assisting David in exercising his right. Rather, it would be preventing David from exercising his right in the most effective way. If this is correct, there seem to be cases in which David’s right to be protected cannot be extended to cover David’s fellow citizens in the way suggested by Victor. These are cases in which the best way for David to exercise his right (as well as to fulfil his duty to protect his family) is to ignore his fellow citizens and not to have William punished.
As far as I can recall, Victor does not address this objection. However, elsewhere in the book he discusses the problem of why the state can punish wrongdoers without the consent of the victims (pp. 297-9), and one of the answers offered to this problem could be invoked by him here. The answer is the following: while David has a right to be protected, he also has a duty to protect his fellow citizens from criminal wrongdoing, provided he can do so at a reasonable cost. This is why David is not simply at liberty of having William punished, but he has a duty to see him punished.
This is a plausible answer, but surely the cost condition cannot be simply assumed to be always met. If David lives in a dodgy neighbourhood, in which the police doesn’t dare to enter and in which the punishment of William is likely to have little deterrent effect, giving up the protective services of William as a security guard might be too costly for David.
Moreover, I wasn’t sure whether the compensatory duties incurred by wrongdoers must be necessarily discharged by protecting their victims. Could they be discharged in some other way? For example, by providing the victim with fundamentally important benefits? Imagine that David’s business is going very badly, to the point that he is not able to provide for his family any longer. Imagine also that having William working for David for free in his business for x number of years would allow David’s business to recover, so that he will be able to provide for his family. Here again, it seems that it would be very costly to give up William’s services for the sake of general deterrence. Can the Duty View explain why William should nonetheless compensate William by being punished?
Hello Massimo
Excellent questions. I think that there are two separate issues here.
1) in principle, is the only way for offenders to satisfy the duties they owe to their victims and others punishment, or might they be required to do something else (either something protective or non-protective)?
2) how should the ‘goods’ of punishment (or other duty satisfaction) be distributed amongst potential beneficiaries.
I am more confident about the answer to 1) than 2). The answer to 1) is that there is no special reason to prioritise punishment over other forms of protection. Obviously, there are good pragmatic reasons why we would not want convicted offenders to act as unpaid security guards, but in principle there would be nothing wrong with this. Compare:
WOLVES: A society has relatively low levels of offending. However, it is subject to many wolf attacks from the surrounding area. This society needs people to patrol the borders to prevent the wolves from coming in. Doing this is boring and time consuming, but because protective clothing can be made available, it is not very dangerous. Too few people are willing to do this work.
In such a society, requiring offenders to patrol the borders would be permitted even if doing this deterred fewer offenders. There is only a relatively weak reason to reduce the offending rate compared with the wolf-attack rate. If the wolf attack rate can be reduced much more by requiring offenders to do this, this policy would be acceptable.
I am less confident about the complete answer to 2). In standard cases, as you note, the ordinary duty of rescue would provide a good reason to choose deterrent punishment over forms of protection that benefit the victim but not others. We can extend these ideas in some ways. For example, citizens together share the costs of apprehending and convicting offenders. They also share the costs of ensuring that duties are satisfied in an acceptable way (unsupervised private punishment is ruled out simply in virtue of the fact that it is likely to be unjust). Hence, citizens can expect to receive some of the benefits of punishment.
There may also be straightforward distributive reasons not to allow private security. Those whose victims are apprehended will have greater security resources available than others, and they may be expected, for reasons of distributive justice, to distribute those resources rather than being entitled to all of them. But there is much more to say about this than I have said.
Your question is in some ways similar to a challenge posed by Kim Ferzan in her essay on Ends in Law and Philosophy. My response is forthcoming.
Thanks, Victor. This is very helpful. Let me try to push you both on 1) and 2).
About 1): although WOLVES involves again protective measures, you mention that in principle it would be permissible to use wrongdoers to provide non-protective benefits. Does this mean that the Duty View could justify the use of criminals by the state to provide free labour? Criminals could either work in factories run by the government, or their manpower could be distributed in a fair way to privately-owned factories.
Also, can the Duty View rule out the possibility of having at least some criminals (i.e. those responsible for having posed a threat to the life of their victims) doing jobs that pose significant risks to their own life, such as handling dangerous chemicals or working in laboratories where there is a risk of being exposed to radiations etc.
In the book you provide a number of arguments as to why criminals cannot be used for organ transplants (pp. 308-9), but as far as I can see none of those arguments can be invoked to answer this objection.
I can think of two replies to what you say in relation to 2):
a) You argue that since all citizens together share the costs of apprehending and convicting offenders, they should all receive some of the benefits of punishment. Two comments: to begin with, it is not clear that allowing William to discharge his duty to David by being his security guard would not also benefit David’s fellow citizens, although to a lesser extent. For other potential wrongdoers presumably will be dissuaded from committing assaults by the fact that they know that, if caught, they would incur the same sort of duty. In other words, allowing William to discharge his duty to David by being his security guard would still have some deterrent effect, which will benefit David’s fellow citizens.
Secondly, it could be argued that what we all receive in exchange for our support of the criminal justice system through taxation is: i) the benefit of crime prevention and ii) the benefit of having criminals apprehended. But once I am victim of a crime and the criminal is apprehended, I am the main beneficiary of the duties incurred by him, and whatever compensation I am owed should not be distributed. The model I have in mind here is the one used in tort law. We all contribute to maintaining the tort law system because we all benefit from having courts, judges etc., but once the judge decides that you must compensate me for destroying my precious garden dwarves, the compensation you owe to me should not be distributed among all those who contribute to maintaining the tort law system.
b) You write that “Those whose victims are apprehended will have greater security resources available than others, and they may be expected, for reasons of distributive justice, to distribute those resources rather than being entitled to all of them.” But the reason why David would have greater security resources than others is because he has previously suffered an unjust attack from William (an attack that others have not suffered). This might explain why reasons of distributive justice do not require that these resources be redistributed. Perhaps here you might be tempted to reply that the reason why some redistribution is nonetheless in order is that given that (as you argue in the book) William will have to provide more than strict compensation, David will be better off once he is compensated than he was before he was attacked by William. We might say that David effectively benefits from the injustice that he is a victim of. But it is not clear to me that he should not be allowed to benefit from this injustice. Consider the following case.
The football player Garrincha, considered by many as the best dribbler of all times, is sometimes said to have owed his fabulous dribbling to the fact that his legs were bent (he was called the “angel with bent legs”.) Imagine that you assault me and break my legs in several parts. Despite the doctors’ best efforts, my legs end up being bent, just like Garrincha’s. As a consequence, once I heal I become a famous football player. Thus, as a consequence of your assault I can be said to be better off than I was before I was attacked by you. However, in this case it is not clear that I’m not entitled to the benefits produced by my misfortune. At the very least, whether I am seems contingent on how well off those around me are. This final observation might seem obvious, but it’s important for my original example. Remember that David lives in a dodgy neighbourhood where the police dare not enter. It’s not clear in this case that distributive justice requires redistributing David’s newly acquired security resources to all his fellow citizens. If we plausibly assume that it’s unjust that David does not get the same level of protection received by citizens living in wealthier areas, redistribution seems unjustified.
Hello Massimo,
These are both good ‘pushes’.
On 1), of course, prisoners have regularly been used to provide free hard labour. The practice was abolished in England in the mid 1950s. There are a number of reasons to support abolition, which are somewhat related to the torture argument. First, I suspect that it was almost completely useless at producing anything. Even when it was used, in more industrial societies where you might have thought that it could be useful, it was deemed so useless that prisoners were often encouraged to perform pointless tasks.
Second, there is a great danger in a society which does allow very useful forced labour from its prisoners. Such a society has a strong incentive to keep its prison population high. We should worry about very cheap prisons in the same way that we should worry about the use of drone strikes in war – once we start to lower the costs of punishment and war, we increase the rate of unjust punishment and war. There is already a tendency to imprison many people who I doubt very much are liable to be punished. Were it not costly, I am certain that this tendency would be exacerbated.
Third, where the benefits of harsh punishment are not much greater than the benefits of rehabilitative communicative punishment, we owe it to offenders to use the latter form of punishment. We have strong reasons to put offenders in a position that they can redeem their lives in other ways. Forced labour is only likely to cause resentment.
Fourth, doing this is not just owed to offenders. If we wish to build a stable society with low rates of offending, it is imperative that we treat all people, even those who are liable to suffer great losses, as having great importance. Treating some people as though their lives are expendable should be done only in the most extreme circumstances, as in the hit man cases.
In principle, though, I think that it is permissible to force people to labour to provide non-protective benefits to victims and others. Consider:
Cave: D, V and V’s child are stuck down a cave together. D rapes V. V’s child suffers from a heart condition. She has managed to get hold of a gun, and is now pointing it at D. There is a mineral in the roof of the cave that will cure the heart condition. D can get the mineral, but doing so will be risky. Doing this will remove the serious risk that V’s child will die of her heart condition before she is rescued from the cave.
I think that it may be permissible for V to compel D to climb up and get the mineral. Do you think that this is not permitted?
On 2), first, in tort cases, the costs of apprehending and enforcing compensatory duties are much less, and court costs are typically paid by the parties. The reason why we do not distribute compensatory benefits is that they only render the victim as well off as he would have been had he not been injured. But even here, there are some distributive pressures. We restrict the availability of the tort system for distributive reasons, for example (John Gardner has been writing on this), and there is some pressure to distribute compensatory benefits to victims who don’t have perpetrators to claim from.
Secondly, I agree that where a person is in an unjust society, they may have a special claim to protection. They may not have a special claim, though, when compared with others who are also unjustly insecure. There would be nothing wrong, and everything right, in counterbalancing the lack of security that an unjustly insecure victim experiences through punishment. In such a case, the rich would have no complaint that they are not receiving the benefits of punishment.
The question, in your case, then, is how to distribute the benefits of security amongst those who are unjustly insecure. I think that there are conflicting pressures here. On the one hand, the victim is the person whose rights have been violated, so there is some pressure for the ‘extra-compensatory’ benefit of punishment to go to him. But there are distributive justice reasons to share those benefits where this can be done. The victim is entitled to be a beneficiary of the injustice that has been done to him, of course. The question is whether he is especially entitled to those benefits and to what degree. I don’t have a complete answer to this, but I think that your excellent Garrincha case doesn’t demonstrate that the victim is entitled to the benefits of his comparative ‘good luck’ in being offended against. I think that other principles of distributive justice will apply in this context as well. Your account has a somewhat libertarian flavour that I would resist.