Suppose, just for the sake of argument, that the version of the dispositional theory of value that I myself prefer is correct (see, for example, Smith 1989, 1994a, 1997; compare Lewis 1989, Johnston 1989): when a subject judges it desirable for p to be the case in certain circumstances C, this is a matter of her believing that she would want p to be the case in C if she were in a state that eludes all forms of criticism from the point of view of reason – or, for short, and perhaps somewhat misleadingly (Copp 1997), if she were fully rational. More precisely, if still somewhat misleadingly, let's suppose that when a subject judges it desirable that p in C this is a matter of her believing that, in those nearby possible worlds in which she is fully rational – let's call these the “evaluating possible worlds” – she wants that, in those possible worlds in which C obtains – let's call these the “evaluated possible worlds” – p obtains.

Once we have supposed this to be so it is, I think, extremely tempting to suppose that we have thereby either explicitly or implicitly taken a stand on certain crucial debates in meta-ethics: tempting to suppose that we must be cognitivists as opposed to non-cognitivists; relativists as opposed to non-relativists; and realists as opposed to irrealists. We must be cognitivists because we have supposed that evaluative judgement is a species of belief.