It is sometimes argued that having inconsistent desires is irrational or otherwise bad for an agent. If so, if agents seem to want a and not-a, then either their attitudes are being misdescribed – what they really want is some aspect x of a and some aspect y of not-a – or those desires are somehow 'inconsistent' and thus inappropriate. I argue first that the proper characterization of inconsistency here does not involve logical form, that is, whether the desires involved (...) have the form 'a and not-a', but rather the possibility of fulfilling all one's desires; and secondly, that the 'essential' conflicts involved in such inconsistencies are quite common and no worse for an agent than contingent conflicts. I draw implications concerning moral epistemology, moral realism and the logic of attitudes. (shrink)

Many believe that the rational evaluation of actions depends on the rational evaluation of even basic desires. Hume, though, viewed desires as "original existences" which cannot be contrary to either truth or reason. Contemporary critics of Hume, including Norman, Brandt and Parfit, have sought a basis for the rational evaluation of desires that would deny some basic desires reason-giving force. I side with Hume against these modern critics. Hume's concept of rational evaluation is admittedly too narrow; even basic desires are, (...) despite their nonrepresentational nature, subject to rational evaluation. But this evaluation is not relevant to their tendency to generate reasons for action, and does not undermine the spirit of the Humean view of rational action. (shrink)