The Problem of Comparative Value

Abstract

Observing that the concept of 'impersonal' value (what other philosophers have called 'objective' or 'intrinsic' value) is central to consequentialist moral systems, Thomas Hurka has recently responded to the charge that this concept is unintelligible. His suggested analysis holds that for S to be impersonally good is for it to be such that 'all moral agents ought morally to desire and pursue it'.1 I am sympathetic to such an account, having urged a similar one elsewhere.2 My aim here is to discuss a difficulty for such an analysis-the difficulty of providing an adequate account of comparative impersonal value. I will, first, raise and then briefly treat three ways of overcoming this difficulty, pointing out serious problems for two of them, and, finally, indicate some of the implications and promise of the third approach. If S's having impersonal value is to be identified with its being the case that everyone morally ought to desire and pursue S, then there are three obvious ways of trying to understand the claim that S1 is impersonally better than S2. On the first, the comparative is built into the desiderative element in such a way that Sl's being better consists in its being the case that everyone morally ought to desire and pursue S1 more than she does S2, where 'desire . . . more' means 'desire more intensely', or 'fervently', or some such. Against such an analysis Brentano argued, following Descartes, that it can never be wrong (contrary to what one ought to do) to endorse a valuable thing as strongly as one can.3 There is no such thing as erring on the side of excess so as to have the wrong amount of desire for a good (although one can desire it in the wrong way, making, for example, poor choices). However, if one cannot be faulted for desiring S2 as much as one can, then it is hard to see how it could be true that everyone ought to desire S, more than she desires S2. For, if one desires S2 as much as one can desire anything, one cannot have a greater desire for S, and that normally