Matt Matravers , Justice and Punishment: The Rationale of Coercion

書名: Justice and Punishment: The Rationale of Coercion

作者: Matt Matravers (Author)

出版社: Oxford University Press, USA (October 5, 2000)

語言: English

ISBN-10: 0198295731

ISBN-13: 978-0198295730

電子書格式：DOC (Word 文檔)

Book Description

He asked a very simple question: Why, and by what right, do some people lock up, torment, exile, flog, and kill others, while they are themselves just like those they torment, flog, and kill? And in answer he got deliberations as to whether human beings had free will or not; whether or not signs of criminality could be detected by measuring the skull; what part heredity played in crime; whether immorality could be inherited; and what madness is, what degeneration is, and what temperament is; how climate, food, ignorance, imitativeness, hypnotism, or passion affect crime; what society is, what its duties are—and so on . . . but there was no answer on the chief point: By what right do some people punish others?

This book attempts to answer the challenge of showing that morality is not a confidence trick or a fetish. It does so by arguing that moral norms are those that rational, self‐interested people could accept. The problem is approached by asking by what right some people punish others, and by comparing recent developments in theories of distributive and retributive justice. The first part of the book considers retributive, utilitarian, and mixed theories of punishment. In the second part, recent theories of distributive justice, especially those of Rawls and Gauthier, are examined. It is argued that these theories cannot give an adequate account of punishment. In the final part, an argument is offered for a genuinely constructivist account of morality—constructivist in that it rejects any idea of objective, mind‐independent moral values and seeks instead to construct morality from non‐moral human concerns; genuinely constructivist in that, in contrast to Rawls, it does not take as a premise the equal moral worth of persons. The conclusion is that a genuine constructivism will show the need for, and justification of, punishment as intrinsic to morality itself.

Contents

Introduction

1 Consequentialism

2 Retributivism I: Fair Play Theory

3 Retributivism II: Resentment, Guilt, and Censure

4 The Scope of Impartial Justice

5 Impartial Justice, Motivation, and Punishment

6 Justice as Mutual Advantage

7 Self‐Interest and the Commitment to Morality

8 A Constructivist Theory of Moral Norms

9 The Moral Community, Justified Coercion, and Punishment

About the Author

Matt Matravers is at the Department of Politics at University of York.

Introduction

Matt Matravers

He asked a very simple question: 'Why, and by what right, do some people lock up, torment, exile, flog, and kill others, while they are themselves just like those they torment, flog, and kill?' And in answer he got deliberations as to whether human beings had free will or not; whether or not signs of criminality could be detected by measuring the skull; what part heredity played in crime; whether immorality could be inherited; and what madness is, what degeneration is, and what temperament is; how climate, food, ignorance, imitativeness, hypnotism, or passion affect crime; what society is, what its duties are—and so on . . . but there was no answer on the chief point: 'By what right do some people punish others?

(L. Tolstoy, Resurrection) 1 

1 Quoted from Pincoffs 1966: epigraph. It is also quoted in Duff 1986: 187.

The aim of this book is to answer Tolstoy's question: why and by what right, do some people punish others? This is not a new question. The problem of punishment is one of the most enduring in political theory. Whilst other topics—political obligation, power, democracy, and the like—go in and out of fashion, there is a seemingly constant and substantial stream of contributions to the discussion of punishment. 2 

2 For useful summaries of the contemporary state of punishment theory see Duff and Garland 1994; Duff 1996.

The result is that penal philosophy can now reasonably make a claim to being a separate sub-discipline of moral and political philosophy and, indeed, it has been treated as such. In part, my argument is that this separation of punishment theory from the wider concerns of moral and political philosophy is damaging to both. For a theory of punishment to be adequate, it must be rooted in a broader moral theory, and that broader moral theory will, I believe, be constructivist.

An advantage of this approach is that it explains a puzzle. Political theory has been dominated over the last three decades or so by the topic of justice. However, recent theories of justice have for the most part concentrated on distributive justice and have excluded retributive or corrective justice. Indeed, the term 'justice' as used in the literature has almost come to be

end p.1

synonymous with 'distributive justice'. Yet the right to punish, and the role of punishment in understanding the state and political legitimacy, was once taken to be central in political philosophy. 3 

3 Michael Davis makes the strange claim that before the latter half of the 18th century punishment was of little or no concern to political and moral theorists. He writes, in support of the idea that punishment theory is independent of political theory, that 'Hobbes was unusual in giving it [the topic of punishment] two chapters of Leviathan. Most political philosophers (for example, Locke in The Second Treatise or Rousseau in The Social Contract) said much less, and what they said was perfunctory' (Davis 1989, 321). In this he is surely mistaken. Locke grants man in the state of nature (only) 'two Powers', one of which is 'the power to punish the Crimes committed against [the] Law [of Nature]' (Locke 1988: ii, § 128), and it is this power—or right—that is given up to the political association and thus forms the basis of the State. Indeed, Locke defines political power to be 'a Right of making Laws with Penalties of Death, and consequently all less Penalties, for the Regulating and Preserving of Property' (Locke 1988: ii, § 3). Rousseau is less explicit in placing punishment at the centre of his concerns but, in the famous opening paragraphs of The Social Contract, it is clear that what is to be 'made legitimate' is the 'chains' on people, which are backed up by the threat of force.

The question then arises, why has distributive justice become separated from retributive justice?

My argument is that this separation has occurred because contemporary theories of justice cannot explain the relationship of justice and morality more broadly conceived. As it is this relationship which a theory of punishment needs to explain, it is in examining the problem of punishment that the limitations of contemporary theories of justice are most starkly exposed. Moreover, those limitations are not confined to the response of contemporary theories of justice to the problem of punishment, but are such as to undermine also the accounts of justice. Put bluntly, the argument pursued here is that it is through the discussion of punishment that the inadequacies of contemporary theories of justice are demonstrated and it is therefore through the discussion of punishment that those inadequacies can be rectified.

The claim that punishment theory needs a wider moral context is controversial. The first part of this Introduction begins the defence of this claim (a defence that is completed by the rest of the book). The second, describes the argument that follows from its being true.

For some, the thought that the question of punishment is 'relatively independent' (cf. Davis 1989) of moral and political theory more broadly is encouraged by what appears most obviously problematic about punishment. This is that punishing someone involves doing to her things that would normally be in violation of her rights. As one commentary puts it, 'it is usually wrong to lock people up, to take their money without return, or put them to death' (Duff and Garland 1994: 2; cf. Knowles 1999: 29).

end p.2

Given this, it might be taken that the central questions for the penal philosopher are, 'when it is appropriate to do such things?'; 'to whom should they be done?'; and 'in what quantities?' However, this is to miss an important point. Punishment follows from the contravention of a rule that has attached to it a provision that those who contravene the rule will be liable to punishment. Before asking whom, under what conditions, and how much, what is needed is, as Stanley Benn puts it, 'to ask what can justify punishment in general'. That is, 'to ask why we should have the sorts of rules that provide those who contravene them should be made to suffer: and this is different from asking for a justification of a particular application of them in punishing an individual' (Benn 1958: 326; cf. Hart 1959).

To give a very rough picture: three distinct sets of questions can be discerned: (i) questions about the rules and their foundations, scope, and authority; (ii) questions about what justifies the threat that those who contravene the rules under certain conditions will be liable to punishment; (iii) questions about what justifies the particular application of the threatened sanctions.

For those about whom Tolstoy complains, the primary enquiry of punishment theory is into (iii). For what is at stake in questions of skull size, heredity, and passion is whether the conditions for liability are met in full by some particular offender. Similarly, consider the following example offered by the Oxford philosopher John Mabbott:

I was myself for some time the disciplinary officer of a college whose rules included a rule compelling attendance at chapel. Many of those who broke this rule broke it on principle. I punished them. I certainly did not want to reform them; I respected their characters and their views. I certainly did not want to drive others into chapel through fear of penalties. Nor did I think there had been a wrong done which merited retribution. I wished I could have believed that I would have done the same myself. My position was clear. They had broken a rule; they knew it and I knew it. Nothing more was needed to make punishment proper. (Mabbott 1939: 42)

Mabbott clearly distinguishes between the rule and the threat of a penalty attached to the breach of that rule. He clearly believes the rule compelling attendance at chapel to be unjustified, but that is for him not relevant to the question of the justification of punishment, which he takes to be the question, 'under what circumstances is the punishment of some particular person justified and why?' (Mabbott 1939: 39). As is clear, Mabbott believes that the punishment of a person is justified by that person having culpably broken a (correctly instituted) rule. The punishment falls on that

end p.3

particular person because it is he who has broken the rule. Presumably having established that a given student did not attend chapel the only question for Mabbott is whether the student was culpable in so doing. This is surely insufficient. Although Mabbott's account might be taken to be a comment on the proper application of punishment—if Mabbott had always pardoned one favourite offender whilst punishing all the others he would have done something unjustifiable—it cannot seriously be taken to be a justification of punishment in the sense in which such a justification is required.

However, it might be thought that whilst an adequate theory of punishment must do more than address (iii), penal philosophy can nevertheless limit its enquiry to (ii) and (iii), but need not engage with those questions grouped under (i). It might be argued, for example, that the content, source, and nature of the rules that govern a community is a matter for political and moral philosophy. Assuming that some set of rules can be justified then it is a reasonable assumption that it would be a good thing, other things being equal, to reduce the incidence of rule breaking. Preventing, or diminishing the number of acts of rule breaking provides the rationale for appending to the rules the threat of sanctions, and penal philosophy concerns the subsequent questions that arise about that threat. Alternatively, it might be thought that whatever the source, nature, and authority of the rules the failure to abide by them on the part of the offender calls for some response. This response may be one of condemnation; or it may be best captured in the language of 'cancelling' or 'annulling' the act of the offender; or in the claim that justice needs to be restored to the relations of the offender and her community.

These two examples of how punishment theory might be thought to be relatively independent gesture towards the traditional way of conceiving of the 'punishment debate': that is, as a debate between consequentialism and retributivism. It is doubtful that this picture was ever accurate and it certainly now fails to capture the subtle and sophisticated ways in which punishment theory has developed. 4 

4 So much so that the labels 'consequentialist' and 'retributive' are of increasingly little use as the theories that they are meant to group together have become so diverse.

However, one reason for the continuing resonance of the consequentialist versus retributive presentation of the problem of punishment is undoubtedly that each position seems to contain something that is essential. Another is that neither retributivism nor utilitarianism need, at first sight, to address those questions grouped under (i). My argument is that neither can, in fact, escape this requirement, but before considering that it is worth commenting briefly on what it is of importance that is captured by each position.

end p.4

Retributivism, in looking to the past act of the offender, connects the punishment of the offender to what it is that he has done and thus captures two important intuitions: that punishment is justified if (and only if) it is deserved—if it is 'of an offender for an offence'—and that the amount of punishment imposed ought to be in some way proportional to the wrong done. Yet, retributivism is also a strange doctrine. It no doubt appeals to a strong intuition that 'the guilty deserve to suffer', but it contradicts another, that 'two wrongs do not make a right'. Moreover, retributivists do not seem to be able to explain why the guilty deserve to suffer, and when asked to do so they often do no more than restate this basic proposition in different, and often 'mystical' terms (Lacey 1987: 410; Chapters 2 and 3 below). The consequentialist argument that if it were certain that punishment yielded absolutely no benefit in either social or individual terms it would be wrong to continue to punish is, for many, quite compelling. 5 

5 Although not, of course, for Kant who notoriously writes that even on the point of the dissolution of society 'the last murderer remaining in prison would first have to be executed' (see Kant 1991: § 333, 142).

For their part, utilitarian justifications of punishment capture the idea that one purpose of having rules that threaten sanctions against those who contravene them is to reduce the incidence of prohibited acts. In addition, as with all consequentialist arguments, utilitarian theories of punishment appeal to two further intuitions: that human well-being matters and that moral rules ought to be evaluated in light of their effect on human well-being (Kymlicka 1990: 11; Chapter 1 below). Where consequentialist theories seem to fail our intuitions is in being entirely forward looking. Taking into account only future consequences appears to ignore the importance of the past act. Furthermore, because of its concern for aggregate human well-being, utilitarianism appears to be blind to distributive concerns. It is unjust to punish an offender's family even if doing so would deter future offending and, thus, bring great gains in aggregate human well-being (Chapter 1 below).

The insights captured by retributivism centre on who should be punished (those who deserve it) and on the quantum of punishment that may justifiably be imposed (that which is in proportion to desert). The problems associated with the account stem from the question of why punish at all. Consequentialism, conversely, appears best to address the question of what justifies having a system of punishment, and its difficulties emerge when considering the distributive questions of whom and how much to punish. One well-established way of responding to these features of the debate is to separate what Hart calls the question of the general justifying

end p.5

aim of punishment from the question of its distribution (Hart 1959: 4; see also Benn 1958; Clark 1997). This allows the combining of consequentialist and retributive approaches in a 'mixed theory', the most common form of which takes the answers to the distributive questions ('Whom?' and 'How much?') to be side constraints on a utilitarian understanding of the general justifying aim of punishment (see Benn 1958; Hart 1959; Clark 1997). 6 

6 For one example of an argument that the right to punish is derived from retributivism, but the quantity from consequentialist considerations, see Bradley 1927: 27: 'Having once the right to punish, we may modify the punishment according to the useful and the pleasant; but these are external to the matter, they can not give us a right to punish, and nothing can do that but criminal desert. This is not a subject to waste words over: if the fact of the vulgar view is not palpable to the reader, we have no hope, and no wish, to make it so.'

Whether such a mixed theory can be successful is a question taken up later, the crucial test being whether the theory can successfully integrate its consequentialist and retributive (or other distributive) components. What is of interest here is whether such mixed theory can sustain the claim of relative independence.

On the usual understanding of mixed theory—a preventive rationale for the general justifying aim of punishment constrained by retributive distributive principles—the argument remains in the domains of questions (ii) and (iii). The argument begins not with the nature, scope, authority, etc. of the rules, but with the need to prevent breaches of those rules (the pursuit of that preventive aim being constrained by distributive concerns). However, despite its very widespread use, Hart's formula contains an ambiguity remarked upon at least as early as 1961. 'It is important to notice', Armstrong argues,

that the moral justification of a practice is not the same thing as its general point or purpose, except in the eyes of those who have travelled so far down the utilitarian road that they never question the means if the end is desirable. . . . An act or practice may have a very sound point indeed and still lack moral justification . . . . There is an ambiguity in a phrase like 'The general justifying aim of punishment', between why we do it, and why it is morally permissible—if it is—for us to do it. (Armstrong 1969: 141; cf. Dolinko 1991: 541) 7 

7 In addition to arguing that Hart's formula of 'the general justifying aim of punishment' is ambiguous, Dolinko claims the separation of the distributive question of 'whom to punish' from the question of the moral permissibility of the practice of punishment to be 'preposterous' for 'to split off . . . the question of who may be punished . . . suggests that we can decide what punishment is and why we engage in it without knowing who is supposed to receive punishment'. As Dolinko correctly notes, Hart surreptitiously builds the answer to the distributive question into 'the supposedly separate "definition" of punishment' as 'of an actual or supposed offender for an offence" (all quotations from Dolinko 1991: 541).

end p.6

What Armstrong points towards is the claim that punishment theory must concern itself with the morality of attaching the threat of sanctions to rules (as well the morality of imposing those sanctions on particular people). And whilst it seems plausible to think that the point of threatening sanctions must have something to do with preventing offending (see Chapter 3 below), that is not the same as arguing that preventing offending through the threat and imposing of sanctions is morally permissible.

Nevertheless, it might still be argued that the question of the moral permissibility of attaching threats to rules can be addressed without the need to consider questions concerning the rules themselves. The argument that this is not so is, in a sense, presented throughout this book. However, that it is not so is surely plausible. That what informs the moral justification, rather than practical rationale, of attaching the threat of penalties to rules is only that such penalties reduce the incidence of rule breaking is implausible. The argument must be more complex and part of the complexity can only be unravelled by considering the rules to which those threats are to be attached and the agents to whom they are to apply. Addressing the question of why it is a good thing to prevent the breaking of some rules would seem to require addressing the question of what is the nature and content of those rules.

Even were prevention the major component of a moral justification of punishment, questions would still arise as to who ought, or has the authority, to threaten such sanctions. In addition, whether it is morally permissible to prevent offending by threat of sanctions, even where that threat is made by a legitimate authority, is, in part, a matter of whether it is morally permissible to alter the behaviour of agents by threatening them. In turn, this question raises the issues of the relationship of prudential and moral reason and of their relationship to agency.

The case is perhaps even clearer when it comes to the retributive claims that some action in contravention of some rule calls for a response, an annulment or cancellation, or other corrective to right the wrong committed. If these claims are to be made meaningful the strategy must be to link them with the wrongdoing in the offending act. Retributivism, if it is not to be either circular and mystical or irrelevant to the moral justification of punishment must show that such responses are morally permissible (or obligatory) by a correct understanding of what is morally owed to responsible agents or by such an understanding of the nature of the rules.

The claim that a moral justification of punishment needs to give an integrated account of questions (i) and (ii), just as it does for (ii) and (iii), is plausible. In the first three chapters that follow the claim is substantiated.

end p.7

It is argued that existing approaches to the theory of punishment have failed successfully to justify punishment either because they have not taken seriously the need to integrate the account of punishment and a broader moral theory or because the moral theory on which they do rely is itself inadequate. Chapter 1 examines the utilitarian account of punishment. Utilitarians, of course, have available a complete moral theory: an account of the rules, of what considerations determine whether those rules should have a threat of sanctions attached to them, and of when the use of those sanctions is morally permitted. Treated as a straightforward teleological theory in which first the good is identified and then the basic moral command is to do that which will maximize that good, utilitarianism can be quickly dispensed with. It makes little sense to claim that there is a moral obligation owed to states of affairs. Treated more seriously as a way of giving content to the demand that equal respect is shown to individuals, utilitarianism is a much more attractive theory than is commonly recognized. However, the counting of all wants in a social decision function raises a problem of external preferences that ultimately undermines the account. The utilitarian account of punishment fails because it relies on a utilitarian moral theory that is ultimately indefensible.

Chapters 2 and 3 focus on retributive theories. Retributivism looks initially to be the most plausible candidate for being an independent account of punishment, focusing as it does on the idea that, in some sense or other, culpable offenders deserve punishment. Chapter 2 examines what is arguably the most important retributive school of recent years, which ties punishment to the restoring of a just balance of benefits and burdens between law-abiding citizens and free-riders. Chapter 3 looks at expressive and communicative retributive accounts. In all these cases the argument is that in order to maintain the relative independence of punishment theory and the claim that punishment is to be understood as deserved by the past act of the offender, retributive theorists separate the understanding of the wrong done in the offence from the wrong to which punishment responds. Such a strategy is untenable. Moreover, insofar as they surreptitiously appeal, or need to appeal, to an understanding of the rules to which the threat of penal sanctions are to be attached, what they provide is inadequate.

Given that the argument of the first three chapters is that much punishment theory does not succeed because it fails to take sufficiently seriously the need to integrate a justifying account of punishment with a larger political and moral theory, the next chapters look to contemporary accounts of justice to provide such a theory. It is here that the puzzle referred to above

end p.8

appears. Although concerned with the topic of justice, recent work on distributive justice by Rawls and his followers has very little to say about punishment theory and penal justice. The argument pursued in Chapters 4 and 5 is that this is because theories of justice, at least those that can be classed as 'impartialist', cannot adequately explain the relationship of justice to morality more widely. In Chapter 4, it is argued that such theories, if they are to give an adequate account of the public domain, must appeal to considerations that fall outside the narrow confines of an impartialist account of distributive justice. However, if they do so appeal then they must give up the claim of impartiality. Chapter 5 focuses on the structure of impartialist theories and brings to the fore a distinction between faux and genuine constructivism that underpins the remainder of the book. Faux constructivists do not, as the name suggests, attempt to construct moral norms, but rather use the techniques of constructivism to explicate the demands made by an initial assumption of the fundamental equality of persons. The result is that the demands of justice, or of morality, appear to the agent as alien restrictions on the pursuit of his good. Although faux constructivists attempt to fill the 'gap' that they open up between the agent's personal perspective and the demands of morality they cannot adequately do so. The result is that the threat of sanctions is needed to attempt to fill this gap. Such threats must be attached to existing moral rules and the justification for so doing must appeal to one or other of the theories discredited in Chapters 1-3. The problem with the account of punishment lies with the faux nature of the constructivism adopted by these theorists.

Given the identification of the problem as lying in the separation of the demands of morality and the agent's self-interest, the argument turns to an attempt to ground moral reasons in prudential ones; to the genuine constructivism of David Gauthier. Gauthier sets out to show that there are rationally compelling reasons for a non-tuist rational agent to adopt the constraints of morality. However, in Chapter 6 it is argued that he does not succeed. Given the resources that he has available to him, Gauthier cannot show that the rational non-tuist has reason to treat moral norms as anything but instrumentally useful to the pursuit of his narrow self-interest. Obedience to those norms is thus a matter of prudential calculation and this is inadequate. A theory of morality must be able to explain moral norms' 'imperatival force' (Charvet 1995: 1): the authority that such norms have over the pursuit of narrow self-interest. Gauthier closes the gap between morality and self-interest only by dissolving the former into the latter.

However, in recent work Gauthier has suggested that meeting the demands of constructivism may not require constructing morality from

end p.9

concerns quite as narrow as those that he allowed himself in his 1986 book, Morals by Agreement. This idea is pursued in the development of a genuine constructivist account of morality in Chapters 7 and 8.

The central claim of these chapters is that the closing of the gap between self-interest and morality does not require that morality can be shown to be rationally compelling even to the egoist. There are no rationally compelling reasons to be moral. However, if the constructivist enterprise begins with an already social agent, standing back from her commitments, asking whether she has reason to endorse the priority of moral reasons, then there surely are reasons that can be given to show that such a commitment to morality is not a mistake or a confidence trick. The commitment to morality must be a commitment that goes beyond what is dictated by reason, but that is not to say that reason is impotent. The argument of Chapters 7 and 8 is that there are important and substantial reasons to be moral that can be understood from the perspective of self-interest. These reasons are not decisive, but are sufficient to meet the challenge that morality is a fraud, a mistake, or a confidence trick.

The constructivist theory sketched in these chapters ties together self-interest and morality, not by reducing the former to the latter nor by surreptitiously appealing to faux constructivist assumptions. It claims that morality must be understood as the product of a commitment by agents to recognize the imperatival force of moral norms. Such a commitment is only possible where each agent understands his good as best pursued together with others in a community regulated by such norms. Of course, this commitment can only be reasonable when it is made together with others. To commit oneself to the authority of moral norms in circumstances in which one is surrounded by those who would exploit and free-ride on that commitment would be nonsensical. In such circumstances, morality is indeed a confidence trick, which the agent would do better to avoid.

Given this understanding of morality and of the moral community, the argument is that the will to commit to morality also commits the agent to renouncing free-riding (understanding moral norms as merely of instrumental value in the pursuit of one's narrow self-interest). Social norms that govern co-operation can only be understood as moral in conditions in which agents are secure and know that the commitment to morality will not lead to their being taken advantage of. Moral norms are, thus, necessarily coercive, for the commitment of those participants in the moral community to the authority of moral norms is also a commitment to coerce those who would jeopardize the moral project by free-riding. Thus, the account of morality integrates the coercive threat of sanctions and the

end p.10

broader moral theory. It does so only because it is able to close the gap between the personal perspective and the demands of moral reasons. Rather than being alien constraints on the pursuit of self-interest (as they are in faux constructivist theory) or being 'evils' from which the rational self-interested being 'would be free' (as they are in Gauthier's account), the agent commits himself, together with others, to understand and to pursue his good through the demands of moral reasons. What is achieved is the integration of rules of justice, coercion, and morality in a single, genuinely constructivist account.

Although the place of coercion is central to the argument of Chapters 7 and 8, the answer to Tolstoy's question is not complete until Chapter 9. In understanding morality as a product of a commitment, the theory must allow that some will reject that commitment and the idea of moral cooperation. Such people stand in no moral relation to others and, thus, although they may be coerced and suffering may be inflicted upon them, this cannot be done in the name of punishment. Rather, these people stand in a state of nature relationship to the moral community. However, those who do make the commitment to morality will, of course, need to be assured that by doing so they do not make themselves vulnerable. In integrating self-interest and morality the account does not obliterate the demands of narrow self-interest and, given that agents will sometimes give in to temptation and behave immorally, an account is needed of the coercion and punishment of those members of the moral community who do not renounce their commitment, but sometimes fail to abide by its demands. Chapter 9 develops such an account. There I argue that in constructing the norms of co-operation, the contracting parties in this constructivist theory would adopt an understanding of punishment which is coercive, which threatens hard treatment, but which is also both communicative and educative.

In short, the strategy pursued in this book is to examine the problem of punishment both for itself and for what it reveals about contemporary theories of justice and of morality. That some such theories cannot give an adequate account of punishment is evidence, or so it is claimed, of deeper flaws in their accounts. Avoiding these flaws ought, therefore, to provide not merely an account of morality, but also an adequate account of punishment. Chapters 1-3 reveal what is needed in an account of punishment. Chapters 4-6 reveal why it is that contemporary theories of justice cannot provide what is needed. Chapters 7-9 attempt to develop a theory that meets the demands of an account of punishment and through this the attempt is made to develop a more adequate moral theory.