The statement is hard to define concisely because it covers so much ground. Everything from bar graphs to contracts to the opening sentence of a novel can be analyzed as a statement. Thus, our sense of what a statement is depends less on compiling a set of specific examples than it does on taking a specific approach to the analysis of a set of signs. If we take a piece of speech or writing as a statement, what aspects must we pay attention to? What do we want to know about signs in their role as statements? Foucault, in fact, devotes most of this chapter to showing what we don't want to know about statements. We don't, for example, want to know about the propositional content of the sentence, 'The golden mountain is in California,' nor are we concerned with its external referent (or about whether or not it 'really' exists). Instead, we would want to know about the position of this statement in relation to other actual and possible statements. Does the statement occur in a conversation? A novel? What other statements make it possible? What forms of authority does it depend on?

The analysis of statements also excludes any consideration of a human subject as a thinking, intending creator of the set of signs in question. For a given group of signs, the Foucauldian method does not ask how they resulted from an individual psychology, or what the motivation of the author or speaker was. The speaking or writing subject is replaced by an authorial function, a function that tells us from where and from what authority the statement comes without telling us anything about the actual, human author of the statement. Anyone can occupy a given subject-function, and any one person can switch between a wide range of such positions. Thus, in the analysis of the statement, we are not really concerned with 'context' in its traditional sense; the field in which the statement gains its identity is not one of physical objects and intending authors, but of positions, institutions, and above all, other statements.

There is a confusing undercurrent that runs throughout Foucault's separation of the statement from considerations of propositions and authors. The problem is that statements, though they involve much more than either of these things, are quite capable of containing both of them. When we analyze the sentence about the golden mountain, we are not concerned with the sentence as a proposition about an external referent, and we are not concerned with the psychology of the speaker. Yet these considerations can reappear at a different level (the level of the statement and its associated field). If the sentence about the golden mountain is spoken by a patient in an insane asylum, we would not, according to Foucault's method, focus on the likelihood of the mountain being wholly fictional (the propositional content) or on the particular madness of the patient (the psychology of the author). But it would still be crucial to know that the statement is, say, part of a series of statements made by insane people, or part of a diagnostic session.

The difference here is really in where we, as historians, are led when we analyze documents. Considering the golden mountain as a product of a particular psychology might lead us to speculate on what's 'really' going on in the head of the author, and perhaps to bolster these speculations with considerations of statements by the same author, or again to examine similarly strange visions in other time periods. Considering the golden mountain as a statement, however, we will be led to consider its place in a different kind of field; we would look for documents written in response to the statement (diagnoses, perhaps, or vehement denials that the golden mountain exists), and our conclusions will be about the rules that define psychiatric discourse rather than about the mind of the particular patient or the actual existence of the golden mountain. Considerations of such psychological or physical elements still exist, but only in other, related statements. The Foucauldian historian takes the field of statements as his or her only and most basic assumption, and the description of statements related in a discourse as his or her only goal.

Popular pages: The Archaeology of Knowledge