Foucault’s version of discourse is the most pervasive theoretical idea in The Archeology of Knowledge. The term has a history as the object of study for a new kind of history, the history of ideas. But Foucault devotes much of The Archeology of Knowledge to refining and winnowing the usual sense of discourse into an object of analysis that is very strictly delimited. The first major alteration that Foucault makes is casting aside of everything but the processes of discourse itself. Thus, his method studies only the set of “things said” in their emergences and transformations, without any speculation about the overall, collective meaning of those statements. Archeology does not describe history through discourse; it describes the history of discourse.

Foucault carries his insistence on discourse-in-itself down to the most basic unit of things said: the statement. Just as discourse is never taken as a partial sign of a greater, partially hidden historical truth, so individual statements are never taken as expressions of a psychology, nor even as vehicles for referential meanings and propositions. Foucault addresses statements only in the specific conditions of their emergence and transformation; these conditions are themselves discursive (and sometimes institutional).

Thus, discourse is not just a set of articulated propositions, nor is it the trace of an otherwise hidden psychology, spirit, or encompassing historical idea; it is the set of relations within which all of these other factors gain their sense (their conditions of possibility). This argument is responsible both for the immense success of Foucault's method and for the most persistent criticisms of it. The idea that discourse can be described in and of itself, not as a sign of what is known but as a precondition for knowledge, opens limitless possibilities for showing that what we think we know is actually contingent on how we talk about it.

Popular pages: The Archaeology of Knowledge