We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 21 September 2015
John Harsanyi has offered an argument grounded in Bayesian decision theory that purports to show that John Rawls's original position analysis leads directly to utilitarian conclusions. After explaining why a prominent Rawlsian line of response to Harsanyi's argument fails, I argue that a seemingly innocuous Bayesian rationality assumption, the continuity axiom, is at the heart of a fundamental disagreement between Harsanyi and Rawls. The most natural way for a Rawlsian to respond to Harsanyi's line of analysis, I argue, is to reject continuity. I then argue that this Rawlsian response fails as a defence of the difference principle, and I raise some concerns about whether it makes sense to posit the discontinuities needed to support the other elements of Rawls's view, although I suggest that Rawls may be able to invoke discontinuity to vindicate part of his first principle of justice.
1 Rawls, John, A Theory of Justice, rev. edn. (Cambridge, MA, 1999)Google Scholar and Justice as Fairness: A Restatement (Cambridge, MA, 2001). In this article I will focus primarily on Rawls's view as presented in Justice as Fairness, which he describes as rectifying some mistakes from Theory and unifying the view from Theory with that developed in many of his subsequent essays. See Rawls, Justice as Fairness, pp. xv–xviii.
2 In Rawls, Theory, the discussants are imagined as choosing principles for a society in which they will (or do) live, whereas in Rawls, Justice as Fairness, they are redescribed as trustees, so that they are choosing principles for a society on behalf of one who will live (or does live) in that society. This distinction does not matter for my analysis.
3 Rawls is, of course, also concerned to argue against other alternatives, such as libertarian principles, particularly ‘right-libertarian’ ones. But, especially in Justice as Fairness, he frames his analysis primarily as an attempt to demonstrate that his principles are preferable to utilitarian ones, and then points out the ways in which the arguments developed for that purpose also suffice to show that his principles are superior to other contenders. The emphasis on utilitarianism as the primary alternative to his view is perhaps most evident from the fact that Rawls structures his argument as involving two ‘fundamental comparisons’, first between his preferred principles and an unrestricted form of utilitarianism and second between his principles and restricted forms of utilitarianism.
4 Hare, R. M., ‘Rawls’ Theory of Justice’, Reading Rawls, ed. Daniels, Norman (New York, 1975), pp. 81–107 Google Scholar; Vickrey, William, ‘Measuring Marginal Utility by Reactions to Risk’, Econometrica 13 (1945), pp. 319–33CrossRefGoogle Scholar. Harsanyi's earliest version of the objection is in Harsanyi, John C., ‘Cardinal Utility in Welfare Economics and in the Theory of Risk-Taking’, Journal of Political Economy 61 (1953), pp. 434–5CrossRefGoogle Scholar. Harsanyi asserts that his own analysis was independent of Vickrey's in Harsanyi, John C., ‘Morality and Theory of Rational Behavior’, Social Research 44 (1977), pp. 623–56Google Scholar, at 634.
5 There is a great deal of debate and literature on this. For an example of an argument of this sort posed as an objection to utilitarianism, see Bernard Williams, ‘A Critique of Utilitarianism’, Smart, J. J. C. and Williams, Bernard, Utilitarianism: For and Against (Cambridge, 1973), pp. 77–118 CrossRefGoogle Scholar, at 134–5. For examples of arguments of this sort incorporated into a utilitarian perspective, see Hare, R. M., Moral Thinking: Its Levels, Method and Point (Oxford, 1981)CrossRefGoogle Scholar, and Railton, Peter, ‘Alienation, Consequentialism, and the Demands of Morality’, Philosophy and Public Affairs 13 (1984), pp. 134–71Google Scholar.
6 Rawls's earliest published formulation of the idea that developed into his original position analysis is in Rawls, John, ‘Justice as Fairness’, The Journal of Philosophy 54 (1957), pp. 653–62CrossRefGoogle Scholar. Harsanyi initially presents his version of an original position (or equiprobability) argument in Harsanyi, ‘Cardinal Utility’, and presents a related axiomatic argument in Harsanyi, John C., ‘Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility’, Journal of Political Economy 63 (1955), pp. 309–21CrossRefGoogle Scholar. He deploys the original position argument as a criticism of Rawls in Harsanyi, John C., ‘Can the Maximin Principle Serve as a Basis for Morality? A Critique of John Rawls's Theory’, The American Political Science Review 69 (1975), pp. 594–606 CrossRefGoogle Scholar, and develops his own view more fully and explicitly, presenting the two arguments side by side, in Harsanyi, John C., ‘Morality and Theory of Rational Behavior’, Social Research 44 (1977), pp. 623–56Google Scholar. For an excellent analysis of Harsanyi's arguments, focusing on the axiomatic version and developing a similar argument of his own, see Broome, John, Weighing Goods: Equality, Uncertainty and Time (Oxford, 1991)Google Scholar.
7 Harsanyi, ‘Morality’, p. 634.
8 This is identified as the ‘standard objection’ to Harsanyi in Broome, John, ‘Can There Be a Preference-Based Utilitarianism?’, Justice, Political Liberalism, and Utilitarianism: Themes from Harsanyi and Rawls, ed. Fleurbaey, Marc, Salles, Maurice and Weymark, John A. (Cambridge, 2008), pp. 221–38CrossRefGoogle Scholar, at 231, and that characterization is endorsed by Marc Fleurbaey, Maurice Salles and John A. Weymark, ‘An Introduction to Justice, Political Liberalism, and Utilitarianism’, Justice, Political Liberalism, and Utilitarianism, pp. 1–67, at 17. The line of objection originates in Sen, Amartya, ‘Welfare Inequalities and Rawlsian Axiomatics’, Theory and Decision 7 (1976), pp. 243–62CrossRefGoogle Scholar, and is developed and clarified in Weymark, John A., ‘A Reconsideration of the Harsanyi–Sen Debate on Utilitarianism’, Interpersonal Comparisons of Well-Being, ed. Elster, Jon and Roemer, John (Cambridge, 1991), pp. 255–316 CrossRefGoogle Scholar, and John Roemer, ‘Harsanyi's Impartial Observer Is Not a Utilitarian’, Justice, Political Liberalism, and Utilitarianism, pp. 129–35. There is also helpful discussion in Fleurbaey, Salles and Weymark, ‘An Introduction’, pp. 17–20 and 30–2.
9 The interpersonal comparisons require not merely cardinality but co-cardinality. Nonetheless, without cardinality neither the intrapersonal nor the interpersonal comparisons are possible.
10 I will focus on the response offered in Broome, ‘Can There Be’. An alternative way of defending Harsanyi's argument, by adopting a substantive account of well-being independent of individuals’ preferences, is suggested in Weymark, ‘A Reconsideration’, pp. 307–15. A version of this strategy, focused on idealized desires, is developed at greater length in Risse, Mathias, ‘Harsanyi's “Utilitarian Theorem” and Utilitarianism’, Nous 36 (2002), pp. 550–77CrossRefGoogle Scholar.
11 Thanks to an anonymous referee for suggesting this way of understanding the objection to Harsanyi. This is also the way the objection is characterized in Broome, ‘Can There Be’, pp. 230–1 and in Fleurbaey et al., ‘An Introduction’, p. 17.
12 Broome, ‘Can There Be’, pp. 230–2. Although Broome defends Harsanyi on this point, he finds other aspects of Harsanyi's argument problematic. Nonetheless, I take Broome to be pursuing a line of argument and analysis closely related to and inspired by Harsanyi, and one that has the potential to pose similar sorts of problems for Rawls. See Broome, Weighing Goods.
13 Doubts about Broome's approach are raised by Fleurbaey, Salles and Weymark, ‘An Introduction’, and Weymark, John A., ‘Measurement Theory and the Foundations of Utilitarianism’, Social Choice and Welfare, 25 (2005), pp. 527–55CrossRefGoogle Scholar.
14 Rawls, Theory, pp. 78, 150, 281–5.
15 Rawls, Justice as Fairness, pp. 57–61 and 127, and Theory, p. 284.
16 Rawls, Theory, p. 78.
17 I have been focusing on Harsanyi's equiprobability argument, but the same conclusion follows for his axiomatic argument. Absent the information needed to make probabilistic judgements about the frequency of different social roles, neither argument can achieve its aims.
18 Rawls, Justice as Fairness, p. 18. Rawls says essentially the same thing in his discussion of the veil of ignorance in Justice as Fairness, pp. 82, 87.
19 James Lenman, ‘Review of Allan Gibbard, Reconciling Our Aims: In Search of Bases for Ethics’, Notre Dame Philosophical Reviews (2009), <https://www.ndpr.nd.edu/news/24177-reconciling-our-aims-in-search-of-bases-for-ethics/> (2014).
20 In part II of Justice as Fairness, Rawls offers some general considerations in favour of his preferred principles of justice, but reserves his ‘more formal and official argument’ (Rawls, Justice as Fairness, p. 39) to part III, which is entirely structured in terms of what it would make sense to choose as parties in the original position. Given that, and the fact that Justice as Fairness is described by Rawls as a restatement and clarification of his view (Rawls, Justice as Fairness, pp. xv–xviii), it is hard to resist the conclusion that Rawls takes the original position argument to be a central and substantive element of his overall analysis.
21 For a classic version of this sort of argument, see Williams, ‘A Critique’, which highlights both counterintuitive conclusions that are supposed to follow from utilitarianism and counterintuitive features supposedly displayed by utilitarian deliberation even when it leads to intuitively plausible conclusions.
22 For an extended argument that utilitarianism avoids most of the counterintuitive conclusions typically associated with it, and, perhaps more importantly, that the intuitive plausibility of conclusions of this sort is not dispositive of utilitarianism, see Hare, Moral Thinking.
23 Gibbard, Allan, Reconciling Our Aims: In Search of Bases for Ethics (Oxford, 2008), pp. 50–1CrossRefGoogle Scholar and 150–2.
24 Gibbard, Reconciling, p. 50.
25 Gibbard, Reconciling, p. 152, writes: ‘if the retort leaves cold someone who genuinely understands what it involves, then I don't know anything to say that would make the person responsive.’ As Gibbard, Reconciling, p. 50 n. 30, notes, Broome and Brian Barry report being unmoved by this general line of thought in Broome, Weighing Goods, pp. 55–9 and Barry, Brian, Theories of Justice (Berkeley, 1989), pp. 334–5Google Scholar. Perhaps, though, it is significant that they are responding to the idea as formulated by Harsanyi rather than to Gibbard's own formulation in terms of a ‘You’d have agreed’ retort, which seems to me to highlight the moral force of the idea.
26 As indicated above, this is not to say that invoking reflective equilibrium in this way makes it impossible to argue against utilitarianism, just that it makes it impossible for that argument to be grounded, in any meaningful way, in an original position analysis.
27 Rawls, Theory, p. 146. Thanks to an anonymous referee for calling my attention to this quote, and the related line of argument.
28 Rawls, Theory, pp. 149–50.
29 I am stipulating that these assertions are uncontested in order to focus attention on a clear case in which opting for an original position with a thin veil of ignorance leads to a different conclusion from opting for one with a thick veil of ignorance.
30 See, for instance, Rawls, Justice as Fairness, p. 101 (in the context of the first fundamental comparison) and p. 120 (in the context of the second).
31 These conditions are formulated on Rawls, Justice as Fairness, p. 98. My presentation of the list is different from his in two ways, but the substance is the same. The differences are as follows. (1) He lists the lack of information required for probabilistic reasoning as the first of three conditions. That does not appear on my list because I am, at this point, attempting to capture only his other conditions. (2) He combines the first and second conditions on my list. Together, they constitute the second condition on his list. I have separated them in an effort to ensure clarity in my subsequent analysis.
32 This also avoids creating the false impression that Rawls wants to derive the difference principle from the maximin rule of choice. Rawls, Justice as Fairness, p. 43 n. 3 talks about this misconception, and admits his own role in creating it. My preferred approach also avoids creating the impression that Rawls is advocating the general use of a maximin principle in deliberation, which he is not, as indicated in Rawls, Justice as Fairness, p. 97 n. 9 and p. 99. It is also worth noting that Rawls, Justice as Fairness, p. 99 acknowledges that the maximin principle is not essential to his analysis, and that a presentation like the one I am offering fully captures his line of thought. Given the extent to which framing the argument in terms of the maximin principle has led to misunderstanding and confusion, and the fact that the argument can be presented more simply without appeal to the maximin principle, I think it is preferable to work with a version of Rawls's view that avoids invoking maximin reasoning.
33 As discussed in sect. I, one could also block Harsanyi's argument by appealing to the standard objection posed by Sen, Weymark and Roemer. That may very well be an effective response to Harsanyi, but, as indicated earlier, it does not shed light on what I take to be the core disagreement between Rawls and Harsanyi.
34 Harsanyi, ‘Morality’, p. 636 n. 17.
35 For a discussion of the sure thing principle in connection with the Allais paradox and related objections, see Broome, Weighing Goods, pp. 94–117. For a discussion of the sure thing principle in connection with the Ellsberg paradox, see Binmore, Ken and Voorhoeve, Alex, ‘Defending Transitivity against Zeno's Paradox’, Philosophy and Public Affairs 31 (2003), pp. 272–9CrossRefGoogle Scholar.
36 Harsanyi, ‘Cardinal Welfare’, p. 313 n. 10.
37 Broome, Weighing Goods, p. 94. Similarly, Risse takes continuity to be potentially controversial only in the context of comparing the costs and benefits of eternal damnation or salvation with earthly goods. See Risse, ‘Harsanyi's “Utilitarian Theorem”’, p. 561.
38 Strictly speaking, the continuity axiom on its own is only sufficient to ensure that there is some very small risk of R such that I am indifferent between running that risk or not running it. The conclusion as stated in the text also requires use of the sure thing principle. Stating it as I have done highlights the potential Rawlsian objection, and it should be clear, even to those who are suspicious of the sure thing principle in general, that the continuity axiom is what is at issue here. (The only work being done here by the sure thing principle is to license the idea that if P is preferable to R, and one is comparing lotteries with P and R as the only outcomes, a lottery with a higher probability of P and a lower corresponding probability of R is preferable to one with a lower probability of P and a higher corresponding probability of R.)
39 Rawls’ second condition, which is that outcomes above the guaranteeable level are only marginally above it, ends up being unnecessary. As I see it, this reveals some equivocation on Rawls's part between two lines of argument. On the one hand, there is an argument like the one I am developing here on his behalf, in which discontinuity is the key to rejecting a standard utilitarian analysis. In that argument, Rawls's second condition is superfluous. On the other hand, there is an argument that I would characterize as an attempt to ground Rawls's preferred principles in standard utilitarian considerations (although Rawls would not like that characterization), and in that argument the second condition is relevant and important.
40 If the chance of being above the discontinuity were tied, then the relative evaluation of outcomes within the tiers would become significant.
41 The example of crossing a road is one of several of this sort mentioned in John C. Harsanyi, ‘John Rawls’ Theory of Justice: Some Critical Comments’, Justice, Political Liberalism, and Utilitarianism, pp. 71–9, at 72. A similar example involving plane travel is discussed at greater length in Harsanyi, ‘Can the Maximin Principle’, p. 595. See also Norcross, Alastair, ‘Great Harms from Small Benefits Grow: How Death Can Be Outweighed by Headaches’, Analysis 58 (1998), pp. 152–8CrossRefGoogle Scholar.
42 To avoid verbiage, if there is a point of value discontinuity and X is below it and Y above it, I will call X ‘discontinuously worse’ than Y.
43 The possibility that it could make sense to act as if there is a value discontinuity even if we suppose that there is not actually any such discontinuity is suggested by the lines of thought developed in Hare, Moral Thinking, and in Railton ‘Alienation’.
44 In contrast, Dale Dorsey has recently argued that discontinuity is a relatively common phenomenon. See Dale Dorsey, ‘Headaches, Lives and Value’, Utilitas 21 (2012), pp. 36–58. In particular, he claims that there is a discontinuity between the value of fulfilling what he calls a ‘deliberative project’, on the one hand, and satisfying ‘momentary desires’ or realizing ‘hedonic achievements’, on the other. His primary examples of deliberative projects are becoming a philosopher or a trombone player, and his primary example of momentary desires and hedonic achievements involves avoiding a mild headache. Dorsey's argument is complicated, and his conception of continuity is not exactly the same as mine, which further complicates matters. His analysis therefore requires a more detailed treatment than I can provide here. Briefly, though, my view is that the analysis from the text applies to his preferred examples as well. If I am told that the only way of relieving my headache poses a risk of interfering with one of my deliberative projects, that will indeed deter me from relieving the headache. But as the risk becomes increasingly slim, I think it eventually makes sense to run the risk for the sake of alleviating the headache. I take it that anyone who takes recourse to ibuprofen or acetaminophen to treat a headache concurs, or at least acts as if they concur.
45 Rawls, Justice as Fairness, p. 120.
46 Rawls, Justice as Fairness, pp. 122–30.
47 Rawls, Justice as Fairness, p. 128.
48 Rawls's analysis depends at this point on an empirical, psychological claim, and scepticism about this claim may very well be warranted. But for the sake of argument I am going to grant Rawls his psychological claim and attempt to show that his argument does not succeed even if he is correct about this.
49 Rawls, Justice as Fairness, pp. 18–19.
50 There is another way of interpreting Rawls's argument, which involves focusing on the idea that the parties in the original position have a ‘higher-order’ interest in securing the two moral powers, as asserted on Rawls, Justice as Fairness, pp. 175 and 177, and grounding the preference for Rawls's principles of justice in that higher-order interest. As I see it, though, if the assertion of a higher-order interest is interpreted as a stipulation built into the original position, then it looks much like the thickening of the veil discussed in sect. II. It is hard to see why we should adopt that original position model over one that lacks such a stipulation. On the interpretation I prefer, the idea of a higher-order interest in securing the two moral powers is itself grounded, if at all, in a value discontinuity between possessing those moral powers and not possessing them.
51 Rawls, Justice as Fairness, pp. 112–13.
52 Rawls, Justice as Fairness, pp. 148–50.
53 Rawls, Theory, p. 266 and Rawls, Justice as Fairness, p. 42.
54 Harsanyi, ‘John Rawls’ Theory of Justice’, p. 73.
55 Harsanyi, ‘Can the Maximin Principle’, p. 602.
56 I would like to thank Blain Neufeld, an audience at the University of Wisconsin–Milwaukee, and a particularly helpful anonymous reviewer for their feedback, comments and advice on earlier versions of this article.
