Cyril Hédoin (REGARDS, Université Reims-Champagne Ardennes, France) has recently published online a working paper entitled “Utilitarianism, Prioritarianism and Nonlinear Social Welfare Functions: Should We Accept Bernoulli’s Hypothesis“.
Abstract: Abstract: Utilitarianism and prioritarianism are currently two of the most important consequentialist theories in ethics. Prioritarians hold that, everything else equal, it is better that benefits go to the worse-off people than to the better-off, i.e. a given amount of good counts more the less well-off are the people it goes to. However, axiomatic defenses of utilitarianism relying on expected utility theory, such as those of John Harsanyi (1955) and John Broome (1991), offer a surprising and counterintuitive objection to prioritarianism: the distinction between an amount of good and how this amount count is empty and meaningless. This objection ultimately depends on the validity of “Bernoulli’s hypothesis” which states that an alternative is better for someone than another one if its expected goodness is higher. This article weights the reasons for and against Bernoulli’s hypothesis. I show that the strongest argument for Bernoulli’s hypothesis comes from the fact that in some cases at least the general goodness function cannot be nonlinear in individual goodness while satisfying the strong independence axiom. However, it is not clear that this problem has an empirical significance. As a result, expected utility theory only weakly support utilitarianism against prioritarianism. This seems to be a particular instance of the more general rule that ethics does not supervene on rationality.
Keywords: Utilitarianism – Prioritarianism – Expected Utility Theory – Bernoulli’s hypothesis – Strong independence – John Broome.