The central argument of this article is that there is no fact of the matter, no evidence, however tentative or questionable, that will serve adequately to identify actions "chosen" or "determined" for the purposes of sociological theory. This argument will be developed with reference to the two theorists of the greatest importance in advocating the sociological value of the concept of agency: Talcott Parsons, with his "voluntaristic theory of action," set the scene for the whole agency and structure debate in modern sociology, and Anthony Giddens, in his theory of structuration, provides the most comprehensive recent account. Both theorists put forward grounds and justifications for their use of the concepts of "choice" and "agency," but it will be argued here that in the last analysis, none of them has any sociological merit.
