LEAP 3 (2015) Response to Blumenson V ic tor ta dros University of Warwick abstract This paper defends some of the central claims in The Ends of Harm: The Moral Foundations of Criminal Law against challenges by Eric Blumenson. Key Words: Punishment, duties, retributivism, desert, proportionality INTRODUCTION I am ver y grateful to Eric Blumenson for his thoughtful, insightful and wide-ranging essay engaging with my book The Ends of Harm: The Moral Foundations of Criminal Law (hereafter, Ends). Blumenson makes many points which he thinks count against the Duty View of punishment (DV) that I defend in Ends which need careful thought. I won’t attempt to address all of his concerns. I will focus on those objections which I think he finds more telling, and where I have said less to defend the view outlined in Ends.1 Before continuing, let me clarif y one feature of the general moral and polit ica l v iew out lined in Ends t hat Blumenson brief ly engages w it h. Blumenson suggests that the account of the means principle in ends is so strict and wide ranging that it implies a form of libertarianism akin to that defended by Robert Nozick in Anarchy, State, and Utopia (Nozick 1974). I would like to clarify the similarities and differences between my approach and Nozick’s. I have some sympathy with Nozick’s general methodological approach to political theory. Like Nozick, I think that the duties and constraints on state action are to be defended in the light of a more basic set of moral principles that govern individual interactions with each other. The state, and the principles that govern it are not, I think, sui generis. Nozick’s account is more d i rect ly rel ia nt on a set of r ig ht s t ha n t hat defended i n Ends. Nevertheless, otherwise there is a broad similarity between us with respect to methodology. 1 I have responded to some points that are similar to those advanced by Blumenson in Tadros 2012, 2013 and 2015. Response to Blumenson 57 LEAP 3 (2015) Furthermore, like Nozick, I think that there is a constraint on compelling some people to provide for others – the constraint on using some as a means to the good of others. Where Nozick and I differ, though, is in our understanding of the content of these rights and duties. Nozick was much more skeptical than I am about enforceable positive duties, and this feature of his moral theory is central to his defence of libertarianism.2 In contrast, I believe that there are enforceable positive duties. The scope and stringency of these duties may depend on a number of factors, including the extent to which people are mutually engaged in cooperative beneficial activities. I do not provide a full account of the scope and source of positive duties in Ends – I leave that as an open question, even though it has implications for criminal justice. Nevertheless, my account of the means principle emphasizes the limits of that principle in a way that Nozick’s account does not. This prov ides t he possibilit y of broad ranging socia l prov ision and development aid, depending on the scope of domestic and international mora l obligations. My ow n v iew is t hat positive enforceable duties are quite extensive, and hence I believe that we have powerful obligations of these kinds that can be enforced by the state.3 This also helps to address one of Blumenson’s concerns – he thinks that relying on the ‘duty of easy rescue’ to secure the permission of the state to punish offenders is inadequate because indiv iduals do not have a ver y powerful dut y to transfer their compensatory rights to the state to ensure that other people are protected from offending. I doubt that this is right. This is so in part because victims of crime are provided with substantial benefits from a system of punishment. Requiring them to allow the state to have control over their punitive rights is not typically onerous (on the contrary) and any cost that they bear as a result is more than compensated for by the benefits they are provided with. Blumenson also complains that funding health care and occupational safety would, on this view, depend on the existence of duties to provide these things through the state. But it is highly plausible that the provision of these things does depend on such duties – when we fail to provide adequate health care or occupational safety, or for that matter adequate security from crime, we fail to do what justice requires, and hence we fail to satisfy our duties to others. When we provide health care beyond these limits, the appropriate complaint does seem to be that citizens do not owe this level of health care to each other. If the state fails to demonstrate that we owe some level of health care to each other, it has failed to justify that level of health care. 2 Nozick 1974, especially 30-3. 3 See, also, Tadros 2011b. 58 Victor Tadros LEAP 3 (2015) I think, therefore, that the provision of social justice depends on citizens underlying duties to each other. This is not a libertarian view, at least if libertarians are committed (as Nozick was) to the non-existence of general enforceable positive duties. 1. METHODOLOGY Blumenson thinks that the fact that there are many duties, with uncertain scope, involved in DV provides some reason to reject that view. This is not a reason to reject DV. If these duties ex ist, we ought to tr y to fulfill them. It may be that we are likely to fulfill them only approximately, but that fact does not free us from these duties. In any individual life, many of us are confronted with a wide range of duties, some of which are in conf lict with others. We have many complex duties as parents, children, professionals, friends, and simply as human beings, and it is often difficult to know whether we have fulfilled these duties. We are likely to do so only approximately at best. This fact hardly frees us of these duties, or makes them irrelevant to the way we live our lives. Blumenson also thinks that the fact that DV has many steps should incline us against it. The success of the argument for each step, he suggests, is necessary for the success of DV. Even if we have a great deal of confidence in each step in the argument, our confidence in the conclusion should be limited if there are many steps in the argument. Unfortunately, Blumenson’s argument would rule out being persuaded by any philosophical argument that relies on many steps, even where we have very powerful reasons to endorse every step in the argument. It is difficult to accept this conclusion. Furthermore, any particular moral claim relies on the truth of many other controversial claims – for example claims about moral claims. Blumenson’s argument would thus lead to more general skepticism about all particular moral claims. Finally, Blumenson’s argument makes philosophical argumentation too easy. Consider the truth or falsity of act utilitarianism. Act uilitarianism is false if any non-act-utilitarian claim is true. There are many non-act-utilitarian claims. Even if we have very low levels of confidence in any particular claim, Blumenson’s view implies that we ought not to endorse act utilitarianism in virtue of the fact that there is some relatively high probability that one of these claims is true. I take it that this argument against act utilitarianism is not valid – to show that we ought not to accept act utilitarianism, we ought to demonstrate that we have confidence in some particular claim that is in conf lict with act utilitarianism. Response to Blumenson 59 LEAP 3 (2015) It is also worth noting that some of the steps in the argument provided in Ends are not necessary to justif y punishment generally – they justif y punishment of particular people for particular purposes. For example, I defend the v iew that it is sometimes permissible to harm a person as a means to avert threats to people other than their immediate victims, even when they are not responsible for posing these threats. If the argument for this claim fails, punishment may still be justified in many cases. It will be justified in cases where offenders, through their offending, create threats to others. This will sometimes be so, because by offending they may undermine the effectiveness of the criminal law. It will also be justified in cases where punishing the offender helps to protect the victim from future offending. I don’t find these limits on punishment plausible, and I argued for a more expansive view. But the question under consideration is not about whether we should accept DV, but rather about what version of DV we should accept. 2. BLUMENSON’S ALTERNATIVE As punishment is difficult to justif y, we should not be surprised that an argument for punishment has many steps to it. Take Blumenson’s justification of punishment. One idea amongst many controversial ideas that Blumenson relies on, in defending his negative retributivism, is that punishment of the innocent is ruled out by something like the following principle: Negative Desert: It is permissible to use D as a means to the good iff D deserves to be used, and in virtue of this fact. Negative Desert is not a very clear principle. We need to know what desert means, and there are many possibilities. We also need to know what it means to deserve something ‘negatively’. I have some doubts that we can make good sense of negative desert in this sense. Like many friends of desert, I think that ‘X deserves O’ implies that there is good reason for X to get O irrespective of any further good that will be secured if X gets O. If so, there is no such thing as negative desert in Blumenson’s sense. Negative Desert is not a coherent view, I claim, because it is a fact about desert, as I understand it, that if D deserves O, the value of giving O to D does not depend on any further value that giving O to D will secure instrumentally. Negative Desert is in conf lict w ith this claim about desert. It makes the reason to give D O depend on something external to D getting O – that O is a means to some further good. Hence, it is not a v iew about desert, as I understand it. Perhaps Blumenson has a different conception of desert in mind in Negative Desert. Without an account of that conception, I continue to find Negative Desert confused. 60 Victor Tadros LEAP 3 (2015) My suspicion that Blumenson’s view is confused was strengthened by his response to the incoherence objection. There Blumenson swithers between two different views. On the one hand, he seems to accept that desert implies that there is a positive reason to give D O, but a reason that can be outweighed. This view accepts that it is intrinsically valuable to give D what he deserves, but that this value can be out weighed. But this v iew is in conf lict w ith Blumenson’s denial that retributivists are committed to the view that it is intrinsically valuable that offenders suffer the burdens of punishment. All sensible retributivists claim that the good of offenders getting what they deserve can be outweighed. So this view is just accepts positive desert, with all of its difficulties. Another view is the ‘rights forfeiture’ view. D, this view claims, loses his right against being punished in virtue of having acted wrongly. This view is not best seen as a form of retributivism at all. One reason is that rights forfeiture does not depend on desert.4 A second reason is that it is not clear how desert adds anything to a rights-forfeiture view. Rights-forfeiture views of punishment claim that punishment may permissibly be inflicted to serve certain ends on condition that a person has forfeited her right against such treatment. They then argue that a person forfeits her right against such treatment by acting wrongly. It would add nothing to such theories also to claim that wrongdoers deserve to lose these rights. The language of desert adds nothing substantial to rights-forfeiture theory – it is not even clear what it means to deserve to lose a right. DV is a version of a forfeiture view. So if, by negative desert, Blumenson simply means ‘rights forfeiture’, we are roughly in the same camp. But traditional rights-forfeiture views of punishment are superficial – the idea that wrongdoers lose rights is, of course, true on any justification of punishment. What needs explaining is why offenders lose their rights against being harmed as a means to the ends of punishment. Claiming that they have lost their rights to be used in this way is a conclusion that needs defending. The duty view offers a defence of this claim: it suggests that wrongdoers lose their rights against being used for the ends of punishemnt because they incur duties to serve these ends. I also claim that nothing, or almost nothing, else is sufficient – a person loses her right not to be used to serve an end, I claim, only if she has a duty to serve that end, or would have such a duty were she able to pursue it. Blumenson does not show that this view is false. If Blumenson wishes to defend a rights forfeiture view of punishment, he ought not to refer to desert. The idea that D has forfeited her right against being treated in a certain way does not depend on desert. If he rejects the 4 See, for example, McMahan 2009: 8-9. Response to Blumenson 61 LEAP 3 (2015) kind of rights forfeiture view defended in Ends, he should supply an alternative explanation how wrongdoers lose their rights against being punished. Furthermore, any convincing and complete account of rights-forfeiture will be controversial. A rights-forfeiture theorist needs both to provide an account of why people lose rights, and an account of the extent to which they lose rights. One reason why the argument in Ends is complex is that it attempts to make substantial progress with both questions. Simple rights-forfeiture views seem comparatively simple, but that is only because they avoid making progress with the difficult questions about why people lose rights, and the extent of the rights they lose. Let’s suppose that there is such a thing as negative desert. Let's suppose that there is such a thing as negative desert. Like DV Negative Desert relies on very controversial premises if it is to play the role in the justification of punishment that Blumenson wants it to play. Defending some particular account of desert against others will be controversial. This is obviously so as there are many different claims that f riends of deser t ma ke about t he nature of deser t.5 Fur t hermore, a ny interpretation of Negative Desert is extremely controversial. Even if there is such a thing as negative desert, it is not uncontroversial that we can deserve to be used. It is not uncontroversial that anything like criminal wrongdoing is the desert basis for being used. And it is not uncontroversial – in fact it is clearly false - that it is only permissible to use a person as a means to the good only in virtue of the fact that the person deserves to be used. That this is clearly false is demonstrated by the fact that we can mount relatively uncontroversial arguments for the permissible using of others without relying on desert. Negative Desert is by no mea ns t he most cont roversia l element of Blumenson’s view, though. It relies on another claim: that the state has an obligation to punish wrongdoers even when this will produce no net social benefit ‘when this is obligatory in order to fulfill the state’s social contract obligations to the victim because no less draconian route is sufficient to do so’. This element of Blumenson’s view also relies on many controversial claims, and so is vulnerable to Blumenson’s own argument. It relies on the success of following argument: The state has social contract obligations to victims. These social contract obligation that the state owes to the victim require the state not to ignore wrongdoing. 5 Even within the camp of the intrinsic goodness desert view, there is an enormous range of possibilities, many of which have gone unnoticed. For an exhausting, but perhaps not exhaustive, exploration of many of them, see Kagan 2012. 62 Victor Tadros LEAP 3 (2015) If serious wrongdoing is not punished, the state ignores wrongdoing. Therefore the state must punish serious wrongdoing. This argument is even more controversial than Negative Desert. First, the socia l cont ract t radit ion is ex t remely cont roversia l. I doubt t hat state obligations are grounded in social contract obligations. More importantly, it is not clear why the state, in failing to punish, necessarily ignores wrongdoing. There are many different non-punitive responses that the state might make to wrongdoing that would demonstrate that the state takes wrongdoing seriously. Given that punishing offenders is necessarily burdensome to the offender, but not necessarily burdensome to the state, it is also not clear why punishing offenders is sufficient to demonstrate that the state takes wrongdoing seriously. If I want to show that I take something seriously, it is I that should demonstrate a willingness to bear burdens for the sake of that thing. In punishing offenders, the state only demonstrates that it is willing to burden someone else – the offender. This may simply show that the state does not care much about the offender, not that it takes the wrongdoing seriously. Furthermore, it is not clear that by refraining from punishing a person who has committed a serious wrong against the victim the state withdraws the status of citizens as valued members of the political community, as Blumenson cla ims. To demonst rate its commit ment to t he v ict im, t he sensible thing to do is to help the victim. Punishing the offender may be one way to help the victim. The argument in Ends is intended to demonstrate that. But it is not the only way. The victim could be helped by being provided with extra protection by the state, or by being compensated in other ways. And the state could publicly express the importance of the v ictim, and educate its citizens about how terrible it was that the victim suffered in the way that she did. None of this requires punishing the offender. If the state does these things, is it really true that the state devalues the v ictim or denies his civic personhood? So spelling out and defending Negative Retributivism requires an argument with many steps in it. As we have reason to doubt each of these steps (in my view, decisive reasons to reject some steps), Blumenson’s ‘diminishing returns’ argument, if successful, defeats his own view as well as mine. Blumenson responds t hat a n a rg ument w it h mu lt iple steps is more credible if its conclusion is independently intuitively attractive. He thinks that negative retributiv ism is intuitively more attractive than the dut y view. I don’t find retributivism intuitively attractive. Blumenson’s version seems unclear. Negative retributivists can simply assert that those who we Response to Blumenson 63 LEAP 3 (2015) intuitively think ought to be punished lack a right against being punished. But then the theory does little more than reporting our considered convictions about punishment rather than explaining them. More standard retributivist views are also unclear, and seem barbaric. Few claims need more careful defence than the claim that it is impersonally valuable that wrongdoers suffer. Though some people believe this claim, it is not by itself intuitive – it rests on the idea that suffering and harm are sometimes to be sought for their inherent properties, and many recoil at this thought. Furthermore, I don’t think that the explanation that DV gives for the permissibilit y of punishment is intuitively unattractive. First, the vast majority of people, when asked why we punish offenders, cite prevention as the aim. DV offers an argument why it is permissible, subject to certain constraints, to harm offenders to secure this aim. But the aim that DV advocates is familiar and attractive. Secondly, the idea that offenders may permissibly be used to secure this aim in v irtue of their w rongdoing is attractive. DV then gives an answer to the familiar objection that offenders may not be used merely as a means to secure this aim – that offenders can be expected to serve the end of protection in virtue of having acted wrongly. This is the most original part of DV, but I don’t think that the response to the objection is itself unintuitive. 3. OTHER WAYS OF FULFILLING THE DUTY DV relies on t he idea t hat t he permissibi lit y of punishing of fenders is grounded in the (primarily protective) duties that offenders incur as a result of their wrongdoing. Blumenson, in challenging Step 6 of his summary, argues that it is a weakness of DV that this would allow uneven punishment of offenders in cases where offenders have discharged the protective duties that they owe to their victims in other ways. I think that this implication of DV is one of its strengths. 3.1. Who Gets to Decide One question that Blumenson raises, that has also been raised by Kim Ferzan, is why the state should be in a position to decide how the duty that the offender owes to the victim is satisfied. Ferzan raised the following objection to DV. She argued that if the offender owes a duty to the victim, it is for the victim to determine what the offender does for her. If she wants protection, she can secure protection from him. If she wants her car washed, she can secure that end (Ferzan 2013). In response, I 64 Victor Tadros LEAP 3 (2015) argued that it would be w rong for the v ictim to impose any significant harm on the offender for the sake of her car. The offender can be harmed for the victim’s sake only if any harm imposed on him is proportionate to the end sought (Tadros 2013). Blumenson does not think that this obviates the problem that Ferzan raises. I do not see why. The first thing to note is that offenders obviously do owe very stringent duties to the victims of their wrongdoing. Blumenson’s challenge of explaining how these duties can best be fulfilled is thus quite general. If punishment does not vitiate these duties, offenders retain them. If Blumenson thinks that offenders retain these duties, how does he think they ought to be fulfilled? Does he think, for example, that after the offender is punished, the offender may nevertheless be seriously harmed again for the sake of the victim? Nevertheless, we should meet the challenge posed by Blumenson. Why does Blumenson doubt that DV can meet it? It is not completely clear, but perhaps Blumenson believes something like this: if D owes a duty to act for the sake of V at some cost n, V may impose n on D for any end whatsoever. This v iew, though, is not credible. It is not generally true of duties. For example, the fact that I have promised to deliver a television to your house implies that I must bear the cost of driv ing to your house. That, though, does not imply that you may impose on me the cost of driv ing to your house for some other end. The most important restriction on the duty imposed on offenders is that any cost that the offender is compelled to bear must be proportionate to the end sought. For example, if D kills one of V’s children, it is plausible that V can kill D as a means to protect another of V’s children if that is the only way to protect the second child. This is so in virtue of the stringent protective duty that D incurs to V, and to V’s child. This does not imply that it would be permissible for V to kill D for fun, or to use D’s skin to make a handbag, for the harm imposed on D would be disproportionate to the good of having fun or getting a handbag. Hence, if the victim has a right to decide how the duty that the offender owes to her is satisfied, she has a right to decide only within a certain range of goods that are sufficiently important to justify the harm imposed on the offender. Now, there might be some circumstances in which the victim does have a right to decide, and the state would act wrongly in making the decision for her. For example, suppose that both V and X, who is V’s husband, are now threatened with death. D can be used as a means to protect either V or X but not both. It is plausible that V is permitted to determine whether to use D to protect V or to protect X. It follows that it would be wrong for the state to use D to protect V if V would prefer that X be protected. Needless to say, Response to Blumenson 65 LEAP 3 (2015) though, the fact that it would be wrong, in these circumstances, for the state to decide for the victim does not militate powerfully against DV, for these circumstances very rarely arise. It is also worth noting that the state does sometimes give the v ictim cont rol over how t he of fender’s dut ies a re sat isf ied. For exa mple, it is common in restorat ive just ice prog ra ms to prov ide t he v ict im w it h a n opportunity to decide what the offender should do in response to having offended. Within some limited range of options, this will sometimes be appropriate. In determining whether the state or the victim ought to decide how D’s duty is satisfied, we must also bear in mind a number of other factors that count against victims making decisions. First, if victims are entitled to make decisions about how the offender’s duty is satisfied, they will be at risk of coercion from offenders.6 Secondly, victims, who will often feel resentment towards offenders, may be inclined to seek retribution by imposing heavier costs on offenders than necessary. Thirdly, the victim owes duties to other citizens, including protective duties, and she may be required to select a particular way of satisfying the duty that the offender owes to her because this will best satisfy these protective duties.7 Fourthly, if the state punishes the offender, certain other values, such as communicative values, can be advanced, and this provides some reason for the state uniformly to determine how the offender is punished. If the victim does not have a strong reason to prefer that the duty that is owed to her is satisfied in one way rather than another, it is permissible to ensure that the duty owed to her is satisfied in a particular way to advance other values. Overall, the idea that the victim should have complete choice over how the offender’s duty is morally abhorrent, unrealistic and impractical. Hence, the fact that victims typically have a right to determine how the duties owed to them are satisfied is not a significant challenge to DV. 3.2 Duties and the Wealthy The second part of Blumenson’s challenge concerns wealthy offenders. There are three features of this challenge that should be separated. One question is whether it is plausible that w rongdoers who have taken on burdens for the sake of their victims ought nevertheless to be punished. DV implies that if these burdens are sufficiently large they ought not to be. A second question concerns the relationship between punishment and compensation. 6 See, also, the discussion concerning giv ing the v ictim control over prosecution decisions in Tadros 2011a: 296. 7 See Tadros 2011a: 297-9. 66 Victor Tadros LEAP 3 (2015) Does DV deny the victim the right to seek monetary compensation from her offender? A third question is whether DV has plausible implications for the punishment of the wealthy, who may have the means to provide a great deal of protection to others at little cost to themselves. To assess the first question, suppose that V is attacked by a gang, including D, X, Y and Z. D assaults V. Let us suppose that it would have been permissible to harm D as a means to avert this threat, harming him to degree x. If the harm that D does to V is large, x will be even larger. This is for the reason that it is generally permissible to harm a person to avert a threat that they culpably pose, even if the harm that is necessary to avert the threat is greater than the harm that the person would do if the threat they pose were realized. D then realizes that what he did was wrong. V still faces a threat from X, Y and Z. D, recognizing the duty that he owes to V, now protects V against X, Y and Z. He is harmed to degree x in the process. Blumenson implies that the state nevertheless ought to punish D. This seems wrong. D has borne a great cost in order to protect V against X, Y and Z. It is wrong to harm D even more for punitive reasons. The idea that we should punish people who have already voluntarily borne great costs for the sake of their victims fails adequately to acknowledge the idea that people can redeem themselves for their wrongdoing through their voluntary actions. In response to Blumenson’s second challenge, there is nothing in DV that rules out the provision of monetary compensation to victims where monetary compensation is available, and can help to ameliorate the victim’s loss. In evaluating the comparative merits of a compensation scheme against a punitive scheme, we ought not simply to consider what would be beneficial to indiv idual v ictims where most offenders are punished. We ought to consider the circumstances of victims were no one punished. In a system where compensation was the only available remedy for theft, for example, people would be very insecure in their property. In those circumstances, a compensatory scheme would be wholly inadequate to secure property rights. Hence, there would be powerful reasons to criminalize theft. Nevertheless, even if theft is criminalized, victims may seek monetary compensation from offenders. To explain why this is so, notice the following feature of DV. Obligations to compensate others are subject to two constraints – a constraint on the maximum amount of harm that can be imposed on the wrongdoer (what I have called the maximum harm threshold) and a constraint on the maximum amount of benefit that can be secured from the wrongdoer (the maximum benefit threshold). DV accepts that there is a maximum harm threshold beyond which the offender may not be harmed, though not a maximum benefit threshold. Below the former threshold, we must consider how the offender Response to Blumenson 67 LEAP 3 (2015) is to be ha rmed, a nd for whose benef it. Ha rm may be imposed on t he offender to provide monetary compensation to the victim, to deter, or a combination of the two. This has the attraction that it limits the overall harm that an offender suffers as a result of his wrongdoing. If the offender provides compensation to the victim, and this is very burdensome to him, the amount that he may be punished is reduced. Blumenson’s claim that, according to DV, offenders who are punished owe no compensation to their victim is thus false. That depends on how much the offender is punished. It is true that if an offender is punished up to the maximum harm threshold, he may not be harmed further to compensate the victim. But this is an attractive implication of DV. If, though, the offender is punished to some degree less than the maximum harm threshold, compensation may be extracted from him for the victim’s sake. Given that the victim will typically have been rendered much worse off than she would have been by the offenders action, there is good reason to ensure that we should aim to improve the victim’s circumstances. The question is how best to do this. The third feature of Blumenson’s challenge concerns the way in which we should respond to the greater protective resources available to the wealthy. As I have said a great deal about this elsewhere, I will make my comments brief. In compensatory justice, the maximum benefit threshold is normally set at the level of full compensation. It is contested how full compensation should be understood,8 but in standard cases, if the wrongdoer has rendered the victim as well off as she would have been had the wrongdoer not wrongfully harmed her, he has fully compensated her. The maximum harm threshold implies that there is a limit to how much the w rongdoer may be harmed in order to prov ide the v ictim w ith full compensation. If full compensation to the victim would require us to harm the wrongdoer a great deal, it is impermissible to extract full compensation from the wrongdoer. With respect to serious wrongdoing, whilst there is a maximum harm threshold, there is no maximum benefit threshold. Wrongdoers cannot satisfy the duties incurred through wrongdoing simply by providing full compensation to victims. If the provision of full compensation is not onerous, they can be expected to do more to protect the victim, and perhaps to protect others. Given this, the fact that the wealthy can provide a great deal of protection to victims at little cost does not imply that they may not be punished to produce further protection if they provide this level of protection.9 8 I discuss this further in Tadros 2014a. 9 For further defence of this view, demonstrating how it comes apart from retributivism, see, Tadros 2011a: 286-91; 2012: 99-102; 2013: 300-9. 68 Victor Tadros LEAP 3 (2015) Blumenson, I think, agrees w ith this verdict, but thinks that this is tantamount to a form of negative retributivism. This is false. First, the scope and stringency of the duties that we owe in virtue of wrongdoing are not, in general, best explained by desert. When we act wrongly, we incur duties. We do not incur them because we deserve to incur them. Of course, we should explain why we have these duties – for example, because we can redeem ourselves by responding appropriately to our wrongdoing, or because we could have avoided having them, or some other explanation. It is difficult to see how desert claims figure in an explanation of the duties that we incur. Of course, we could define retributivism in a way that corresponds to the duties that we incur through wrongdoing. We could simply claim that when a person incurs a stringent duty in virtue of having wrongdoing, that just is her getting what she deserves. But if the retributivist claims this, she just accepts DV, but mangles the terminology. Now, Blumenson looks for a n ex pla nat ion why serious w rongdoers cannot satisfy their duties simply by providing full compensation to their victims. He complains that my defence of this is vague. I agree that there is more to say about this issue than I said in Ends. There I noted the fact that it is implausible that the duties of serious wrongdoers are satisfied when full compensation is provided, and gave an explanation for this by considering complaints that wrongdoers and victims could make to the scope of these duties. I also showed that the view that I endorse is intuitive in protection cases such as Three Threats. I have since done more to explain this idea, and I will say no more about it here. But whatever the merits of my explanation, Blumenson’s reference to desert does nothing to help. It is simply a bad redescription of the idea that wrongdoers incur stringent duties that are not satisfied by providing victims with full compensation. It does nothing at all to explain the source of these duties. The explanation that I provided in Ends may have been vague and incomplete. Blumenson offers no explanation at all. 4. EMPIRICAL SUPPORT DV justifies punishment only if punishment is effective in deterring crime. I am poorly placed to do the empirical work to that is necessary to determine whether it is effective, and in which circumstances. This is partly due to my lack of empirical skills. It is also due to the fact that the empirical work could not realistically be done. W hat would be required would be a study that compares reasonably just large-scale liberal societies w ith and w ithout Response to Blumenson 69 LEAP 3 (2015) systems of punishment. But there is no reasonably just large-scale liberal society without a system of punishment. The question is whether dismantling a system of punishment and replacing it w ith some intrinsically better a lternat ive wou ld be det rimenta l to t he crime rate. If it wou ld not be detrimental to the crime rate, our system of punishment is unjustified.10 The fact t hat DV justif ies punishment only contingent ly on it being effective in deterring crime and on it not having disproportionate bad side effects is no objection to it. Any sensible theory of punishment is contingent in this way. Any system of state punishment is enormously costly. It will harm not only offenders, but innocent people as well. These costs need to be justified. The idea that they can be justified by anything other than crime reduction is implausible. Even if it is true, as retributiv ists claim, t hat punishment is impersonally valuable, it is very difficult to believe that any impersonal value that it has is sufficiently great fully to justify the costs of any realistic criminal justice system. If state punishment is ineffective in reducing the crime rate, state punishment ought to be abandoned.11 5. PROPORTIONALITY Blumenson thinks that DV cannot explain why some factors that intuitively ought to affect the sentence that we ought to impose on an offender are relevant to punishment. I am not sure why he thinks this. Though it is true that I ought to have said more about this issue in Ends, it is highly plausible that both a person’s duties to avert the threats that he poses and the strength of the duties he incurs through his wrongdoing depend on the kinds of factors that Blumenson outlines as relevant to punishment. Furthermore, retributivism seems to me less well placed to explain these factors. Blumenson rightly claim that retributiv ists t y pically think that punishment shou ld ref lect t he g rav it y of t he crime a nd t he of fender’s blameworthiness for it. W hat they lack is an explanation why this is so. There is little reason to think that, on the best view of desert, what we deserve is determined by the properties of our actions. It is more plausible to think that what we deser ve depends on our v irtues and v ices.12 But this more plausible view of desert has troubling implications for a theory of punishment. Retributivists typically tailor their theory of desert to provide plausible 10 For further discussion, see Tadros 2012: 91-3. 11 For a compelling argument for this view, see Husak 2010. 12 See, further, Tadros 2011a, ch.4.II. It is not uncommon for friends of desert in moral philosophy to think that virtue and vice over a whole life is the proper desert basis. See, for example, Kagan 2012: 6-12. ch.4.II Response to Blumenson 70 LEAP 3 (2015) implications for punishment. But if so, their justification of punishment is unsatisfactory. It is no good to begin from a conventional view about what punishment ought to be imposed for which crimes under which circumstances and justify this conventional view simply on the basis that doing what we do is impersonally valuable. What is needed is an independent argument that shows that desert explains why punishment should fit the gravity of crimes and blameworthiness for them. Such arguments are hard to find, and Blumenson offers none. Blumenson is also wrong to think that the necessity constraint on self- defence creates problems for the relationship bet ween self-defence and punishment outlined in Ends. I w ill restrict myself to an evaluation of Blumenson’s first case, as I think this the most interesting: Unnecessary Defence. A threatens V with a knife. V is able to retreat safely and therefore defensive force is unnecessary; but V fails to retreat, parries A’s thrust unsuccessfully, and dies. As harming A was not necessary to avert the threat he poses, it might be argued that A is not liable to defensive harm. Blumenson then concludes that DV implies that he is also not liable to be punished. This argument is much too quick. First, it is not clear that A is not liable to defensive harm. It is a matter of dispute whether harming A wrongs him where harming him is unnecessary.13 I am inclined to the view that V wrongs A if V unnecessarily harms A. But this does not imply that A incurs no duties as a result of wrongfully killing V. The fact that V wrongly chose to attempt to harm A rather than retreating does not vitiate the duties that A incurs for wrongfully harming V. It is also false that A is not liable to defensive harm to avert the threat that he later poses in Unnecessary Defence.14 To see this, consider a variation on Unnecessar y Defence where X, a t hird pa r t y, cou ld inter vene a f ter V attempts to parry A’s thrust to avert the threat that A poses to V. If X harms A to avert the threat that he poses, X does not wrong A. Although V acts wrongly in attempting unnecessarily to harm A, this does not vitiate A’s liability for threat that he later poses.15 Furthermore, whilst it is true that, if V successfully harms A in Unnecessary Defence, V wrongs A, it is false that it is wrong for V to harm A given that V 13 For some discussion, see, for exa mple, McMa ha n 2009 ; Firt h a nd Quong 2012; Frowe 2014. 14 Blumenson also considers the problem of criminal attempts and the significance of intentions to punishment. I say more about these issues in Tadros 2013: 313-22, so I leave them aside here. 15 For related discussion, see further Tadros 2014b. 71 Victor Tadros LEAP 3 (2015) has decided to stand his ground rather than to retreat. Suppose that V stands his ground. Suppose, also, that by standing his ground he gives up the opportunity that he had to retreat. He must now decide whether to harm A, parrying the blow that A aims at him, or to allow himself to be harmed by A. He may recognize that he was wrong to stand his ground, but this does not vitiate his permission to harm A. He wrongs A in virtue of the fact that he had another option which rendered his defensive force unnecessary. But despite the fact that the existence of this option renders his defensive harm wrongful, given his wrongful act he is permitted to defend himself. It follows that there is no asymmetry between liability in self-defense and liability in punishment of the kind that Blumenson’s argument relies on. CONCLUSION Obviously, there is a great deal more that would need to be done fully to meet Blumenson’s interesting and important objections to the arguments in Ends than I have done here. I hope, at least, to have shown that some of Blumenson’s objections can either be met by clarifying DV, or by showing that the seemingly counterintuitive implications of DV are more attractive than he thinks. DV is a new theory of punishment. I do not claim to have worked out all of the details of the theory in Ends. I continue to think that it has a great deal of promise, despite the powerful objections that have been mounted against it. BIBLIOGRAPHY Ferzan, KK., 2013: "Rethinking The Ends of Harm", Law and Philosophy 32: 177-198. Firth, J. and Quong, J., 2012: “Necessity, Moral Liability and Defensive Harm”, Law and Philosophy 31: 673-701. Frowe, H., 2014: “Non-Combatant Liability in War”, in How We Fight: Ethics in War, ed. H. Frowe and G. Lang, Oxford: Oxford University Press. Husak, D. N., 2010: “Why Punish the Deserving?”, in The Philosophy of Criminal Law: Selected Essays, Oxford: Oxford University Press. Kagan, S., 2012: The Geometry of Desert, Oxford: Oxford University Press. McMahan, J., 2009: Killing in War, Oxford: Oxford University Press. Nozick, R., 1974: Anarchy, State and Utopia, Oxford: Blackwell. Tadros, V., 2011a: The Ends of Harm: The Moral Foundations of Criminal Law, Oxford: Oxford University Press. — (2011b) “Independence Without Interests?”, Oxford Journal of Legal Studies 31: 193-213 — (2012) “Replies”, Jerusalem Review of Legal Studies 5: 89-109. — (2013) “Responses” Law and Philosophy 32: 241-325. Response to Blumenson 72 LEAP 3 (2015) — (2014a) “What Might Have Been” in Philosophical Foundations of Tort Law, ed. J. Oberdiek, Oxford: Oxford University Press. — (2014b) “Resource Wars”, Law and Philosophy 33: 361-389. — (2015) “Answers”, Criminal Law and Philosophy 9: 73-102.