The set of relations between a discursive formation and the sciences is called an episteme. This is not a single thing, like a world-view or a stage of human reason, but rather a set of different relations. Neither is the episteme stable; it shifts along with the discursive formation. Of a given science, the analysis of the episteme 'questions not its right to be a science, but the fact that it exists' (i.e., the discursive conditions of its emergence and transformation).
At this point, it is unclear whether an archeological analysis can address anything besides epistemes. Foucault can imagine the outline of such an project, however; one hypothetical example would be the archeological analysis of sexuality, which would aim more at the 'ethical' than at an episteme. Other examples might include analyses of painting or politics. Archeology has so far been practiced only in the realm of the episteme (the sciences in their discursive formations), but this is only because the episteme is such a large and important target: in our culture, 'discursive formations are constantly becoming epistemologized.'
This chapter is a broad examination of why Foucault's archeological method seems especially suited to the history of the sciences. In one sense, this is simply because our culture has tended to push all discursive positivities toward the status of knowledge and then toward the status of science (and sometimes across the threshold of fully formalized science). But this account only points to a more serious and complex attempt on the part of the archeological method to revise our understanding of knowledge itself.
Foucault frequently claims Kant as the reason why his method can truly be called 'analytic.' Kant revised how we think about thinking by showing that thought operates through a set of categories. Foucault's archeology further revises our ideas about knowledge, this time by showing how knowledge takes shape as an effect of discourse. Thus, knowledge is shifted from its place as the ultimate aim of discourse to a possible effect within the field of discourse: knowledge is 'that of which one can speak in a discursive practice.'
Such a formulation is both extremely liberating and extremely problematic for the study of history. It has been the catalyst for a massive body of revisionist writing that seeks to challenge what we take to be true by showing that truth to be contingent on a historical discourse; but it has also been the major focus for critics of Foucault's early work. The basic criticism asks of Foucault's method how it can simultaneously show knowledge to be contingent and claim somehow to separate its own claims from that contingency. Foucault will address this problem briefly in his Conclusion.
The place of science in discourse is described in much the same way as the place of knowledge in discourse. Each is a stage in the transformation of a discourse, not from blindness to self-recognition or from unconsciousness to self- consciousness, but from one kind of regularity to another. This is a complex process; notions of a progressive series of thresholds or a series of concentric circles (with discourse containing knowledge containing science) are helpful but ultimately simplistic. The series of thresholds is in fact extremely variable in terms of chronology, and even in terms of its order. Discourse does not 'contain' knowledge and science; they are effects both of and in discourse. As usual, archeology seeks not to show the generalizable form of the interdependence of these three levels, but to describe that interdependence in its diversity.