This paper deals with an aspect of the commitment-theoretic account of evaluative compounds that Simon Blackburn has recently offered. The main point of the paper is that the special account of disjunction is flawed because it fails to validate certain very simple patterns of inference. This point is brought out by considering two examples. A reply on behalf of Blackburn is considered, but it is shown that this reply is defective because it makes use of an unacceptable inference-rule. In the (...) last section, the relatively technical point about disjunction is placed in a broader context. It is argued that without an acceptable account of disjunction, expressivism entails revisionism about everyday moral discourse. (shrink)