A final discovery: that the objects of these diverse and specific causes are human bodies. The forces that drive our history do not so much operate on our thoughts or social institutions, or even our environment, as on our individual bodies. So, for example, punishment in the 18th century is a matter of violent assaults on the body, branding, dismemberment, execution, whereas in the 19th century it takes the apparently gentler but equally physical form of incarceration, ordered assemblies, and forced labor.
But unlike psychoanalysis, Foucault's history is not hermeneutic, that is, it does not try to interpret what we hear and read in order to recover its deeper meaning. It deals with texts, but treats them not as documents, but in the manner of an archaeologist, as monuments. Archaeologists of knowledge, in other words, do not ask what Descartes' Meditations mean, that is, what ideas Descartes was trying to express in them; rather, they use what Descartes and many other writers, famous or not, of the same period wrote as clues to the general structure of a system in which they thought and wrote.
Foucault's idea of an archaeology of thought is closely linked to the modernist literary idea that language is a source of thought in its own right, not merely an instrument for expressing the ideas of those who use it.
…Foucault begins with the fact that, at any given period in a given domain, there are substantial constraints on how people are able to think.
…what the archaeologist of thought is interested in is a further set of constraints that, for example, make it 'unthinkable' for centuries that heavenly bodies could move other than in circles or be made of earthly material.
Such constraints seem foolish to us. Why couldn't they see such things are at least possible? But Foucault's idea is that every mode of thinking involves implicit rules (maybe not even formulable by those following them) that materially restrict the range of thought. If we can uncover these rules, we will be able to see how an apparently arbitrary constraint actually makes total sense in the framework defined by those rules. He suggests that our own thinking too is governed by such rules, so that from the vantage point of the future, it will look quite as arbitrary as the past does to us.
When authors write, much of what they say is a product, not of their distinctive insight or ability, but the result of the language they are employing. For much of the text, it is just language that is speaking. One standard Romantic idea sees the author as straining against the structures of language to express unique individual insights. Here, the assumption is that an author has access to a personal, pre-linguistic vision, the expression of which must work against language's tendency to merely conventional expression.