`Ought' Conversationally Implies `Can'

The principle that 'ought' implies 'can' is often taken for granted in discussions of determinism, moral dilemmas, and other areas of practical reasoning. Yet the principle is seldom discussed critically or in detail. A complete discussion would include analyses of the controversial terms 'ought' and 'can', but I need not here undertake such enterprises. Any plausible analyses which do not beg the question are compatible with what I have to say, except where noted.' My focus will be on the relation between the terms 'ought' and 'can'. The relation is supposed to be an implication, but there are many kinds of implication. I will discuss three kinds of implication, although even finer distinctions could, of course, be drawn.2 The three kinds of implication are entailment, presupposition, and conversational implication. As I use the terms, entailment and presuppostion are semantic, or concern the truth conditions of what is said. In contrast, conversational implication is pragmatic, or concerns the effects of saying what is said. Which kind of implication holds between 'ought' and 'can' affects the truth value of a judgment that an agent ought to do something the agent cannot do.3 If 'ought' entails 'can', and an