I have presented versions of this article at the British Society for Ethical Theory, UCL, the Hebrew University, Balliol College, Oxford, and the Universities of Bristol, Manchester, and Vermont. I would like to thank the members of these audiences for their comments. I would also like to thank Alex Voorhoeve for extensive discussion of the claims and arguments I advance in this article and Tyler Doggett, Peter Graham, Peter Vallentyne, Andrew Williams and Gabriel Wollner for their written comments.
1 Rawls, John, A Theory of Justice, rev. edn. (Cambridge, Mass., 1999), p. 27.
2 Rawls, A Theory of Justice, p. 26.
3 See Nozick, Robert, Anarchy, State, and Utopia (New York, 1974), p. 41. Nozick presses a separateness-of-persons objection to utilitarianism (pp. 32–3).
4 See Parfit, Derek, ‘Equality or Priority?’, The Lindley Lecture (Lawrence, Kan., 1991). Henceforth, all references to ‘prioritarianism’ are to the ‘pure’ version of this view that Parfit presents and considers in this lecture.
5 Otsuka, Michael and Voorhoeve, Alex, ‘Why It Matters That Some Are Worse Off Than Others’, Philosophy and Public Affairs 37 (2009), pp. 171–99.
6 This example is inspired by some remarks of Nozick's in Anarchy, State, and Utopia, p. 33.
7 This is true even though the diversion into his own pension fund might be regarded as paternalistic.
8 See section I and pp. 181–2 of that article.
9 See ‘Why It Matters’, p. 188.
10 At least, it does so on a standard interpretation, on which the actual rather than the expected lifetime utility of individuals is that which receives prioritarian weight. In section VIII, I consider a non-standard interpretation of the view, which attaches prioritarian weight to expected rather than actual utility.
11 Note that we're considering the interests of the actually impaired in isolation from the interests of other actual people.
12 Cf. Thomas Nagel's objection to utilitarianism that it ‘depends on an application to interpersonal conflicts of the same principles which are used to settle conflicts between reasons arising from the interests of a single person. The conditions of choice corresponding to this principle are that the chooser should treat the competing claims arising from distinct individuals as though they all arose from the interests of a single individual, himself. He is to choose on the assumption that all these lives are to be amalgamated into one life, his own. . . . But this . . . completely distorts the nature of the competing claims, for it ignores the distinction between persons’ (Nagel, , The Possibility of Altruism (Princeton, 1970), p. 138). In his ensuing discussion on pp. 138–40, Nagel appears, however, to blend elements of each of the two separateness-of-persons objections that I have just distinguished.
13 ‘Why It Matters’, pp. 183–4. I have slightly modified this passage in order to make it more directly relevant to the cases under discussion in this article.
14 More precisely: either benefit one or more persons or benefit another or more other persons.
15 Not only are these two complaints distinct, but the complaint that prioritarianism is insensitive to luck egalitarian unfairness should be regarded as a further way, distinct from the two that I have already identified in this article, in which prioritarianism fails to take the distinction between persons seriously. See Alex Voorhoeve and Marc Fleurbaey, ‘Egalitarianism and the Separateness of Persons’ (this issue), for a discussion of the precise way in which luck egalitarian considerations of unfairness are sensitive to the distinction between persons.
16 In order to show that the competing claims complaint is distinct from the complaint of luck egalitarian unfairness, one can also construct cases in which there is a complaint of unfairness but no competing claims complaint against treatment for the slight impairment. One such case is the two-actual-person case in which either of two possible persons might be very severely impaired (case 5) that I introduce in section VIII.
17 Among such cases in which they have no significance are those Parfitian non-identity cases in which it is because of our choices that these merely possible people have no chance of being actual. For a discussion of non-identity cases, see Parfit, , Reasons and Persons (Oxford, 1986), ch. 16.
18 But what, one might ask, if we choose to bring the person with the very severe impairment into existence and treat his impairment? Couldn't someone press a valid complaint on behalf of the merely possible person with the slight impairment even though our choice ensures that he never come into existence? The answer is no, since a possible person doesn't have any claim to be brought into existence, either on grounds that such existence would be very good in absolute terms or on grounds that his existence would be better than the existence of someone else we instead create. Only those who either do exist or stand a chance of existing have morally significant claims, where these claims are tied to their interests in the event, as chance or certainty would have it, that they exist, and the weight of these claims is a function, among other things, of the likelihood that they will exist.
19 This is a slightly simplified version of a case that we present on p. 197 of ‘Why It Matters’. In his PhD thesis, Gabriel Wollner constructs an analogous two-person case involving inversely correlated risks and notes that, in such a case, the option that increases ex post inequality between the worse off and the better off can nevertheless be justified to each as in his ex ante self-interest. Wollner shows that, while Voorhoeve and I were aware of this feature of this sort of case, we failed to see how it implies that the ‘separateness of persons’ objection involving insensitivity to prudential justifications that we advanced in ‘Why It Matters’ is separable and distinct from our ‘separateness of persons’ complaint that prioritarianism is insensitive to relational and comparative considerations. (See Gabriel Wollner, ‘Egalitarianism with a Human Face’, UCL PhD thesis (2010), pp. 58–61.) Here and elsewhere in this article, I attempt to remedy this earlier failure.
20 For a similar argument against the maximization of expected utility in cases such as this one, see Marc Fleurbaey and Alex Voorhoeve, ‘Decide as You Would with Full Information! An Argument against Ex Ante Pareto’, Health Inequality: Ethics and Measurement, ed. Nir Eyal, Samia Hurst, Ole Norheim and Dan Wikler (Oxford, forthcoming).
21 In a section of ‘Why It Matters’ entitled ‘Have We Correctly Described How the Priority View Deals with Choice under Risk?’ (pp. 195–8), we label this the ‘ex ante Priority View’. This section builds upon and in some respect revises that section.
22 See Rawls, A Theory of Justice, p. 56.
23 See ‘Why It Matters’, p. 196.
24 Let us assume that the lives of those who are very severely impaired or severely impaired would be worth living and that such people would not regret their existence.
25 Unlike our two-person case involving certainty (2), this case shares the following feature with our one-person case involving risk (1), in which it is reasonable to provide treatment for the slight impairment: one person has a claim to the treatment for the slight impairment, and that claim doesn't compete with anyone else's claim for treatment for the very severe impairment.
26 He registers no difference for the following reason. Since these cases involve no uncertainty regarding who would benefit, and to what degree, under any given course of action, expected utility is identical to actual utility. Moreover, treatment for the slight impairment yields just as much actual utility in each case, as does treatment for the very severe impairment.
27 In Michael Otsuka, ‘Risking Life and Limb’ (unpublished manuscript), I offer an elaboration of this point regarding the significance of epistemic versus objective risks.
28 I am assuming that one's ex ante self-interest is calculated on the basis of epistemic chances. I understand ‘epistemic chances’ to be those that are constituted by one's justified belief, given the evidence, in what the objective chances are. Such justified belief need not be true.
29 The ex ante prioritarian might offer the following ad hominem rejoinder: ‘In so downgrading the significance of expected utility, you undermine the first of your two separateness-of-persons objections to prioritarianism, since, in making this objection, you rely on the claim that the fact that a given policy maximizes the expected utility of an individual provides a justification to that person. How then, can you consistently turn around and belittle the significance of mere expected utility in questioning the prioritarian's move from an ex post to an ex ante view?’ My reply is that if expected utility is all that we have to go on, then it's justifiable to be guided by it. But if, as in the case of inversely correlated risks, we can overcome the ignorance on which the expected utility calculations are based, we should do so. We should do so for reasons that Fleurbaey and Voorhoeve spell out in ‘Decide as You Would with Full Information!’.