The Idea of a Justification For Punishment KEVIN MAGILL The argument between retlibutiuists afld conseqaeitialists about whal morully iustifies the punishneflt of offe ers is incoherent. If we were to discouer tbat all of tbe contetding jrctifications were mistakefl, tbere is to realistic prospect ,hat tbk wotld lead us to abandon legal punishment. lustification of words, beliefs and deeds, car only be intelligible ol the sssumption tbat if one's justification werc fornd to be inualid and there was no other justification, ote uould be prepaled to stop sayiflg, believing or doing ohat one has attetflpted to iustiff. Therefore tbe moral standing of our practices of punisbing offendets sbould not be sought in a iust;fication of ir. Because punishment involves the infliction of harm or deprivation on an offender, and because inflicting harm or deprivation on an individual is otherwise generally thought to be a bad thing to do, punishment appears to call for a moral justification.' Theories about what it is that justifies a sociery in punishing offenders typically contain two elements. First, there are various principles or criteria that must be satisfied by the practice if it is to be morally iustified; that it is an economical means of maximising securiry, well-being or satisfaction; that it should only be inflicted on an offender; that there should be some relationship of proportionality between penalty and offence; that an offence should have been knowingly and voluntarily undertaken; that penalties be mered out equitably and impartially, and so on. Second, rhere are arguments which aim to demonstrare that the stated principles are necessary and sufficient to justify punishment. Advocates of theories of punishment have given little attention to the intelligibility of the general enterprise in which they are engaged: the enterprise ol iustification. To think abour rhe enterprise of justification is to think about what criteria and presuppositions must be satisfied if a justification - any justification is to count as a justification. CRISPB Vol.1, No.1 (Sprin8 1998), pp.86-101 PUBLISHED BY FRANK CASS, LONDON JUSTIFICATION FOR PUNISHMENT 87 It is worth noting, ro begin with, that there is one kind of justification for punishing offenders that can properly counr as such. Punishment can be rationally iustified according to its beneficial social consequences. A rational justification of punishmenr will provide reasons for believing that punishing offenders is the most effective means of achieving some desired end, such as keeping law breaking below a tolerable level (either by deterring law breaking and denying offenders the opportunity to reoffend or by the reform or education of offenders), so that, assuming we want law breaking to be kept to a tolerable level, or that we want somerhing else that depends on it, punishing offenders is shown to be somerhing we have reason to do. A moral justification, by conrast, must give reasons for believing that punishing offenders is a good or ngDt rhing to do. Consequentialism converts the rational justificarion of punishment into a moral justification by adding something like rhe following: that keeping law breaking to a tolerable level maximises freedom and securitS which in turn maximises satisfaction or well-being, and maximising satisfaction or well-being is an over-riding good. The objections to strict consequentialist jusrifications of punishment are well-known. If punishment is justified according to its beneficial consequences in maximising well-being or satisfaction by deterring crime and mitigating social insecuriry scapegoating acts of victimisation againsr non-offenders can be justified in the same way. Moreover, since many kinds of crime are committed predominantly by people from economically and educationally deprived backgrounds, and since there is reason for judging that deprivation is one kind of cause for such crimes being committed, a justification of punishment in terms of beneficial social consequences would, it is thought, be a hideous justification of unequally adding legal penalties to an already unequal distribution of deprivation and benefit. Additionally, justifying punishment according to maximisation of well-being or satisfaction faces difficulties in articulating a widely shared sense that there should be some relationship between culpability and severity of sentence. In short, any stricrly consequentialist justification of punishment entails a jusrification of legal victimisarion of both offenders and nonoffenders, which conflicts with widely shared and deeply entrenched moral attitudes or sentiments. CRISPP 88 A further reason for scepticism about whether a justification of punishment given strictly in terms of beneficial social consequences can, simply by stipulating that maximisation of well-being or securiry is morally desirable, ever provide an adequate moral justification, concerns the relative social privileges of those who might offer or assent to such a justification. If those who are prepared to see punishment of offenders continue on the basis of a strict consequentialist justification are privileged relative to those on whom the burden of punishment mostly falls, and their privileged staus is dependent on the continuing punishment of offenders, it is difficult to see how this can be clearly distinguished from a prudential self-serving justification. Since the moral attitudes or sentiments with which srricr consequentialism conflicts are recognised as reributive in character, dissatisfaction with consequentialist justifications of punishment, and with various standard defences of them, have led most contemporary punishment theorists to opt either for some form of retributive justification of punishment or for some form of compromise between consequentialism and retributivism. The guiding ideas of retributive theories of punishment have been that punishment is justified when an offender has voluntarily broken a law,' and that those who voluntarily break the law deserue to be punished.' Advocates of pure retributive theories of punishment allow that punishment has beneficial consequences, but argue that such consequences are irrelevant to the issue of why we are morally justified in punishing offenders. To punish with the aim of deterring crime or reform of the criminal, some argue, would be to fail to respect the autonomy, digniry or rights of those who are punishedr to use them, as we do objects and animals, as mere means to social ends. The punishmenr of an offender is intrinsically right or good, it is argued, and not as a means to some further end.' Compromise theorists take the view that to punish offenders as an end in itsel( without thought to what public good can come of it, would be to engage in infliction of suffering for its own sake and gratuitously to add to whatever harm has already been done by the offender (and often enough to what has been done to the offender by the unequal distribution of deprivations and benefits). The compromise they aim at is between social benefir as rhe general iustifying aim of punishment and rerributive sentiments about who may rightly be punished and in whar way.' JUSTIFICATION FOR PUNISHMENT 89 The details of the various retributive and compromise theories are unimportant for the purposes of this paper: they are all alike in what distinguishes them from strict consequentialism - that they treat retribudve principles as ineradicable from a satisfactory justificarion of punishment - in which case they all share one crucial defect in respect of their status as moral justifications of punishment. If we were to discoyer that punishment cannot be iustified other than strictly in terms of its beneficial social consequences, and if it is already accepted that punishment cannot be given an acceptable stricr consequentialist jusrification, then we would have to accept that there cannot be a moral justification for punishment.' Punishment cannot be justified retributively (in whole or in part) because justification of any kind involves a presupposition rhat any retributive justificarion of punishment necessarily lacks. We can and do attempt to justify all kinds of things, including actions, practices, beliefs, claims and arguments. To justify something is to provide reasons for saying it, doing it, believing it, or whatever; and to provide reasons for saying, doing or believing something is to say that that is utlry one says, does or believes it. If I offer a justification of something I have done (or am doing, or propose to do), I necessarily imply that the reasons I offer in justificarion are my reasons /or having done it.' Often enough we have several over-riding reasons for doing something, and our justification for doing ir need not include all of those reasons. When we have several reasons for doing something and our justification of ir does not cover all of the reasons for doing it, if the justification is discovered to be faulry we might continue to do it for whateyer reasons the justificarion did not cover and still be able to justify doing it according to those other reasons. But if one's justification for doing something covers all of one's reasons for doing it, then if the justificadon were found to be invalid and could not be rectified or improved upon, one could have no furrher reason or justification for acting in that way, in consequence of which ohe should stop doing it.' If I offer a justificadon for something I do, and I insist that I am acring for no other reason than the reasons I give in jusrification, and it is pointed out to me that the justification is invalid (because ir is based on a falsehood, sa5 or it involves a fallacy, or that there are overriding reasons I had not considered against doing it), and I 90 CRISPP recognise that the justification is invalid bur carry on doing what I am doing, it will be concluded that my justification was dishonest or confused: thar I am not really acting for the reasons I offered in justification. In that case one should also conclude that my justification was not merely invalid in the sense rhat it was based on falsehood or fallacy, but that it could not even be counted as a genuine attempt at justification, since my real reasons for acting were clearly other than those offered in justification. In general, then, we may say that a justification (for doing somerhing, believing something, saying something, or wharever) can only count as such if the reasons it contains are really rhe reasons for doing, saying or believing the thing for which a justificarion is being offered. Ve have reason to think that a jusrification does not satisfy this requirement whenever we have reason ao suspect that a person would continue to say or do whateyer he or she is arrempting to justify even if the justification were shown to be invalid and where no other reasons or justification have been offered.' If someone would continue to do something in the absence of any acceptable jusrification for it, then any justification he or she does offer for it can not properly count as a justification. Let us call this the Principle of Relinquishment in Light of Failure or PRLF: A reason or reasons for doing (saying, believing, etc.) something can only provide a justification of it, if ir is the case that one is prepared not to do ir,'n or to stop doing it, should those reasons be found not to be reasons for doing it and rhere are no orher reasons for doing it." To engage in justification - any justification - presupposes PRLF. Any justification of punishmenr, therefore, if it is to be a justification, must satisfy PRLF. It is possible that the kind of rational justificarion of punishment mentioned earlier might satisfy PRLF. If ir were found that punishing people is nor the most economical means of minimising law breaking, then it is conceivable that we might abandon it in favour of whatever means are mosr economicalLikewise, if for some unforeseeable reason we genuinely ceased to care about well-being and security or that laws be obeyed, it is a safe bet that we would no longer go to rhe trouble of punishing law JUSTIFICATION FOR PUNISHMENT 9l breakers. Ar any rate, and as I shall argue, for all or most of us ir does matter that laws are obeyed and there is no alternadve we know of to punishment as a means of deterrence and enforcemenr.,, While a rational justificarion for punishment will give (most of) us reasons for assendng ro the punishment of offenders, it can not show that we are right to do so. For rhat, so it seems, we need a moral justificarion; and a moral justification for punishment, musr, in rhe absence of an acceptable srrict consequentialisr lustification, be wholly or partly retribudve. PRLF requires of any arrempr to morally jusrify punishing offenders, that if the reasons offered in justification were discovered not to provide sufficient moral reason for punishment, and there were no other moral justification for doing it, that we ought nor to punish offenders. Can this requiremeni be satisfied? If legal punishment is, as its rarional justification requires, the most economical means of minimising law-breaking and, thireby, of maximising well-being and security, and if, moreover. the alternatives ro legal punishmenr are so .uneconomical' as to be conremplated only with dread, then there is no reason to believe that we would ever contemplate giving it up on the Brounds that it lacks a moral iusrification. Much depends, then, on what the alternarives to legal punishment are and on whether they would be as dreadful as we vaguely imagine rhem to be. The principle recommendarions of legal punishment as the most economical means of maximising well-being and security are that it deters law-breaking and that it denies opportunities for ieoffending. fu yet we have no idea how any sociery mighr realisrically hope to live wirhout some system of dererrence; which is to say, we have no idea how we mighr effecr a sufficiently radical reorganisation of social relationships, values and individual desires as to make it possible to live without a sysrem of deterrence and enforcemenr. Historical evidence - enough of it within living memory - is thar the breakdown of legal aurhoriry wilt be quickly succeeded by a new constitutional apparatus of deterrence and enforcement, or by local warlordships, each with its own rough and ready apparatus. If we can imagine that we mighr bring the practice of punishing offenders to an end because it lacks a moral jusrification, we certainly cannor, as things stand, realistically imagine bringing deterrence to an end.,, Can we imagine a sysrem of deterrence and enforcement that 92 CRISPP would not also be a system of punishmenr? The familiar alternative to punishment as a means of securing obedience to law is some or other therapeutic model involving rrearment or re-education of offenders. Let us suppose thar, on discovering that punishment lacks a moral justification, we embark on a sysrem of therapeutic correction in which the rerributive vocabulary of punishmenr is abolished and offenders are understood rhereafter as suffering from impairment or maladjustment and brought to understand themselves as being helped and enabled by imprisonmenr and correction. Something like this has been the dream of various social reformers, therapists and utopian novelists. Prisoners mighr, in such a system, come to see their imprisonment and the various hardships it involves as a cure for rheir criminal sickness, enabling them, eventually, to rejoin the communiry as well-adjusted and law-abiding citizens. Imprisonment and suffering would be welcomed or accepted, it is hoped, as curative, in just the same way thar a patient accepts the pain of surgery.'' While a system such as this would represent a radical shift in artirudes towards penal servirude, it is misraken to think that it would no longer be a sysrem of punishment. In the firsr place, the therapeudc model it involves is drawn from rhe religious idea of redemption through purgarive suffering and while the secular version of the idea may be intelligible, it will only be so on rhe assumprion that the suffering to be undergone is understood by rhe penitent as deserved. While modern therapeutic models may have dispensed with rhe immortal Soul, or the better self, as an object of redemption or treatmenr, and while offenders may be taught ro see their crimes as products of inappropriate behavioural rraining rarher than innate sinfulness, to be reformed the offender must be given reasons for thinking that he has choices about how to behave, that those he has harmed ought not to have been harmed, that sociery has the right to protecr irself against him, and that his suffering has a moral and legal standing such that he can see it as contributing to his redemption. I can see no way in which this could be intelligibly accomplished without the retributive vocabulary of responsibility, deserr, punishment and atonemenr. Without the idea of responsibility, for example, the offender would have no way of seeing the harm he has done in a way that is at all differenr from the way he might view the JUSTIFICATION FOR PUNISHMENT 93 harm done by anyone else, and no way of seeing what he has done as something he might have chosen not to do. Mthout the idea of desert he has no reason to give any weight to the interests of others, and no way of seeing his punishment as somerhing that sociery has a right to inflict on him: as being anything orher rhan an imposition. And without the concepts of desert, punishment and atonement he has no way of understanding his suffering as something that enables him ro feel properly entitled to rejoin the communiry. The point of a therapeuric model for the treatment of offenders, to the extent rhat it has ever worked or can ever be made to work, is nor to replace punishment but to give punishment, properly so called, a positive outcome in getting the offender to behave responsibly towards others and to accept responsibiliry to orhers for whar he does. Further, if such a system were to continue to serve as a deterrent - as something a potential offender would regard as wonh avoiding - it could not be regarded by the population at large as anything other than a sysrem of punishmenr. It is possible that the architects o1 such a system might seek to remove all ralk of punisbrnent, penalty, desert and retributiott from the adminisrration and enforcement of the law, but if the system were one in which rhe rreatment would still have to be feared, this would be an exercise in ideological redefinition: naive at best, cynical at worst. Logical retributivism norwirhsranding,'r a system of punishment will never cease to be so through euphemistic relabelling. It is possible, of course, to imagine thar a therapeutic programme of correcrion and re-education of offenders might form part of a larger system of behaviourist utopian condirioning in which all but a few are so well-adjusred and content rhat they lack the impulse to disobey the law In such a sociery there would be no need of deterrence, and in that case its treatment of offenders would, presumably no longer counr as punishment. But as I have argued, such a state of affairs is not a real alternarive to our current prictice of punishing offenders, since we have no realistic idea of how it could be accomplished. In Skinner's Walden Two,," for example, beyond a vague reference to behavioural reinforcement, rhere is a deafening silence abour how the utopia is able to change its citizens sufficiently for it to prosper without any need to punish wrongdoers. The omission is all the more striking when one considers that the great 94 CRISPP social experiment is carried out in a small and relatively closed communiry in which the imaginative challenge of a model for life without punishment would be, so we should think, a small mamer compared to rhat of re-engineering an endre nation or a world.'' Moreover, even if we did know how we might go about realising a behaviouristically re-engineered sociery ir is hard to imagine what reasons we might have for choosing to do so, and even if we did, the idea that the primary reason for so choosing would be that punishmenr lacks a general justificarion is scarcely credible.,, As things srand, then, we have no realistic idea of how we might bring it about that there no longer be a system of dererrence. The only alternative to our current constitutional systems of legal deterrence, is the rough and ready justice of warlords and militia.,' If we are to consider the abandonment of legal punishment on grounds of its lacking a moral justification, warlordism could not, on those grounds, be considered an alternative. We have no realistic model of consritutional law enforcement involving dererrence that would not also be a system of punishment in all bur the names given to it by utopian administrators. The supposed therapeutic alternatiye to punishment is not an alternative at all, but rather one in which retriburion would be geared towards reformative ends. Therefore, for us, the only real alternative to the practice of punishing offenders for their crimes would be one - warlordism - thar would be vastly more unjust - disastrously so - than any real or imagined injustice in our current pracrice.']. If this is the only real alternative ro olrrent practice, we cannot coherently suppose that we might give up that pracrice if the reasons offered in (wholly or partly) retributive justifications were found not to provide sufficient moral reason for punishing offenders. It follows that no (wholly or partly) retributive moral justification of punishment can sarisfy PRLF. Now it might be argued that all that all rhat PRLF requires of a moral justificarion of punishment is that if rhe reasons offered in justification were to be found not to provide sufficienr moral reason for punishing offenders, and there were no other justification - rzoral or othenise - for doing it, thar we ought to cease punishing offenders. But if there is a non-moral jusrification for punishing people, then PRLF is satisfied. After all, why should the fact rhat we have a non-moral justificarion for punishing people rule our our JUSTIFICATION FOR PUNISHMENT 95 having a moral lustification for it as well? [t has already been conceded that one can have more than one reason for doing something, and there is no reason to suppose that having moral reasons for doing something is incompatible with having other reasons for doing it. This is true, provided that there are no moral rcasons against doing something which, in the absence of any moral justification for it, should lead us to stop doing it. But this is not true of punishmenr. As I mentioned ar the beginning, the reason a moral justification is sought for punishment is that punishment involves the infliction of suffering or deprivation and inflicting suffering or deprivation is ceteis paribus a bad rhing to do. Any partially or wholly retributivist justification of punishment therefore commits its advocates to the view rhat if there were no moral justificarion for punishing offenders, we ought not to do it, even if there were non-moral reasons for continuing to do it. The argument does, however, suggest that, because there can be non-moral justifications for doing things for which moral justification is sought,'' PRLF should accordingly be qualified as moral PRLF or MPRLF: A reason or reasons for doing (saying, believing, erc.) something can only provide a moral justification of it, if it is the case that one is prepared not to do it, or to stop doing it, should those reasons be found not to be reasons for doing it and there are no other moral reasons for doing it and in that case there are other moral reasons for not doing it. Retributivists and compromise theorisrs might reply at this point that what they have been attempting, in their various ways, is to provide an ideal lustifrcztion of punishment: one involving the supposition that ideally moral agents would abandon the practice if the justification were found to be invalid. I doubt that this has been anyone's intention, but if it were, all that it could succeed in justifying is an ideal system of punishment. If there is no realistic possibiliry that any existing or foreseeable society might abandon punishing offenders on the grounds that the practice is unjustified, no ideal justification of punishment can count as justifying that practice in any existing or foreseeable sociery. In that case any ideal iustification of 96 CRISPP punishment would be a frivolous undertaking. If the notion of an ideal justification of punishment cannot save the enterprise, defenders of rhe idea of a justification for punishment might argue insread that theories of the justification of punishment are jusr that: theories. A theory or an analysis, of rhe iustification of punishment can no\ qua theory or qua analysis, amempr to justify punishment, for otherwise there would be no distinction terween the theory or analysis and what it is a theory or analysis of. A theory or analysis of rhe justification of punishment will take it for granted that such a justification exists and attempr ro clarify what iiis. Thus, it might be argued, retributivisrs and compromise theorists have not been. attempting to justify punishmenr tut rather to analyse and clarify the precise moral standing of the practice; rherefore MpRLF, which applies to justifications rarher rhan theories or analyses, inapplicable. is I grant thar some theories and analyses of the justification of punishmenr have made useful and illuminating contriburions to our understanding of the moral standing of the pracrice, which would continue to be useful and illuminating even if the idea of a justification were abandoned (since what is useful and illuminating abolt them does not depend on that idea). But whatever else may bi said of or for it, if a rheory or analysis of the justification of punishmenr assumes thar a justification exists, then the theoretical or analytical enterprise must depend on the validity or coherence of thar assumprion. Rerriburive and compromise theories also assume rhat the justification cannot be a stricrly consequentialist one. If MpRLF rules out the possibiliry of a justificarion oi the kind retributive and compromise theories assume to exist and seek to specify, then MPRLF applies to them just as much as ro any straightforward artempt to justify punishment.r, A further objecrion ro the applicabiliry of MpRLF concerns rhe epistemic status of the counterfactual it contains. If MpRLF requires of any putarive jusrification of punishment that if the reasons oflered in justification are found not to be reasons for doing it, and if there is no orher moral justification, then we should be prepared to srop doing it. But suppose thar all the would-be jusrifications we know of are shown ro be invalid. How could we be sure thar there is not some valid justification that we have failed to consider? If we cannor be JUSTIFICATION FOR PUNISHMENT 97 sure of that, rhe objection continues, how can we be sure that we ought to abandon punishment? The epistemic objection might then be pressed into service as a second-order stand-in justification, of the form: 'we have a second-order moral reason for punishing offenders for as long as we have not ruled out the possibility that there is no valid first-order reason for doing so'. It is hard to credit that uncertainfy about how we would know that we have exhausted the possible justifications should stand in the way of our being able to say what we would do if we were sure thar we had exhausted rhem. But let the objection stand: there are at least rwo replies. In the first place the objection fails to consider that there might be a proposition P that demonstrates conclusively thar all moral justifications for punishment are necessarily invalid. \i[e have no more reason to deny rhe possible existence of P than we have to rule out the possibiliry that there might be a valid justificarion of punishment we have never considered. I( for all we know, P is possible and if knowing that P would not lead us to abandon punishment, then we can imagine ourselves to be in a situation where all putative moral justificadons of punishment have been shown to be invalid but which would nor lead us ro give up punishing offenders; in which case MPRLF continues to apply and retributive and mixed justifications cannot satisfy it. A second reply is that, eyen if an advocate of rhe objection could find a way of ruling ou. the possibility of P, rhe objection would only serve to confirm the argument, since it claims that MPRLF does not apply to iustificarions of punishmenr, whereas MPRLF must be satisfied by any coherent attempt at moral justificarion. In what other conrext, after all, would we be prepared to count a person as genuinely attempting to justify an action or a prac.ice if we knew that, were we to prove the iustification wrong, he or she would carry on doing what they were doing anyway on the grounds rhat 'there might be a moral reason for doing this that I haven't thought of'? We come to the conclusion that the practice of punishing offenders cannot be given a moral justification. Does it follow from this rhat no sociery or communiry can ever be justified in punishing an offender, or that we ought to abandon the practice (even though we will not)? Not if there is a plausible alternative, in considering rhe moral standing of punishment, to seeking a justification for it. One 98 CRISPP possible alternative would be to recognise the principle that as other moral sentiments or attitudes that are associated with it, as expressing a foundational attirude or sentimenr in our shared moral outlook., If it is a foundation of a moral outlook we all share (in varying degrees, perhaps, or with varying qualifications) thar it is right and just to punish (some) wrongdoers, rhen the question of seeking a justificarion for the practice does not arise. Such an approach might also lead philosophical discussion about punishmenr to focus more on what can properly count as a jusr, humane and effective system for punishing offenders and about which kinds of wrongdoing ought to be punished. The case for such an approach, however, will not be pursued here. Whatever other possible moral grounding for punishment there might be, the enterprise of attempting a gineral moral justification for punishment should be abandoned. It is possible to imagine that a time may come when offenders are no longer punished or even that rhere are no laws left to be broken, but it is a misrake ro rhink thar might happen through any collective act of choice, still less rhat such a choice might be justified by a failure to find a moral justification for punishment. To seek a justification for punishmenr, as such, is a meaningless or incoherent undertaking; and the same is likely to be rue of any of those larger enterpriies of attempting to rest the legitimacy of the modern srate, and of the various duties, obligations and rights thar are bound up with it, on some form of contractual obligation which involves or depends on a justification for punishment." wrongdoers should be punished, as NOT well ES l. Honderich, T., Punishrrrent: The Supposed J*ttilications (Oxford: poliry 1989), pp.1l-14. 2. CL._Ten 'Crimc and Punishmenr', in p Singcr (ed.), A Cozpanion to Ethics lOxford: Blackwcll, l99l). 3. Honderich (nore l), pp.25-31. 4. The most well-known advocate of rhis view is Kant: see The philosoplry of Law ltrans. V Hastie, Edinbtrrgh: T & T. Clark, 1887), pp.t9.r-205. 5. This vicw was advanccd in scveral articles by H. L. A. Hart, collecr ed in his punishment and Responsibility \Oxford: Oxford Univcrsity press, 1958). For a useful discussion of this approach, and of thc scveral varicrics of pure rcrriburivism, s€e Honderich op crt; H.B. Acton led,.l,The Pbilosopb ol punshment lLondon: Macmillan, li69). sce also JUSTIFICATION FOR PUNISHMENT For 99 recent defence of Han's approach, see M. Clark, 'The Sanctions of rhe Criminal Ptoceedings 6f the Aistotelian Society, XCVII, I, pp.25-39, 1997). 6. This is to overlook the posibility rhat there might be a jusrification of punishment rhat is neirher retributive nor cons€quentielisr.Some philosophers have claimed to offer rheori€s of punishment, which although incorporating elements of retriburive and consequentialist rheories are differenr from either (see, for example, J. Hampton, .The Moral Education Theory of Punishment' in A.J. Simmons, M. Cohen, J. Cohen and C.R. Beirz (eds), Prrrishirent (Princeto Princeron Universiry Press, 1995). l[hether such theories are as different from rhe standard alternatives as rheir advocates take them to be is arguable, but, in any case, since such theories offer justifications of punishment which conrain, or purporl ro do iusrice to, rhe retriburiv€ principles that classical consequentialism violates, my argument will apply just as much to thim as to the standard alternatives to classical consequentialism. The only non-consequentialisr justification I can imagine that would escape the argument is a religious juitification accordinS to which punishment of offenders is iustified because God has decreed that it is iust, and in defence of which ir is argued that it would be impious ro expec Cod to provide a justificarion for His decrees. 7. D. D^vidsan, Actions dnd Euerrts (Oxtodt Oxford Universiry Press, 1980), p.9. 8. The force ot should here will depend on wherher the jusrification offered is a rational or e moral one. 9. In which cas€, we assume, she has a reason for what she does, but not one she would offer in iustificarion. 10. The requirement that one be prcpared not to do it, rarher than that one just not do is intended to take care of hypothetical cases such as Harry Frankfurt's infamous and ingenious Dr Black, who is ready and able to intervene ro cause a person to commir a contemplared act murder should they choose nor to carry it our (Ahernare Possibiliries and Moral Responsibility', Tbe,fouraal of Philosoplry (1969), pp,829-391, a La*'in t, of so rhat the person would commit murder wherher or not rhey chose io. What is importanr for PRLF are a justifier's inrendons in respect of failure of lustificatron, rather rhan whar rhey will actually do. 11. A complicarion: sometimes justifications are offered for actions in which the reasons cited are offered as reasons for orhers not ro interfere, withour any necessary implicarion rhar those are the reasons for which the acrions are performed,Alex, ler ui say, wishes his mother ro allow him ro srry up and w,rtch television unril 9 pm. He points out that the film he wants to watch is abour rhe Arn€ncan War of Independence, which he is currenrly studying in history class. If he could choose for himself, Alex would stay up and watch rhe film wirh or without rhe educational recommendation. The recommendation can nevertheless srand as a jusrification for his mother's permission. Whether Alex likes it or nor, however, in offering this iustificarion he commiri himsel[. on pain ot inconsisrency. ro acrepring lhar rf his molher promise5 lo record the film, and he has no fall-back yustifrcarion. he rs bound acceprher refusal with good grace. At any rare, and in the absence of avenging angels, jusrificarions of punishmenr are not offered in order ro obrain permission to place offenders under lock and key. A further complication is that justifications are somerimes offered for the acrions of others, in which cases rhere can be no implicarion, should a jusrification fail, that those who offer justification should be prepared to stop doing what they have attempted ro justify. Indeed, mosr jusrificadons for punishment mighr be construed in this way, although rhey are rypically offered as iustificarions of the penal practices o[ sociedes and communities of which punishment rheorists are members. At any rate, where reasons are offered in iustificarion of the acrions of others, PRLF rcqnire\ mutatis mutandis, th^t rhe justificarion can only counr as such, if ir is the case that one rs prepared to say rhar such acdons should nor be performed, or should cease, should CRISPP 100 those reasons b€ found not to be reasons for rh€ actions and there are no other reasons for them. The same applies to retrospective justifications, but with the qualificarion that we would be wrong to say thet an agent should not have acred in a way she believed to be iustified if she could not reasonably be expected to have known that rhe justification was invalid. 12. Although there might be alternarives in particular cases of law breaking and even for particular categories of crime. 13. I had previously imagined that the consequences of abandoning punishment, and with it deterrence, would be an utterly lawless state of affairs. My mishke was in thinking that an utterly lawless state of affairs could persisr on any significant scale for a significant lengh of time, and that we really have any choice about whether to live with a system of deterrence. 14. Perhaps like Samuel Butler's Mr Nosnibor who willingly accep$ a diet of bread and milk, together with a monthly scourging, as a cure for ihe moral sickness that has led him to embezzle a widow (Erchwon, Edinburgh: Ballantyne, 1908, PP.101f0. By the strange standards of Erehoon it is the widow, in allowing herself to fall into misfortune, who is the real villain of the piece, whereas Nosnibor's sickness is no cause for blame. 15. See A M Quinton,'On punishmenC in Acron (note 5). 15. B. F. Skinner, Wdld?n Two,2nd edn (London: Macmillan, 1976). 17. Utopian and dystopian novels fashion imaginative possibilities from features of our exisiing culture and technology. ln doing so they enable us to consider afresh features of our lives and our world and to consider where we might be headinS. There is no realistic constrainr on the p€rspectives and imaginative possibilities with which such works provide us. Correspondingly, they cannot enable us to arrive at definite judgments and conclusions, whether abour the future or the here and now, on which we might hope to agree. Philosophical and theoretical argument, by contrast, does require us at least to make the attempr to arrive at agreeable iudgments and conclusionr and therefore requires a degree of realism and precision about details that utopian novels do not. 8fi1)e New Wo d, with its imagined Senetic, chemical and social technologies, may suggest ways of thinking philosophically about our exisrinS moral lives and pracrices, but it do€s not and cannot provide us with sufficient material for reaching well rhought out conclusions. 18. Those familiar with Peter Strawson's'Freedom and Resenrmen.'will recognise rhis argument as an adaptation of his reiecrion of the norion that we might come to abandon the inter-personal reactive aftitudes because of an acceptance of the truth of determinism: If we could imagine what we cannot have, viz. a choice in this matter, then we could choose rationally only in the lighr of an assessment of the gains and losses ro human life, its enrichmenr or impoverishment; and rhe rruth or falsiry of a general rhesis of determinism would not bear on the rationaliry of this choice ('Freedom and Resentment', in G. Warson led.], Frce Will [Oxford: Oxford University Press, 19821 19. p.7 0). This should be kept in mind by those who see only ideological pretence in philosophical discussions abour punishment. It is easy to be led from noticing the inequalities and injustices of exisring legal and penal sysrems, and the powerful inrerests they unequally serve, ro thinking that law and punishment can only serve the interests of the wealthy and powerful. If the law serves most of us less well than ir does the wealthy and powerful, and even if it is somerimes shamelessly used in rhe service of the few against rhe many, most people, including many who are relatively badiy served by our existing legal and penal sysrems, would be considerably less well_served by a system of warlordship. Nevertheless, the objection that philosophical discussions JUSTIFICATION FOR PUNISHMENT 101 alout the iusrification of punishmenr have side-stepped more pressing quesrions about the iustice of pracrices of punishint offenders is imporranr. 20. The comparative injustice of the alternative to our currenr practices might be rhoughr of as providing sufficient moral iustification for conrinuing with our currenr pracdc;s. Bur this would be no more rhan a negarive variarion on the srricr consequentialist justification and therefore open to rhe same obiecrions. 21. C[ the disrinction made by K. G. Armstrong,('The Retributivisr Hits Back,, in H. B. Acton fed.l, (nore 5), pp.141 between the quesrion of r.//ry we punish and why it is morally permissible for us ro do ir. 22. A different but related objection has been put to me by David Cockburn in respect of rhe argument from analogy about other minds. Philosophers who think that thire is a need for an argument from analogy might not think that belief in the mental lives of others is owed to such an argument and may accept rhat the belief would hold even if all arguments from analogy were found to be invalid. Instead, they may see the argumenr as needed to provide a justification for what we believe on instincr. Cockburn suggesrs that ir'can be importanr to how one feels about what one is doing that one can see it to be rarionally iusrified even if there is no question of abandonin! that way of thinkint or acting'.I am doubrfulabout whether defenders of rhe argument from analogy would accept this. The srandard purpose of the argument from ;nalogy is to establish that belief in other minds has epistemic warranri in other words. thaiit can counr as knowledge. If someone has a rrue belief and rhere are good reasons for them to have the belief bur rhe good reasons are nor their reasons, their having rhe belief falls short of knowing it to be true. If were true that belief in other minds is insrinctive, rhen the argumenr from analogy might stand as a jusrification of ir if ir could be shown that the insrinct were somehow based on the argument from analogy. If instead the insrinct was unrelared ro rhe argumenr, then all rhe argument could sh6w is rhat the belief is harmless: nor rhat it is lustified. 23. See K. Magill, F/"€dom and Expeimce: Self-Deterrninationwithout lllusions (Lordo1: Macmillan and New York: St Martin's Press, 1997), Chapter 2 for an examination of the implicarions for treating our practices of holding agents morally responsible, prairrng, blaming. erc. rs foundational. 24. Earlier versions of this paper were presented ar seminars and meerings at Edinburgh, Lampeter, Swans€a and Wolverhampron, and at the 1995 Political Studies Association Conference in Glasgow, collected in Contemporary Political Studies: Proceedings of the Polirical Srudies Aslociation conference, J. Stanyer and I. Hampsher-Monk (ids). (Oxford: Blackwell, 1995). I am grarefirl to all of those who took pait in discussron on those occasions, and for commenrs from Alan Apperly, Bryn Browne, David Cockburn, Oswald Hanfling, Kimberly Hutchings, Presron King and Harry Lesser I am also grateful ro Presron King for helpful editorial advice and for the criticisms and ruSSestions of two anonymous re.rders. t