1 My concern with Hume's theory of responsibility in this context will be largely subservient to my primary concern which is Hume's general theory of punishment. I will not, accordingly, provide any detailed analysis of Hume's theory of responsibility. For a more detailed account of Hume's views on responsibility see my ‘On the Naturalism of Hume's “Reconciling Project”,’ Mind 92 (1983), 593-600. Of related interest see also ‘Causation, Compulsion and Compatibilism,’ American Philosophical Quarterly 25 (1988), 313- 21.

2 There are, in general, two closely related elements involved in Hume's naturalism. The first is that which concerns Hume's project of ‘science of man’ (i.e. his empirical, scientific investigation into the principles of human nature). The second is that which concerns Hume's emphasis on the role of feeling in human life. Both these elements constitute major themes inHume's philosophy and, as I will show, they both play an important role in shaping Hume's approach to the problem of punishment. On Hume's naturalism in general see, e.g., Barry Stroud, Hume (London: Routledge & Kegan Paul1977), esp. Chs. 1 and 10. See alsoP. F. Strawson, Skepticism and Naturalism: Some Varieties (London: Methuen 1983), esp. Ch. 1, sect.3.

3 References will be to the Selby-Bigge edition of Hume's Treatise of Human Nature, revised by P.H. Nidditch (Oxford: Oxford University Press 1978); to the SelbyBigge edition of the Enquiries, also revised by P.H. Nidditch (Oxford: Oxford University Press 1975); and also to Essays Moral, Political and Literary, T.H. Green and T.H. Grosse, eds., 2 vols. (London: Longmans 1875). The Treatise is abbreviated as THN; the first and second Enquiries are abbreviated as EHU and EPM respectively.

4 See, e.g., THN, 410: ‘Tis indeed certain .. .'; EHU, 97-8: ‘All laws being founded .. .'; THN, 609: ‘Men have observ'd .. .'; EPM, 187: ‘When any man … .’ It is an interesting question why Hume has relatively little to say on the subject of punishment. Suffice it to say, however, that for Hume, it would seem, it is the sense of honour and shame (i.e., pride or humility) that attaches to virtue and vice, rather than the prospect of reward or punishment, which is the principal motive to virtue. In other words, it is how we feel about ourselves, rather than our beliefs regarding how others will treat us, which is the sure path to virtue.

5 Cf. THN, U, 2, esp. sect. 7. Some commentators, probably with these passages in mind, interpret Hume's views on punishment as being fundamentally utilitarian in nature and thus ‘as necessitating the rejection of desert’ Oohn Kleinig, Punishment and Desert [The Hague: Nijhoff 1973), 49). By and large, however, commentators on Hume's philosophy have little or nothing to say on this subject. We find, in other words, that the thinness of Hume's remarks on this subject is reflected in the relevant secondary literature.

6 See, in particular, Hobbes's Leviathan, C.B. MacPherson, ed. (Harmondsworth: Penguin 1%8 [1651]), Ch. 28; Schlick, ‘When is a Man Responsible?,’ reprinted in B. Berofsky, ed., Free Will and Determinism (New York: Harper & Row 1966), 60.

7 That is to say, most philosophers would accept John Mackie's claim that Hume articulates the ‘classic defence of compatibilism’ O.L. Mackie, Ethics [Harmondsworth: Penguin 1977), 245). Hume's discussion of the free will problem appears at THN, 11,3,1-2 and EHU, Sect. 8. For a standard or orthodox interpretation on Hume's views on free will see, for example, Barry Stroud, Hume, ch.7.

8 Thus Isaiah Berlin, for example, speaks of the ‘Hobbes-Hume-Schlick doctrine': Four Essays on Liberty (Oxford: Oxford University Press 1%9), xv. Similarly, Jonathan Glover argues that ‘almost identical versions of this doctrine [sc. compatibilism] are to be found in Hobbes, Hume, Mill, Russell, Schlick and Ayer': Responsibility (London Routledge & Kegan Paul 1970), SOn.

9 Hobbes, ‘Of Liberty and Necessity,’ selections reprinted in D.O. Raphael, ed., British Moralists: 1650-1800 (Oxford: Oxford University Press 1%9), 63-5; Schlick, ‘When is a Man Responsible?,’ 59-61. The fundamental point which both Hobbes and Schlick are concerned to make in this context is that rewards and punishments serve to cause people to act in some ways rather than others and are, hence, of considerable social utility (cf. Hume, T,410). Schlick goes so far as to analyse responsibility in these terms. That is to say, on his account the issue of responsibility is to be interpreted in terms of the question regarding who it is of utility for us to punish (60). (A similar ‘pragmatic’ or ‘forward-looking’ approach is pursued in P. Nowell-Smith, ‘Freewill and Moral Responsibility,’ Mind 57 [1948], 45-61; and J.J.C. Smart, ‘Free Will, Praise and Blame,’ reprinted in G. Dworkin, ed., Determinism, Free Will and Moral Responsibility [Englewood Oiffs, NJ: Prentice Hall 1970), 208-13.)

10 I have argued elsewhere that this general perspective on Hume on free will is, in important respects, mistaken. (See the papers referred to in note 1.) In this paper I show that this mistaken perspective on Hume leads directly to confusion about his views on punishment.

11 See, more generally, THN, II, 2, 6. I follow Pall Ardal in thinking that Hume holds that the moral sentiments are calm forms of love and hatred. See his Passion and Value in Hume's Treatise (Edinburgh: Edinburgh University Press 1966), Ch. 6. Other relevant passages can be found at T,348, 418 and 439; EPM,302 and also Essays, Vol. 2, 139.

12 D.O. Raphael, ‘Hume and Adam Smith on Justice and Utility/Proceedings of the Aristotelian Society 73 (1973), 99; reprinted in justice and Liberty (London: Athlone 1980).

13 Essays, Vol. 2, 399-406

14 Hume repeatedly points out the difficulties involved in attributing human passions or sentiments to God: Dialogues Concerning Natural Religion, N.K. Smith, ed., 2nd ed. (Edinburgh: Nelson 1947), 156, 212, 226.

15 Of course, this is a separate issue from the question concerning the utility of (religious) belief in the existence of such a future state. That is to say, as Hume recognized, many would claim that we have utilitarian reasons to promote and encourage belief in the doctrine of future rewards and punishments. In reply to this (distinct) claim Hume maintains that beliefs of this general nature are, at best, of little social utility and that they in fact tend to undermine our natural human commitment to morality and the laws of society (D, 219-24; see also THN, 113-15; EHU, 147).

16 My approach here draws heavily on H. L.A. Hart's influential statement of the teleological retributivist (i.e. ‘mixed’ or ‘reconciliationist’) theory of punishment in ‘Prologomenon to the Principles of Punishment,’ reprinted in Punishment and Responsibility (Oxford: Oxford University Press 1968), 1-27.

17 Cf. THN, 348: ‘Nothing is more evident … .'

18 Hume's position on this issue may be contrasted with Austin Duncan-Jones's account of Butler's position. He reports Butler as taking the view that the notion of desert suggests ‘both that a man's conduct is bad and that to penalize him is useful, or both that his conduct is good and to praise or reward him is useful’ (Butler's Moral Philosophy [Harmondsworth: Penguin 1952, 140; my emphasis]).

19 John Rawls, ‘Two Concepts of Rules,’ reprinted in P. Foot, ed., Theories of Ethics (Oxford: Oxford University Press 1967), 144-70

20 A naturalistic interpretation of Hume's views on responsibility is defended in my ‘On the Naturalism of Hume's “Reconciling Project”.’ A more detailed account of Hume's description of the ‘mechanism of responsibility’ is provided in that paper.

21 In this way, while we may, on Hume's account, suspend or abandon the institution or practice of punishment, there is no question of us abandoning or suspending the whole framework of the moral sentiments.

22 See T,474: ‘These sentiments are so rooted in our constitution and temper .. .'; EPM,273: ‘But the sentiments of humanity … ‘; EPM,322: ‘Every one may employ terms in what sense he pleases … .'

23 Hume's views regarding the relevance of freedom to ascriptions of responsibility must be understood within this framework. When an agent acts according to the determination of his will (i.e. when he enjoys 1iberty of spontaneity’) it is possible for us to infer his character from his action. In these circumstances our moral sentiments are naturally aroused. By contrast, where and when an agent is subject to external force or violence (i.e. lacks liberty of spontaneity) or his actions are due to ‘indifference’ (i.e. are uncaused) it is impossible for us to infer his character from his action. It is, according to Hume, a matter of psychological fact that in these circumstances no moral sentiment would be aroused in us and, hence, the agent would not be regarded as responsible.

24 In light of these observations it may be argued that in so far as Hume's views on punishment have been misunderstood and misrepresented the root difficulty lies with confusion about his views on responsibility. More specifically, the general tendency to place Hume squarely in the empiricist-compatibilist tradition, alongside Hobbes and Schlick, has led commentators to overlook Hume's concern with the role of moral sentiment in this sphere. Oearly, as our analysis reveals, without an adequate interpretation of Hume's views on responsibility it is impossible to develop a proper understanding of his position on punishment.

25 See, e.g., C.A. Campbell's objections to ‘Schlick's essentially “forward looking” interpretation of punishment and responsibility': 1s “Freewill” a Pseudo-Problem?,’ reprinted in Berofsky, Free Will and Determinism, 112-18.

26 Smith, The Theory of Moral Sentiments, D.O. Raphael and A.L. Macfie, eds. (Oxford: Oxford University Press 1976); abbreviated as lMS. Also relevant is Smith, Lectures on Jurisprudence, R.L. Meek, D.O. Raphael, and P.G. Stem, eds. (Oxford: Oxford University Press 1978); abbreviated as LJ. Smith states (TMS, 77): ‘Let it be considered too, that the present inquiry is not concerning a matter of right, if I may say so, but concerning a matter of fact. We are not at present examining upon what principles a perfect being would approve of punishment of bad actions; but upon what principles so weak and imperfect a creature as man actually and in fact approves of it.'

27 In general, on Smith's account we sympathize with another person when we imaginatively place ourselves in his position and feel what he feels (i.e. we find that our sentiments ‘correspond’ with those of the person involved). In these circumstances we regard his sentiments as proper or appropriate to their object, and thus we approve of them.

28 Smith's views on this issue, it should be noted, are not entirely uniform. More specifically, Smith does say that on some occasions ‘we both punish and approve of punishment, merely from a view to the general interest of society, which, we imagine, cannot be otherwise secured’ (TMS, 90; my emphasis, and also LJ, 105). It is not entirely obvious, however, how we are supposed to interpret such exceptional cases; nor how Smith's remarks in this direction square with his analysis of the standard case. On this see Raphael, ‘Hume and Adam Smith', 96-7.

29 LJ, 104. Here Smith probably has Hume primarily in mind - although it is ‘Grotius and other writers’ to whom he actually refers. See the editors’ remarks at TMS, 87-8.

30 It may be noted that Smith's claims regarding the indirect utility of retributive practices are highly questionable. That is, Smith claims that the 1aw of retaliation’ will effectively ‘regulate’ punishment in such a way that it indirectly secures our utilitarian ends (deterrence, reform, etc.). At least two difficulties arise for this claim: (a) in many cases such practices may serve no such ends - they may even be counter-productive; (b) from a utilitarian perspective there may be more effective alternative strategies available to us.

31 Within the confines of this paper it is not possible to explore the significance which this general interpretation of Hume's theory has for the more specific or narrower remarks that Hume makes on the subject of punishment. Two passages, however, merit brief comment: (1) At T,608-9 Hume, notoriously, suggests that natural abilities and moral qualities are ‘on the same footing’ as regards their tendency to arouse moral sentiments in us. In other words, according to Hume people are held accountable or responsible for both. There is, nevertheless, he suggests, a distinction to be drawn between them in so far as our moral qualities, but not our natural abilities, ‘may be chang’ d by the motives of rewards and punishments, praise and blame.’ Oearly there is much in this account to question. However, Hume's position on this more specific issue is, I suggest, more intelligible when considered within the framework of his general theory of punishment. (2) At EPM,322 - developing points raised from the passage at T,608-9 (taking issue, in particular, with the suggestion that the distinction between voluntary and involuntary be made the foundation of a theory of morals) - Hume objects to those philosophers and/or divines who treat ‘all morals as on a like footing with civil laws, guarded by the sanctions or reward and punishment.’ The general point that Hume is concerned with here is that it is crucial that we distinguish between situations where our moral sentiments have been legitimately aroused and circumstances in which we are justified in distributing rewards and punishments on the basis of these sentiments. This fundamental point is firmly embedded in Hume's general theory of punishment. Indeed, it lies at the heart of his objections to the theological doctrine of future rewards and punishments.

32 ‘Freedom and Resentment,’ reprinted in G. Watson, ed., Free Will (Oxford: Oxford University Press 1982), 59-80. See also Strawson, Skepticism and Naturalism: Some Varieties, Ch.2. Unlike Hume (and Smith), Strawson speaks of ‘reactive attitudes and feelings.’ He notes, however, that it is ‘a pity that talk of the moral sentiments has fallen out of favour’ as that phrase would be a ‘good name’ for the network of attitudes and feelings he is concerned with (79). It should also be noted in this context that Strawson's paper is in large measure a critique or attack upon Schlickean, ‘forward-looking’ accounts of responsibility and punishment.

33 ‘Savage or civilized,’ Strawson claims, ‘we have some belief in the utility of the practices of condemnation and punishment .. .’ (78). For Strawson, however, this is not the central issue. Rather, the point which Strawson is concerned to establish is that ‘to speak in terms of social utility alone is to leave out something vital in our conception of these practices’ (viz. our moral sentiments).

34 Even though Strawson's paper has generated a great deal of comment and criticism very little has been said about the problem of punishment as it arises for him. Instead, commentators have focused their attention, almost exclusively, on the narrower issues of freedom and responsibility. This situation is particularly surprising in light of the fact that Smith (in whose footsteps Strawson clearly follows) devotes a great deal of his attention to the problem of punishment. It is fair to say, I believe, that this is indicative of the extent to which Strawson's naturalistic theory of responsibility has been cut off from its historical roots.

35 While Hume's views on the subject of punishment have received little attention over the years Smith's naturalistic account of positive retributivism has, by contrast, succeeded in attracting a number of followers. The most important of these have been Edward Westermark, Ethical Relativity (London: Kegan Paul 1932; reprinted Westport, Cf: Greenwood 1970); see esp. Ch. 3, and, more recently, John Mackie, ‘Morality and the Retributive Emotions,’ reprinted in Persons and Values (Oxford: Oxford University Press 1985), 206-19. The Smithian flavour of Mackie's general position is quite obvious - indeed, it can be directly traced to Smith via Westermarck. In order to preserve the coherence and direction of my own discussion, however, I will continue to focus attention on Smith's views.

36 Without going too deeply into the complexities of Hume's system two further points should be noted: (1) Some of Hume's remarks may be interpreted as suggesting that our moral sentiments are incapable of justification (e.g. T, 413-18, 455-70). Nevertheless, Hume makes it clear that our moral sentiments, like other passions, may be said to be indirectly reasonable or unreasonable, justified or unjustified, in so far as they arise out of true or false beliefs. (On the role of reason in morals see esp. EPM, Sect. 1.) (b) The beliefs which generate or give rise to the moral sentiments concern the supposed pleasurable or painful qualities of mind or character of the individual in question. Hume suggests that it is our expectations about the tendencies of an individual's character which shape or condition our moral sentiments. The grounds of such expectations are inevitably rooted in our past experience and knowledge of the action and behaviour of the individual. The crucial point, however, is that our moral sentiments are not to be justified in terms of their future consequences or utility. Whether or not such sentiments are of utility they will naturally arise in the relevant circumstances (see, e.g., EPM, 273 and 322). In this sense they are neither capable of, nor require, a consequentialist rationale.

37 Regarding the relationship between blame and punishment see, e.g., J.E.R. Squires, ‘Blame,’ reprinted in H. B. Acton, ed., The Philosophy of Punishment (London: MacMillan 1%9), 204-11; and Richard Wasserstrom, ‘Some Problems in the Definition and Justification of Punishment,’ in A.l. Goldman and J. Kim, eds., Values and Morals (Dortrecht & Boston: Reidel1978), 307-8. (Wasserstrom, I note in passing, raises several interesting objections against the sort of ‘mixed’ or teleological retributivist theory of punishment which I attribute to Hume. Suffice it to say, in this context, that I am not persuaded that his objections tell against Hume's position.)

38 Consider the following tale told about Zeno of Gtium, a necessitarian, by Diogenes Laertius: We are told that (Zeno] was once chastising a slave for stealing, and when the latter pleaded that it was his fate to steal, “Yes, and to be beaten too,” said Zeno’ (Diogenes Laertius, Lives of Eminent Philosophers, 2 vols. [London: 1925], Vol.2, 135). In many respects, Smith's general position seems to be akin to that of Zeno.

39 Smith may be understood to be suggesting - in a more utilitarian vein - that the point of punishment is that it gratifies our desire for retribution. This line is pursued by James Fitzjames Stephen, Liberty, Equality, Fraternity (Cambridge: Cambridge University Press 1%7 [1873]), 152. Stephen argues that common crimes are punished, ‘not only because they are dangerous to society, but also for the sake of gratifying the feeling of hatred - call it revenge, resentment or what you will … .'It follows, Stephen suggests, that the criminal law is ‘an emphatic assertion of the principle that the feeling of hatred and the desire for vengeance … are important elements of human nature which ought in such cases to be satisfied in a regular public and legal manner.'

40 Consider, for example, the theological implications of Smith's account, and how it contrasts with Hume's position. That is, granted the ontology presupposed in the doctrine of a future state, Smith cannot follow Hume in regarding the existence of Hell as a barbarism without point or purpose. Indeed, Smith takes quite the opposite view. See TMS, 91 and 163f. (Although Smith would grant, no doubt, that Hell involves the excessive punishment of the guilty.)

41 Strawson, ‘Freedom and Resentment', 78-9; Hart, Punishment and Responsibility, 8-11

42 This theme is pursued in some detail in Mackie, ‘Morality and the Retributive Emotions.'

43 Implausible as this strong view may be, several of Mackie's comments ('Morality and the Retributive Emotions,’ 216-19) lend themselves to this interpretation. In general, Mackie blurs the distinction between retaliating without a view to consequences and retaliating spontaneously or instinctively without making any ‘conscious choice’ on the matter.

44 So described, Smith's position may be (loosely) interpreted in terms of Harry Frankfurt's account of free will ('Freedom of the Will and the Concept of a Person,’ in Watson, ed., Free Will). That is, our desire for retaliation is such that whatever our reflective attitude may be with respect to these desires they will nevertheless be effective in leading to action. In other words, it is impossible for us to restructure our will in such a way that our resentment does not lead to retaliation or retribution. We cannot alter or change our will in this respect. This is consistent with the (obvious) fact that retaliatory actions are done of our own will and thus we enjoy freedom of action in respect of them.

45 Oearly, for Hume, someone who is willing to intentionally inflict pain and suffering for no purpose other than vengeance is simply vicious.

46 This issue - along with others closely related to it - is discussed and analysed in an interesting fashion by Jeffrie Murphy and Jean Hampton in Forgiveness and Mercy (Cambridge: Cambridge University Press 1988). It is worth noting that Hume's complex theory of punishment provides an interesting point of contrast and criticism from which to judge and evaluate Murphy's and Hampton's (divergent) contributions to the contemporary debate. It may also be argued, from another perspective, that the interpretation and analysis of Hume's discussion provided here reveals the extent to which Hume anticipated - and perhaps indirectly shaped - some of the more interesting recent developments in this sphere.

47 For helpful suggestions, remarks, and/or discussions concerning various aspects of this paper (early drafts of which date back to 1983) I would like to thank Cora Diamond, Ross Harrison, John Marshall, Neil MacCormick, Bernard Williams and, particularly, D.O. Raphael and the editor and referees for the CJP: David Copp, Annette Baier and Marcia Baron. I would also like to thank audiences at Virginia, Chicago and Stanford for their comments on a paper of related interest. The final draft of this paper was prepared while I held a Mellon Fellowship at Stanford University.