¹ As will become clear, I take these terms to describe the same project. A project which has taken contractualist form over the last 25 years.

² Page numbers in parentheses are all to Barry, Brian, Justice as Impartiality, Oxford, 1995Google Scholar.

³ Larmore, C., Patterns of Moral Complexity, Cambridge, 1987, p. 23CrossRefGoogle Scholar. ‘There are many viable conceptions of the good life that neither represent different versions of some single, homogeneous good nor fall into any discernible hierarchy’.

⁴ Scepticism à la Barry is, I take it, compatible with a monistic metaphysics of value.

⁵ Scanlon, T. M., ‘Contractualism and Utilitarianism’, Utilitarianism and Beyond, ed. Sen, A. and Williams, B., Cambridge, 1982, 103–28CrossRefGoogle Scholar.

⁶ Kymlicka, W., ‘Liberal Individualism and Liberal Neutrality’, Ethics, ic (1989), 883–905CrossRefGoogle Scholar. Kymlicka distinguishes ‘consequential neutrality’ which commits the state to ‘seek to help or hinder different life-plans to an equal degree’ from ‘justificatory neutrality’ which ‘allows that government action may help some ways of life more than others but denies that government should act in order to help some ways of life over others … the state does not justify its actions by reference to some public ranking of the intrinsic value of different ways of life, for there is no public ranking to refer to’, 884–6.

⁷ I am grateful to John Charvet for this point. As Charvet puts it, in an unpublished paper, ‘“to each according to his relative bargaining power” … is a completely impersonal principle. It applies to everyone quite impartially’.

⁸ He cannot, because he takes it as given that an agreement based on relative bar-gaining strength is ‘coerced’ (see, e.g., p. 50).

⁹ This latter type of argument resembles that offered by Kymlicka in his comments on Nozick's entitlement theory. Kymlicka argues that Nozick is most plausibly read as trying to show that the entitlement theory follows from an initial commitment to treating people with equal respect. Kymlicka goes on to show that not only is the theory internally flawed, but that when the consequences of its practice are taken into account it fails to capture what matters to us about the initial commitment, Kymlicka, W., Contemporary Political Philosophy: An Introduction, Oxford, 1990, pp. 118–25Google Scholar.

¹⁰ See Justice as Impartiality, ch. 3, and Barry's, Theories of Justice, Hemel Hemp-stead, 1989, pp. 196–201Google Scholar.

¹¹ This is to avoid, for the moment, difficulties that arise in Barry's application of Scanlon's formula (which concerns our understanding of morality) to the problem of justice.

¹² Barry, , ‘A Commitment to Impartiality: Some Comments on the Comments’, Political Studies, xliv (1996), 329Google Scholar.

¹³ Duff, R. A., ‘Penal Communications: Recent Work in the Philosophy of Punishment’, Crime and Justice: A Review of Research, xx (1996), 27Google Scholar. This account of the function of punishment can be found in Morris, H., ‘Persons and Punishment’, The Monist, lii (1968), 475–501CrossRefGoogle Scholar; Finnis, J., ‘The Restoration of Retribution’, Analysis, xxxii (1972), 131–5CrossRefGoogle Scholar; Murphy, J. G., ‘Marxism and Retribution’, Philosophy and Public Affairs, ii (1973), 217–43Google Scholar; von Hirsch, A., Doing Justice: the Choice of Punishments, New York, 1976Google Scholar; Sadurski, W., ‘Distributive Justice and the Theory of Punishment’, Oxford Journal of Legal Studies, v (1985), 47–59CrossRefGoogle Scholar, and ‘Theory of Punishment, Social Justice, and Liberal Neutrality’, Law and Philosophy, vii (1989), 351–73Google Scholar; Sher, G., Desert, Princeton, 1987, ch. 5Google Scholar; Dagger, R., ‘Playing Fair with Punishment’, Ethics, ciii (1993), 726–52Google Scholar. For subsequent retractions see Murphy, J. G., ‘Retributivism, Moral Education, and the Liberal State’, Criminal Justice Ethics, iv (1985), 3–11CrossRefGoogle Scholar; von Hirsch, A., ‘Proportionality in the Philosophy of Punishment: From “Why Punish?” to “How Much?”’, Criminal Law Forum, i (1990), 259–90, esp. 264–9CrossRefGoogle Scholar, and Censure and Sanctions, Oxford, 1993, pp. 7–8Google Scholar. For a good summary of many points of criticism see Duff, R. A., Trials and Punishments, Cambridge, 1986, ch. 8Google Scholar.

¹⁴ Barry, , ‘A Commitment to Impartiality’, 329Google Scholar.

¹⁵ Cf. Aristotle, , The Nicomachean Ethics, Oxford, 1980Google Scholar, Bk. V, i–ii (esp. 1129a25–1129b6 and 1130b10).

¹⁶ Barry, , ‘A Commitment to Impartiality’, 329Google Scholar. They are also, therefore, injustices because doing harm is contrary to the demands of justice. This is the position which Richard Dagger also endorses in his defence of fair play theory. Dagger, like Barry, believes that an account of wrong as unfairness cannot accommodate certain wrongs. ‘All crimes’, he writes, ‘are in some sense crimes of unfairness. They may be more than crimes of unfairness, as rape, robbery and murder surely are’. Dagger, 479. Emphasis in original.

¹⁷ Barry, ‘A Commitment to Impartiality’, § 3.

¹⁸ Ibid., 332.

¹⁹ Barry correctly notes the need for this proviso; some people regard suffering as part of their conception of the good (Barry notes that ‘Mother Teresa thinks that suffering is a great gift from God’, ibid., 332). Ian Shapiro has pointed out to me that people committed to a conception of the good which we might characterize as a commitment to ‘martial virtue’ may also be such that they not only believe that suffering is a great good but that inflicting suffering is also of value. The importance of this point is made clear below in note 26 and the text surrounding it.

²⁰ Barry, , ‘A Commitment to Impartiality’, 332Google Scholar.

²¹ I have come to the conclusion after many conversations with colleagues and students that liberals would do well to adopt the language of thickness and thinness and drop that of neutrality, fairness, or impartiality. The reason for this is that, despite the valiant attempts of Kymlicka and others to distinguish justificatory and consequential neutrality, it still seems natural to allege that liberalism is not really neutral (because it treats Nazis, child-molesters etc. poorly). Likewise claims that Rawls and other ‘liberal impartialists’ try to build on neutral foundations are misleading. Adopting the thick/thin terminology has the great advantage that it makes it more difficult to misunderstand the aspirations of liberal theory. The sentence ‘But is this really thin to Nazis?’ comes, in my experience, far less quickly to the mouths of students than the same sentence (or variants of it) using the language of fairness, neutrality or impartiality. Similarly, asking ‘how thin are Rawls's (or Barry's) assumptions?’ I have found attracts far better responses – from both students and colleagues – than asking the same question using the idea of neutrality.

²² Thus, I am using ‘contractualist’ to mean an approach to the questions of political philosophy that may be contrasted with ‘realist’ and ‘relativist’. This is why I think that there is an intimate link between ‘liberal impartialism’ and ‘liberal contractualism’. Hobbes and Locke are not contractualist (in my sense). They both use the contract to explain obligation but both build on independently given accounts of morality. Any reader who thinks I am not entitled to the rights of Humpty Dumpty (to give meanings to words just as I wish) may simply replace ‘contractualist’ with ‘constructivist’ in my paper. I choose not to do so because it tends to result in (even more) inelegant sentences.

²³ See, e.g., the essays in Part I, ‘The Original Position’ in Reading Rawls: Critical Studies on Rawls' A Theory of Justice, ed. Daniels, N., New York, 1975Google Scholar; and Barry's, own The Liberal Theory of Justice, Oxford, 1973Google Scholar.

²⁴ That is putting aside questions of self-defence and punishment.

²⁵ I should add that this ‘Aztec’ is created to make a point, it is of no relevance whether historical Aztecs did believe in a religion that commanded human sacrifice.

²⁶ Barry has, on occasion, expressed a distaste for the use of fanciful examples in moral philosophy. However, the Aztec seems to me no more outrageous than the orthodox Thomist whose claims Barry is prepared to consider seriously. However, the point is a theoretical one and it holds independently of how we class the Aztec, Thomist, pursuer of ‘martial virtue’, or Nozickean. The point is that any account of the good that includes a proposition of the form ‘it is morally permitted/obligatory to harm another’ (other than in exceptional circumstances), is, on the Barry account of morality, false (for it is wrong to do such a thing). Acting upon such beliefs is doing something immoral. Both of these claims stand independently from the contractualist procedure and both need independent justification.

²⁷ It often seems to me that Barry writes in a way that suggests that he believes the basic premise of liberalism (which he takes to be fundamental equality) is widely shared by (at least some) people in all (or almost all) cultures and, therefore, working out the consequences of this shared understanding is a proper task of political philosophy.