Rereading what I said earlier about Rawls and utilitarianism, I think it needs a correction. Rawls really is proposing something different in its approach from classical utilitarianism. In particular, his approach focuses attention on the idea that we might want to pay attention to things like the relative position of different individuals in society rather than simply aggregating utilities across individuals. The most popular way of making this operational has been with social welfare weights that depend on an individual’s rank-order according to some welfare measure. I’m very attracted to this idea, partly because it’s analogous to an idea I developed in the theory of choice under uncertainty called rank-dependent (expected) utility.
Confusion arises with Rawls’ treatment because the maximally risk-averse (or inequality-averse) form of rank-dependence is maximin, just ‘as it is for expected utility. There’s a nice paper on this by Udo Ebert ‘Rawls and Bentham reconciled? in Theory and Decision 24, 215?223.
The other issue is whether you view rank-dependent models of this kind as being alternatives to utilitarianism or generalisations/variants. In the theory of choice under uncertainty, some people (including me) like to talk about generalised expected utility theory while others talk about non-expected utility theory.