Abstract

This thesis defends the proposition that society should be organized to maximize the long run weighted sum of its citizen's von Neumann-Morgenstern utility functions. The theory is preference based in that it is individuals' degrees of preference--and not Benthemite hedons or Sidgwickean mental states--which are to be used to rank alternative social states; a preference theory is seen to have intuitively more acceptable consequences than either a hedonistic or mental state theory. ;It is argued that a long run average utilitarian position would be selected if an individual had to choose principles by which society would be governed and he did not know what his place in society would be. I. M. D. Little's and John Harsany's methods of interpersonal comparisons of utility are discussed with a view towards overcoming the normative problems Rawls' raises in regard to the unity of expectations when computing expected average utility in the original position. Harsany's method is seen to solve such problems and is in accord with Rawls' own position on deliberate rationality; Little's method is inadequate in this respect. ;A long run utilitarianism can recommend that tastes be changed in order to maximize happiness. These taste changes enable utilitarianism to offer minorities protection because by any reasonable long run utility calculations, happiness is not maximized by persecuting religious heretics or racial minorities but rather is maximized by inculcating more tolerant preferences. Sen's objection that utilitarianism would leave some people, e.g., cripples, without any concrete protection, and his weak equity axiom, are argued against in this regard. ;Appendix 1 proves Harsany's theorem on von Neumann-Morgenstern utilities without the assumption that individuals' utility functions are linearly independent. However, it is argued that this theorem does not provide a useful argument for utilitarianism or a useful method for arriving at interpersonal comparisons of utility