¹ Williams, Bernard, ‘Internal and External Reasons’, repr. Moral Luck, Cambridge, 1981CrossRefGoogle Scholar. ‘Internal Reasons and the Obscurity of Blame’, repr. Making Sense of Humanity, Cambridge, 1995Google Scholar.

² A subsidiary aim will be to detach it from ‘Humeanism’ of the kind defended, for example, by Smith, Michael in The Moral Problem, Oxford, 1994Google Scholar. See esp. n. 8 below The question of the relation between Williams's position and that of Hume is one of several insightfully discussed in an unpublished paper by John Skorupski, ‘Internal Reasons and the Scope of Blame’, unpub. ms. This relation is discussed at 8–14.

³ Thus, ‘internal’ and ‘external’ are not predicates of reasons, per se, but predicates of statements about reasons, and are hence in the formal, not the material mode. This is a qualification I will ignore in the sequel, speaking for convenience of ‘internal reasons’ and ‘external reasons’.

⁴ Williams, , ‘Internal Reasons and the Obscurity of Blame’, p. 35Google Scholar.

⁵ It must be borne in mind that this set can contract as well as expand in the course of revision; by vividly imagining the way you intend to satisfy your desire for an enjoyable evening by going to the theatre, you may lose your desire to go to the theatre. So ‘revision’ of S into S* is not necessarily ‘expansion’.

⁶ ‘It has been generally recognized that the concept of a reason for action stands at the point of intersection, so as to speak, between the theory of the explanation of action and the theory of the justification of action.’Woods, Michael, ‘Reasons for Action and Desires’, Proceedings of the Aristotelian Society, suppl. vol., xlvi (1972), 189CrossRefGoogle Scholar.

⁷ Williams, , ‘Internal Reasons’, p. 103Google Scholar.

⁸ As Skorupski points out at 4 f., the issue is the insertion of such principles into an agent's S. If they are included in the idea of a sound deliberative route by definition, for example if they are either a priori or necessary, then they are going to be reached from anyone's S as they will be part of any route from such an S. Williams explicitly argues against this alternative interpretation of the idea of a sound deliberative route at p. 37 of ‘Internal Reasons and the Obscurity of Blame’.

⁹ By contrast, Thomas Scanlon, in his interesting discussion of Williams's argument, takes Williams's thesis to be essentially a substantive normative claim, not an analysis of the idea of a practical reason. See What We Owe To Each Other, Cambridge, Mass., 1998, ‘Appendix’, p. 365Google Scholar.

¹⁰ Gibbard, Allan, Wise Choices, Apt Feelings, Oxford, 1990, p. 161n6Google Scholar.

¹¹ Smith, p. 95.

¹² McDowell, John, ‘Might There Be External Reasons?’, in World, Mind and Ethics: Essays on the Ethical Philosophy of Bernard Williams, ed. Altham, J. E. J. and Harrison, T. R., Cambridge, 1995Google Scholar.

¹³ Korsgaard, Christine, ‘Scepticism about Practical Reason’, Journal of Philosophy, lxxxiii (1986)Google Scholar.

¹⁴ Harrison, T. R., ‘Transcendental Arguments and Idealism’, Idealisms: Past and Present, ed. Vesey, G., Cambridge, 1982Google Scholar. This interpretative line was initially developed by Prauss and consolidated in Anglophone Kant scholarship by H. Allison. See his Kant's Transcendental Idealism, New Haven, Conn., 1983Google Scholar.

¹⁵ A project developed in Korsgaard, , The Sources ofNormativity, Cambridge, 1996CrossRefGoogle Scholar.

¹⁶ Interestingly, Korsgaard herself changed her position on these questions significantly; see ‘The Normativity of Instrumental Reason’, in Ethics and Practical Reason, ed. Cullity, G. and Gaut, B., Oxford, 1997Google Scholar. Korsgaard now places some distance between her own considered view and a view she now calls ‘pure rationalism’, writing on p. 240 that: ‘According to dogmatic rationalism … there are facts, which exist independently of the person's mind, about what there is reason to do … The difficulty with this account in a way exists right on its surface, for the account invites the question why it is necessary to act in accordance with those reasons, and so seems to leave us in need of a reason to be rational … we must still explain why a person finds it necessary to act on those normative facts, or what it is about her that makes them normative for her. We must explain how these reasons get a grip on the agent’. Korsgaard traces this defect to dogmatic rationalism's project of deriving reasons for an agent from the reasons of an ideally rational agent. Korsgaard now seems to express the requirement that a source in pure practical reason is insufficient because it cannot explain how the reasons statements it generates for an individual agent say something distinctive about that agent. This looks like the adoption of the internal reasons requirement, not its rejection. For this reason also, Korsgaard now no longer views it as acceptable simply to stipulate that rational agents are such as to be motivated by principles of reason. This seems to be a tactical withdrawal from the argument directed against the internal reasons theory in ‘Skepticism about Practical Reason’.

¹⁷ McDowell does not make the error of treating Williams as an instrumentalist about practical reasoning. He recognizes that the conception of practical reasoning Williams deploys reflects the influence on both Williams and Wiggins of Aurel Kolnai. For the very unpromising line of argument that Williams is covertly an instrumentalist about practical reasoning, see Milgram, Elijah, ‘Williams's Argument Against External Reasons’, Nous, xxx (1996)Google Scholar. Williams's n. on p. 104 of ‘Internal and External Reasons’ cites the anti-instrumentalist view of practical reasoning presented by Kolnai, Aurel in ‘Deliberation is of Ends’, Ethics, Value and Reality, London and Indianapolis, 1973Google Scholar.

¹⁸ McDowell, , Mind and World, Cambridge, Mass., 1994Google Scholar, esp. lecture 4.

¹⁹ Given McDowell's sympathies both with Williams's Davidsonian approach to the mental and to his normative conception of practical reasoning, this charge would have to rest on the importation into Williams's argument of a stereotypically ‘Humean’ position which I would argue has little bearing on the internal reasons argument.

²⁰ Williams, , ‘Reply’, Altham, and Harrison, (ed.), p. 190Google Scholar.

²¹ Price, A. W., ‘Reasons and Desires’, unpub. ms. Presented at a University of London seminar on practical reasoning, 1995Google Scholar.

²² A point that I owe to Brad Hooker.

²³ A corollary is that Williams is sceptical of the external reason theorist's accusation that a person who does not ‘see’ the external reason true of them is thereby ‘irrational’.

²⁴ Onora O'Neill, ‘Constructivisma in Ethics’, repr. Constructions of Reason, Cambridge, 1989, esp. pp. 209 f.

²⁵ For an interpretation of this kind, see Freeman, Samuel, ‘Contractualism, Moral Motivation and Practical Reason’, Journal Of Philosophy, lxxxviii (1991)Google Scholar. For a contrary-view of contractualism as basically Humean and externalist, see Wallace, Jay, ‘How to Argue about Practical Reason’, Mind, xcix (1990)Google Scholar.

²⁶ I take the distinction from Craig, Edward, Knowledge and the State of Nature, Oxford, 1990, pp. 33 fGoogle Scholar.

²⁷ Rawls, John, Political Liberalism, New York, 1993, pp. 24–8Google Scholar.

²⁸ Jeremy Waldron invokes the idea of placing appeals before an individual's ‘tribunal of reason’ in his ‘Theoretical Foundations of Liberalism’, Philosophical Quarterly, xxxvii (1987)Google Scholar. The term ‘reasonable rejection’ is, of course, Scanlon's: the first presentation of this form of contractualism was presented in ‘Utilitarianism or Contractualism’, Utilitarianism and Beyond, ed. Sen, A. and Williams, B., Cambridge, 1982CrossRefGoogle Scholar.

²⁹ I take the useful phrase ‘perspectival ascent’ from Sacks, Mark, The World We Found, La Salle, Illinois, 1989, p. 95Google Scholar.

³⁰ Dancy, Jonathan, Moral Reasons, Oxford, 1993, pp. 147–53Google Scholar.

³¹ Williams himself seems to have been led to reject an impartialist account of moral reasons because he associated it, quite plausibly, with a Cartesian model of objectincation in Nagel, Thomas, The Possibility of Altruism, Oxford, 1970Google Scholar. In that text moral reasons are ‘objective’, achieved by a process akin to ‘perspectival ascent’ and can be interpreted as both ‘external’ and impartial. Nagel replied that he did not conceive of the objective standpoint as capable of generating motivations in its own right but merely as a heuristic device and has now clearly adopted the more defensible position, in The View From Nowhere, Oxford, 1986Google Scholar, ch. 9.

³² Thus I do not want to rehabilitate the unpromising idea of an externalist ‘desire to be moral’, recently defended by Svavarsdottir, Sigrun in ‘Moral Cognitivism and Motivation’, Philosophical Review, cviii (1999)Google Scholar. Any such idea is vulnerable to Michael Smith's objection that while we may desire to do the right thing, we do not desire to do so under that description.

³³ I cannot enter here into the intriguing current debate over a priority, well represented by the essays in New Essays On The A Priori, ed. Boghossian, P. and Peacocke, C. A. B., Oxford, 2000CrossRefGoogle Scholar. But in that anthology Michael Friedman extends his defence of a distinction indebted to Reichenbach's as an account of the a priority of mathematics in ‘Transcendental Philosophy and A Priori Knowledge: A Neo-Kantian Perspective’. In the same volume Phillip Kitcher accepts that this is a viable notion of the a priori, but makes two further points that seem to me to be true and interesting: that this notion cannot do all the work that the tradition demanded of a stronger conception of a priority (one that Kitcher favours) and in particular it cannot establish an authoritative, experience independent framework for future enquiry (pp. 75–7). Relatedly, Kitcher's second point is that the assumption doing the work here is of the truth of historicism, pp. 90–1. While I recognize that Kitcher is pointing to controversial features of mathematical knowledge, it seems to me that in the case of moral reasons recognition of these two features makes the account more, not less, plausible.

³⁴ Here is another analogy that I owe to Adrian Moore. What role does the norm of consistency play when you investigate the structure of the physical world? A pre-supposed role: you do not explicitly adopt a principle of consistency, but if you accepted inconsistent representations of the world, investigating what there is, is not the project you could intelligibly be engaged in. I suggest that the disposition to contractualist impartiality is presupposed in a similar way.

³⁵ Thus I am in no way suggesting that the constraint of impartiality determines any specific content for moral reasons. I mean two more precise things by this. I do not believe that moral reasons are essentially ‘public’ and hence altruistic in content. Nor do I believe that impartial reasons can be explained as both agent-neutral and universal, as I do not believe that the agent-relative/agent-neutral distinction is a tenable one. I argue for both of these claims in ‘The Scope of the Agent-relative’, unpubl. ms.

³⁶ Harrison, esp. pp. 215–8.

³⁷ Larmore, Charles, Patterns of Moral Complexity, Cambridge, 1989, ch. 4, pp. 85–90Google Scholar; The Morals of Modernity, Cambridge, 1996, pp. 35–40Google Scholar.

³⁸ This is the neo-Kantian strategy replicated in the case of his theoretical philosophy, for example by Reichenbach, in The Theory of Relativity and A Priori Knowledge, Berkeley, CA, 1965Google Scholar.

³⁹ Larmore, , Patterns, pp. 87–9Google Scholar.

⁴⁰ Korsgaard, in ‘Scepticism about Practical Reason’, esp. at 32, takes the moral of Nagel's, ThomasThe Possibility of Altruism, Oxford, 1970Google Scholar, to be that moral philosophy can teach us psychology, which should come as news to psychologists. For the alternative proposal of interpreting Nagel as having isolated the role of structural motivation in moral psychology, see Scheffler, Samuel, Human Morality, Oxford, 1992Google Scholar, ch. 5. I agree with Scheffler that this is one of the most promising lines of argument in The Possibility of Altruism and the current paper can be seen as a development of one of Nagel's lines of thought.

⁴¹ Scheffler, pp. 86–8.

⁴² I refer to a ‘higher order disposition’ as it is plausible to maintain that particular dispositions of character, such as virtues, have distinctive patterns of motivation associated with them. The disposition I am concerned with has a regulative status towards such particular motivations. Williams himself has a similar story to tell about how, given the truth of the internal reasons thesis, blame actually operates. His account invokes a ‘proleptic mechanism’ whereby one appeals in moral criticism to ‘the desire to be respected by people whom, in turn, one respects’. This invokes a reason proleptically as it makes it true that a person has a reason motivated indirectly by this desire. The difference between his account and the one I have offered here is that the disposition I cite is more minimal: The motivation to want to be respected by others is a substantial ethical motivation, whereas the desire to advance reasons to others without reasonable rejection goes less deeply into ethical motivations. But merely because my notion is the more minimal, it does not, for that reason, allow one simply to presuppose its universality of scope and in that I agree with Williams. (By way of contrast, Scanlon, in What We Owe To Each Other, takes the universality of moral reasons as a premiss from which one can argue against the idea that moral reasons are internal.)

⁴³ Freeman makes a similar suggestion at 298 f. concerning the role of the central disposition contractualism attributes to agents, but takes his proposal to vindicate the Kantian origins of the theory as it demonstrates that such motivations have a ‘formal basis in practical reasoning’. I do not see that the sense of ‘formal’ Freeman explains has this consequence; I agree with Scheffler that we are dealing with structures which are relativized a priori and compatible with a broadly naturalistic account of motivation. I leave it open whether this position is better traced back to Hume or to the philosopher that Grice once called ‘Kantotle’.

⁴⁴ The analogy is described in n. 34.

⁴⁵ I should add that as I regard my view not as supplying particular motivations but rather shaping the architecture of deliberation, it is distinct from neo-Kantian attempts to make impartial reasons a class of reasons, but a class with a distinctive ‘failsafe’ role in moral deliberation. I have in mind here Herman's, Barbara influential position in ‘On the Value of Acting from the Motive of Duty’, repr. The Practice of Moral Judgement, Cambridge, Mass., 1993Google Scholar.