¹ See, for instance, Dworkin, Ronald, ‘Liberalism’, A Matter of Principle (Cambridge, 1985)Google Scholar.

² See, for instance, Rawls, John, A Theory of Justice (Cambridge, 1971)Google Scholar.

³ Robert Goodin makes a similar point. See Goodin, Robert, ‘The Political Theories of Choice and Dignity’, American Philosophical Quarterly 18 (1981), pp. 91–100Google Scholar.

⁴ Kant, Immanuel, Groundwork for a Metaphysic of Morals (Cambridge, 1998)Google Scholar.

⁵ Hill, Thomas, ‘Humanity as an End in Itself’, Ethics 91.1 (1980), p. 97Google Scholar.

⁶ Goodin, ‘Choice and Dignity’, p. 95; emphasis in original.

⁷ I leave it open whether everyone, and not just the state, has an obligation to respect a person's preferences. While I think that we all have a moral reason to respect each other as persons, and that we therefore have a moral reason to respect each other's autonomous preferences, these reasons do not straightforwardly generate moral obligations. I assume, however, that the state has an obligation to respect its citizens as persons and not just a moral reason to do so. My aim in this article is to show that this obligation creates a further obligation to respect people's autonomous preferences. If, however, all moral reasons entail moral obligations, or if all of us have an obligation and not just a moral reason to respect each other as persons, then my conclusion would apply to all persons and not just to states. For more on the connection between moral reasons and moral obligations, see Shelley Kagan, ‘Defending Options’, Ethics 104 (1994), pp. 333–51. Kagan argues that if there is a moral reason to do X, and it is not outweighed by other moral reasons, then doing X is morally required.

⁸ While it is doubtful that a preference could be constitutive of one's personal identity on bodily continuity views, it seems that it could be on psychological continuity views. If that which makes you the same person that you were five minutes ago is a matter of psychological continuity, then an especially deep preference could be constitutive of your personal identity. For a defense of the bodily continuity view see, for instance, Van Inwagen, Peter, Material Beings (Ithaca, 1990)Google Scholar. For more on the psychological continuity view see, for instance, Parfit, Derek, Reasons and Persons (Oxford, 1984)Google Scholar. For the sake of the discussion I will assume that preferences can be constitutive of one's personal identity, but I will show that this is irrelevant for determining which preferences are worthy of respect. If, however, preferences cannot be constitutive of one's personal identity, then an important competitor to my view is eliminated.

⁹ Modern conceptions have parted ways with the related though distinct notion of moral autonomy developed primarily by Kant. On the former, being autonomous essentially amounts to living according to one's freely chosen conception of a worthwhile life. The Kantian conception, on the other hand, is essentially a view about the nature of morality.

¹⁰ Many who write on autonomy accept some version of the capacity view. See, for instance, Richards, David, ‘Rights and Autonomy’, Ethics 92 (1981), pp. 3–20Google Scholar; Dworkin, Gerald, The Theory and Practice of Autonomy (New York, 1988)Google Scholar; Frankfurt, Harry, ‘Freedom of the Will and the Concept of a Person’, Journal of Philosophy 68 (1971), pp. 829–39CrossRefGoogle Scholar.

¹¹ I describe these capacities as ‘latent’ because this view's proponents consider the mere possession of these capacities as sufficient for autonomy whether or not they are ever exercised.

¹² Virtually everyone writing on autonomy appeals to some such metaphorical description. Among those who use the ‘author of your life’ metaphor are Joseph Raz, Thomas Nagel, Richard Lindley, and, to some extent Joel Feinberg. Among those who use the idea of being self-governing are Gerald Dworkin, Harry Frankfurt, Robert Young, and Lawrence Haworth. See Raz, Joseph, The Morality of Freedom (New York, 1986)Google Scholar; Nagel, Thomas, The View From Nowhere (New York, 1986), p. 114Google Scholar; Lindley, R., Autonomy (London, 1986)Google Scholar; Feinberg, Joel, Harm to Self (New York, 1986)Google Scholar; Dworkin, The Theory and Practice of Autonomy; Frankfurt, ‘Freedom of the Will’; Young, R., Personal Autonomy: Beyond Negative and Positive Liberty (New York, 1986)Google Scholar; Haworth, Lawrence, Autonomy: An Essay in Philosophical Psychology and Ethics (New Haven, 1986)Google Scholar.

¹³ See, for example, Christman, John, ‘Autonomy and Personal History’, Canadian Journal of Philosophy (1991), pp. 1–24Google Scholar; Elster, Jon, Sour Grapes: Studies in the Subversion of Rationality (New York, 1983)Google Scholar.

¹⁴ Harry Frankfurt, Laura Ekstrom, and Gerald Dworkin are among those who hold structural theories. See Frankfurt, ‘Freedom of the Will’; Ekstrom, Laura, ‘A Coherence Theory of Autonomy’, Philosophy and Phenomenological Research 53 (1993), pp. 599–616CrossRefGoogle Scholar; Dworkin, The Theory and Practice of Autonomy.

¹⁵ Recently, George Sher and Sigurdur Kristinsson have proposed substantive theories of autonomy where a necessary condition of being autonomous is being responsive to reasons. See Kristinsson, Sigurdur, ‘The Limits of Neutrality: Toward a Weakly Substantive Account of Autonomy’, Canadian Journal of Philosophy 30.2 (2000), p. 257Google Scholar. George Sher, ‘Liberal Neutrality and the Value of Autonomy’, Social Philosophy and Policy (1995), pp. 136–59.

¹⁶ Indeed, I think these views are more plausibly understood as accounts of autonomous preferences than of autonomous persons. To see this, note that, in their current form, none of the achievement conceptions can account for the non-autonomy of the prisoner who is involuntarily chained to a dungeon wall. After all, he can have motivations that developed in the right way, that presently cohere with other elements of his character and values, and he can be fully responsive to reasons. Thus, as accounts of autonomous persons, these views are incompatible with paradigm cases of non-autonomy. As accounts of autonomous preferences, however, they each have some merit.

¹⁷ Clearly, this preference violates the historical condition. In addition, we can suppose that it does not resonate with other elements of Jim's character, and that Jim has no powerful reasons for preferring detective fiction to biographies. To be sure, one might ask whether Jim's desire to read detective fiction provides him with a reason to do so. Perhaps it does, but on the reason-oriented view, desires cannot straightforwardly give rise to autonomy-conferring reasons because otherwise all desire-based actions would be autonomous.

¹⁸ This preference is a good candidate for autonomy on all three views. Nothing in its history suggests a ‘wrong’ origin, it seems to cohere with other elements of her character and values, and we can assume that Mary has good reasons for choosing this career.

¹⁹ I selected a preference for detective fiction and for painting because neither seems to have intrinsic worth or, for that matter, any moral dimension at all. If, however, the reader thinks there is something intrinsically good about painting and not about reading detective fiction that may distort our judgment, she can substitute a more neutral preference for it, or just assume that Mary and Jim have the same preference though they acquired it differently. This shouldn't affect the intuition.

²⁰ It is unclear whether mere preference satisfaction is a component of well-being. It is not on mental state theories of well-being, but it may be on desire-satisfaction theories and objective list theories. In any case, nothing here hinges on this question. At this stage, I am merely attempting to generate intuitions that the state has a reason to promote the satisfaction of autonomous preferences. As will become clear, I do not think that the state's obligation to help persons satisfy their autonomous preferences is grounded in considerations of well-being. Note, however, that my position is incompatible with the view that mere preference satisfaction is not an element of well-being and that only considerations of well-being provide reasons for state action.

²¹ Here I am denying what some refer to as the Slogan, or the idea that one situation cannot be better (or worse) than another if there is no one for whom it is better (or worse). I am claiming that one world can be better than another even if it is not better for anyone in that world. For a defense of the Slogan, see, for example, Broome, John, Weighing Goods (Cambridge, 1991)Google Scholar; Doran, Brett, ‘Reconsidering the Levelling-Down Objection against Egalitarianism’, Utilitas, 13.1 (2001), pp. 65–85Google Scholar. I, however, find the arguments for rejecting the Slogan entirely persuasive. For this, see Temkin, Larry, ‘Harmful Goods, Harmless Bads’, Value, Welfare, and Morality, ed. Frey, R. G. and Morris, C. (Cambridge 1993), pp. 290–324Google Scholar. See also Crisp, Roger, ‘Equality, Priority, and Compassion’, Ethics 113 (2003), pp. 745–63CrossRefGoogle Scholar.

²² See, for example, Barkow, Jerome, Cosmides, Leda, and Tooby, John, The Adapted Mind (New York, 1992)Google Scholar; Pinker, Stephen, How the Mind Works (New York, 1997)Google Scholar.

²³ The analogy between object and preference ownership is not perfect. In order to be a preference's owner, you must make it your own. Objects, however, can become yours accidentally, such as when you are bequeathed a large sum of money by an unknown relative.

²⁴ To be sure, these kinds of preferences are not necessarily non-autonomous because it is possible to exercise some control over them.

²⁵ Perhaps one could autonomously choose such a life, but this choice cannot be passive in the above sense.

²⁶ This case was suggested to me by Larry Temkin.

²⁷ While I do not offer an account of the distinction between positive and negative duties, on any reasonable account a duty to help someone satisfy their autonomous preferences must be classified as a positive one.

²⁸ See, for example, Dworkin, Ronald, A Matter of Principle (Cambridge, 1985)Google Scholar; Larmore, Charles, Patterns of Moral Complexity (Cambridge, 1987), pp. 59–66Google Scholar; Rawls, John, ‘The Priority of Right and the Ideas of the Good’, Philosophy and Public Affairs 17 (1988), pp. 251–76Google Scholar; Mason, Andrew, ‘Autonomy, Liberalism, and State Neutrality’, The Philosophical Quarterly 40 (1990), pp. 433–52CrossRefGoogle Scholar. While my arguments call this view into question, they are not necessarily incompatible with it because I focus on preferences in general rather than on conceptions of the good life.

²⁹ Dworkin, Ronald, ‘Reverse Discrimination’, Taking Rights Seriously (London, 1977)Google Scholar.

³⁰ Dworkin, ‘Reverse Discrimination’, p. 233.

³¹ Dworkin, ‘Reverse Discrimination’, p. 234.

³² Dworkin, ‘Reverse Discrimination’, p. 235; emphasis added.

³³ Dworkin backs away from this argument in his response to Herbert Hart. I consider this momentarily.

³⁴ Here we must assume that having certain local political preferences facilitates the satisfaction of your national political preferences, and vice versa. They have to be mutually reinforcing.

³⁵ Here I am assuming that there is more to welfare than mere preference satisfaction.

³⁶ Dworkin, Ronald, ‘Rights as Trumps’, Theories of Rights, ed. Waldron, J. (Oxford, 1984)Google Scholar. See also Hart, Herbert, ‘Between Utility and Rights’, Essays in Jurisprudence and Philosophy (Oxford, 1983)Google Scholar.

³⁷ The other argument concerns whether a utilitarian can be neutral between the truth of his own theory and that of competing non-egalitarian theories such as the theory that certain people's preferences should be weighed more heavily than others. Though interesting, this is not immediately relevant for my purposes.

³⁸ Dworkin, ‘Rights as Trumps’, p. 366; emphasis added.

³⁹ See Raz, The Morality of Freedom, p. 380.

⁴⁰ Equality may be an important distributive ideal even if it is not important in every context. See, for instance, Temkin, Larry, ‘Egalitarianism Defended’, Ethics 113 (2003), pp. 764–82CrossRefGoogle Scholar.

⁴¹ See Rawls, A Theory of Justice, pp. 126–30.

⁴² See Raz, The Morality of Freedom, ch. 14.

⁴³ I would like to thank Jim Griffin, Ruth Chang, Doug Husak, Howard McGary, Anthony Ellis, and especially Larry Temkin for their extremely helpful comments on earlier versions of this article.