A theory of language use must be able to shew how it is that the meanings of the symbols and expressions of a language can be learnt by human beings, or, what comes to the same thing, how one person can test whether another person has the same meaning in mind. The semantics we have so far been presenting have had little to say on this problem. E.g., consider the motivation given on pp. 23-5 for thinking of a proposition as a set of possible worlds. The impression left by that passage might seem to be that if one person wishes to tell another what proposition he has in mind he must in some sense shew that other person each possible world in turn and tell him whether the proposition is true in that other world or not. Our purpose in Chapter One was of course to give a preliminary account of what a proposition might be, and nothing in the present chapter is intended to affect the ontological status of these entities, but it would be very implausible to suppose that the procedure described above is the way we go about teaching the meanings of words.