The paper considers a version of the problem of linguistic creativity obtained by interpreting attributions of ordinary semantic knowledge as attributions of practical competencies with expressions. The paper explains how to cope with this version of the problem without invoking either compositional theories of meaning or the notion of 'tacit knowledge' (of such theories) that has led to unnecessary puzzlement. The central idea is to show that the core assumption used to raise the problem is false. To render precise argument possible, the paper first identifies and removes some relevant semantic indeterminacy in philosophical talk of 'semantic knowledge' and 'information'. This yields rules for attributing the two to human speakers and information-processors, respectively. The paper then shows, first, that ordinary speakers qualify as possessing all along an other than finite and definite stock of semantic knowledge and, second, that a very simple information-processor running a procedural semantics qualifies as possessing an analogous stock of semantic information. The second result is used to bring out that the first is neither unduly impressive nor particularly puzzling.