What would a utilitarian agent look like? Some have taken the answer to describe an agent so incompetent and perverse that it casts doubt on utilitarianism itself. In this paper, I develop the strongest form of this 'self-effacingness' objection to utilitarianism, based on the idea of a constitutive link between rationality and normally competent agency. Assuming this understanding of rationality for sake of argument, I then suggest two ways to defend utilitarianism. One appeals to a Railtonian 'sophisticated' or two-level utilitarian psychology, though I suggest some potential problems for this approach. The second involves showing how we can develop a direct utilitarian psychology within rational constraints. In the course of distinguishing these two alternative paths, I make a distinction between dispositions that are 'extrinsically desirable' and those that are desirable in virtue of being 'well calibrated for action' -- a distinction that I then employ to illuminate the Gauthier-Parfit debate about whether it's rational to act on rationally desirable dispositions.
Comments welcome! (Especially since I expect to use this for my job talks, with it being a fairly natural sequel to my writing sample on 'The Fitting and the Fortunate'.)