Abstract
Though the last decade has seen a growing interest in pragmatic approaches to scientific explanation, the question as to the real nature of these attempts has not been answered as yet. Three possibilities are investigated: 1. the pragmatic turn is a kind of linguistic turn where "pragmatic" refers to linguistic pragmatics; 2. the pragmatic turn is a semiotic turn in that the term "pragmatic" should be understood in the sense of semiotic pragmatics; 3. the pragmatic turn is in fact a methodological turn and thus does not consist in relativizing explanations to linguistic situations or interpreters, but it manifests itself in applying a well known but in the theory of science unusual way of model-formation. Assumption is argued for by contrasting two possible senses of the concept "model". As a result, the dividing line must not be drawn between "pragmatic" and "non-pragmatic" accounts of scientific explanation, but between models which constitute necessary and sufficient conditions for correct explanations, and abstract analogical models which use knowledge about a certain fragment of natural language to infer new information about the structure of explanations. The role of pragmatics in the latter type of models is characterized by referring to recent results concerning the question-answer relationship. Finally, the properties of abstract analogical models seem to suggest that the methodological-pragmatic turn is possibly progressive with respect to the special theory of science