Abstract
In this article, I develop the so‐called ‘problem of linguistic creativity’ for two object‐languages, one finite, the other infinite. I then employ an approach first outlined by Ludwig Wittgenstein’s collaborator Friedrich Waismann, to ‘dissolve’ that problem, in a sense made precise by working through the example. This is to unsettle the computational picture of linguistic understanding as turning on the generation of semantic information about sentences on the basis of semantic information about their constituent parts, and to provide a reasonably clear conception of to what such ‘philosophical dissolution’ of a problem amounts.