The goal of this paper is a reassessment of the contributions provided by John L. Austins book How to Do Things with Words to pragmatics. It discusses some assumptions belonging to the received reading of the volume, as regards its aim and structure, the conceptions of illocution and of perlocution, and the alleged exclusion of non-seriousness. Against the received reading, it is argued that How to Do Things with Words is structured as a proof by contradiction of the claim that all speech should be considered as action, that in illocution a major role is played by the conventionality of effects, that perlocution presupposes a conception of action as responsibility, and that Austin had reasons not to deal with non-seriousness in detail, albeit recognizing the issue as relevant to the study of the uses of language. In the conclusions, the tenets attributed to Austin are neither crtiticized nor defended, but an attempt is made to say what are their implications for research into language use and for philosophy.
