¹ Sidgwick, H., The Methods of Ethics, 7th edn., London, 1907, pp. 496–509Google Scholar.

² If there is a dualism of practical reason, then neither the egoist nor the ethical theory is rationally required. The focus of this paper, however, is on one element of the dualism, namely that the ethical theory is no better justified than the egoist theory. The other element of the dualism, that the egoist theory is not rationally required, is not discussed. Nevertheless, the arguments put forward to show that the ethical theory is not rationally required do not suggest that the egoist theory is better justified than the ethical theory; they merely show that the ethical theory is no better justified than egoism.

³ It is irrational to act contrary to reasons that you yourself acknowledge as reasons; it is not irrational to act contrary to reasons that you do not so acknowledge. An agent who does not act for moral reasons but also does not acknowledge moral reasons as genuine is not irrational. But if ethics is rationally required, moral reasons apply to her, and since she is not acting for the strongest reasons that apply to her, she is unreasonable.

⁴ An egoist will not act differently from a utilitarian if, for example, there are strong social sanctions against acting as an egoist. But under any plausible set of social arrangements, it is likely that at least sometimes an egoist can profitably act as a ‘free-rider’, benefiting from those arrangements whilst avoiding their costs. In some circumstances, an egoist will free-ride, but a utilitarian will not.

⁵ Agent-neutral reasons give different agents the same aim: each agent has reason to promote everyone's good. There are different kinds of agent-relative reasons (see Nagel, T., The Possibility of Altruism, 2nd edn., Princeton, 1978, pp. 90–8Google Scholar; Scheffler, S., The Rejection ofConsequentialism, 2nd edn., Oxford, 1994, pp. 2–6CrossRefGoogle Scholar; McNaughton, D. and Rawling, P., ‘Value and Agent-Relative Reasons’, Utilitas, vii (1995)Google Scholar, and Skorupski, J., ‘Agent-Neutrality, Consequentialism, Utilitarianism…A Terminological Note’, Utilitas, vii (1995)Google Scholar. In this paper all references to agent-relative reasons are references to egoist reasons: each agent has reason to promote his own good.

⁶ Nagel, pp. 116–24. Nagel claims that the inability to extend to others the kind of reasons judgements one makes in the first person is a form of solipsism (pp. 123 f.).

⁷ Sturgeon, N., ‘Altruism, Solipsism and the Objectivity of Reasons’, Philosophical Review, lxxxiii (1974)Google Scholar; Darwall, S. L., Impartial Reason, Ithaca, 1983, pp. 117–29Google Scholar; Nagel, T., The View from Nowhere, Oxford, 1986, p. 167Google Scholar.

⁸ An alternative egoist theory would claim that my own happiness is the only value: it grounds reasons for action for everyone; no one else 's happiness grounds any reasons at all. This version of egoism does posit an arbitrary difference between my happiness and everyone else's happiness, and so it is very implausible.

⁹ Egoism might nevertheless seem to distinguish arbitrarily my happiness and your happiness as values: from my perspective, my own happiness contributes to the value of the world, but yours does not. Egoism should not therefore make any claims about the value of the world, but just about what reasons for action there are.

¹⁰ Parfit, D., Reasons and Persons, Oxford, 1984Google Scholar, and Brink, D. O., ‘Sidgwick and the Rationale for Rational Egoism’, ed. Schultz, B., Essays on Henry Sidgwick, Cambridge, 1992Google Scholar, consider whether metaphysical questions concerning the nature of persons could cast doubt on egoism. I have no room to discuss this issue here.

¹¹ A similar theory is defended in Scheffler.

¹² For example, rather than thinking that it is both reasonable to promote my own happiness and to promote everyone's happiness, sometimes I may think that it is reasonable to promote only my own happiness, and sometimes think that it is reasonable to promoteonly everyone's happiness. That is, sometimes I may think that the correct perspective on practical reasoning can only be individualistic and personal I must decide what I have most reason to do - and sometimes I may think that the correct perspective can only be the impersonal ‘point of view of the universe’. These two intuitions, the two perspectives on reasoning, cannot be captured together by any theory, including utilitarianism with an egoist bias.

¹³ Sidgwick thought that an enlightened egoist would follow common sense morality for the most part (Sidgwick, pp. 151–75). Gauthier argues that it is rational for an egoist to become a ‘constrained maximizer’, that is, ‘to comply with mutually advantageous moral constraints, provided that he expects similar compliance from others’( Gauthier, D., Morals by Agreement, Oxford, 1986, p. 15Google Scholar).

¹⁴ Sinnott-Armstrong, W., ‘Moral Skepticism and Justification’, Moral Knowledge? New Readings in Moral Epistemology, ed. Sinnott-Armstrong, W. and Timmons, M., Oxford, 1996, pp. 31–6Google Scholar; G. Sayre-McCord, ‘Coherentist Epistemology and Moral Theory’, ibid., pp. 137–90. In addition, the common-sense beliefs may be revised in the light of the theory accepted (Daniels, N., ‘Wide Reflective Equilibrium and Theory Acceptance in Ethics’, Journal of Philosophy, lxxvi (1979)Google Scholar; Rawls, J., A Theory of Justice, Cambridge, Mass., 1971, pp. 46–53)Google Scholar.

¹⁵ Sayre-McCord, p. 177; Brink, D. O., Moral Realism and the Foundation of Ethics, Cambridge, 1989, pp. 103 f., 142 fCrossRefGoogle Scholar.

¹⁶ Sidgwick thought that there were some self-evident axioms (concerning ultimate reasons for action) that supported utilitarianism but were denied by egoism, for example, that it is rational to weigh equally equal pleasures that occur in different people (Sidgwick, p. 382). The egoist denies this axiom, for she gives no weight to pleasures that occur in other people (except in so far as they affect her own pleasure). If the axiom were genuinely self-evident, the egoist should accept it once she had understood it properly and reflected on it, but it is obvious that an egoist can and will reject the axiom. A utilitarian might reply that an egoist could not both understand the axiom and reject it without irrationality, that is, an egoist could not reasonably reject the axiom. But this would obviously assume what was supposed to be proven, that utilitarianism is rationally required.

¹⁷ Sinnott-Armstrong, pp. 25–31.

¹⁸ This way of expressing the dualism is similar to O'Neill, O., ‘Sidgwick on Practical Reason’, Proceedings of the British Academy, cix (2001), 84 f., 88 fGoogle Scholar.

¹⁹ As an egoist about friends, I may be interested in others' welfare for their own sake, but only when those people are my own friends and when our friendship requires such an interest.

²⁰ I would like to thank Onora O'Neill, Jimmy Altham, and Brad Hooker for helpful discussion of these issues.