Many different kinds of statements constitute the discourse of medicine in the nineteenth century (the field that remains our central example). What laws 'operate behind' this set of statements, linking them together? What 'place' do they all come from? Once again, there are three levels of approach to this particular set of questions. First: 'who is speaking?' What is the position from which the doctor speaks? A long list of factors is involved here, ranging from 'criteria of competence and knowledge' to relational systems such as professional and pedagogical hierarchies to the shifting role of the doctor as a guardian within society as a whole. Second: from what site is he or she speaking? A hospital, a laboratory, or a library? What were the changing functions of these sites in the nineteenth century? Third: What is the position of the subject with regard to 'various domains or groups of objects [physical things, not objects of discourse]?' This is a question about perceptual positioning, modes of and ideas about 'seeing,' observation, about instruments that act as perceptual intermediaries, and about the level of thing observed (body, organ, cell, and so on). It also addresses the position of the doctor as an 'emmitter and receiver' of observations, case histories, data, theoretical propositions, clinical decisions, etc. Again, this set of possible positions changed radically in the nineteenth century.
Thus, the question of where a given statement comes from implicates another complex set of relations. The advent of clinical medicine in the nineteenth century cannot be understood solely as a result of the advent of the autopsy or the teaching hospital (which had, in any case, significant precursors), but only as 'the establishment of a relation.' In an important sense, this relation between diverse elements is 'effected' by the clinical discourse itself; the relation only exists as a relation by virtue of the set of local enunciations that comprise the discourse. This set of enunciations, however, is not 'unified'; it does not cohere in a single rational project or a striving toward a single, unattainable goal. Neither does it imply a single positionality from which the transcendent subject of the discourse speaks: 'instead of referring back to the synthesis or the unifying function of a subject, the various enunciative modalities manifest his dispersion.' Just as the regularity of discursive objects does not depend on words or on things, the regularity of a set of enunciations does not depend on 'recourse to a psychological subjectivity.'
Although the conceptual architecture of something like 'grammar' has been found actually to split into a number of temporally based and conditioned grammars (see Section two), it is still possible to try to assimilate them to a single, totalized structure of grammar. Foucault's historical project, however, must address a 'broader scale,' and the structure that emerges is not one of concepts that fit into a rigorous whole, but concepts in their historical specificity. How can we describe 'the organization of the field of statements' in which concepts emerge and circulate?
First of all, this organization depends upon forms of 'succession.' One enunciation of a concept follows another, and each element in this series depends on the others in a myriad of ways (ways which do not simply map onto the order of succession). Natural history in the seventeenth and eighteenth centuries, for example, was not simply an invention of concepts like 'mammal,' but rather 'a set of rules for arranging statements in a series, an obligatory set of schemata in which the recurrent elements that may have value as concepts were distributed.' Second, we must look to forms of 'coexistence' that mark out a given set of concepts. These forms include: the 'field of presence' (those statements accepted by a given discourse at a given time as central or foundational concepts, and defined as much by exclusions as inclusions); the 'field of concomitance' (comprised of statements outside the discourse that serve as points of analogy or higher authority, such as cosmology for natural history); and the 'field of memory' (statements no longer accepted, but seen as precursors). Finally, there are 'procedures of intervention' to consider; these procedures determine the ways in which statements can be translated, systematized, redefined, rewritten, and so on, and they vary from one discourse to another.
The relations between these elements of succession, coexistence, and intervention define a 'system of conceptual formation.' An analysis of these elements does not provide a theory about the system or progression of the concepts themselves, but rather their 'anonymous dispersion' at a 'preconceptual' level. This field of dispersion, and the rules t which it is subject, is part of what characterizes a particular discourse (here Foucault gives an example of this preconceptual analysis from his discussion of grammar in The Order of Things). Such an analysis is not concerned with recovering the processes of thought or invention that gave rise to a particular progress of conceptual development. Rather, the pre-conceptual analysis describes 'the discursive regularities that have made possible the heterogeneous multiplicity of concepts.' Just as the analysis of the rules of formation for objects is not an analysis of words or things, and just as the analysis of the formation of enunciative types does not study individual psychology, so the analysis of the formation of concepts is not concerned with the progress of conceptual ideas per se.
Foucault continues his series of reconstructions after the rejection of his four 'hypotheses.' The third hypotheses is that discursive unities can be identified by virtue of a shared style or point of view, such as a certain 'descriptive' quality that characterizes all nineteenth-century medical discourse. This proved, of course, to be to simple. The idea of a single, consistent, perspective-based stylistic relation between statements in a discourse is here replaced with a notion that will be the focus of most of Part II of the book: the enunciation. The difference, for Foucault, lies largely in psychology; namely, the enunciative function doesn't need to assume any such thing. When Foucault asks, 'who is speaking?' or what the 'place' is from which a group of statements come, he proposes nothing about the speaker apart from his or her situation in relation to a network of institutional structures, norms of expression, and the groups of other statements in which these are constituted. In this series of 'formation' chapters, it is important to avoid any attention to the interiority of the individual thing itself. Thus, the 'object' is identified by its emergence from discursive relations and not by its nature as a physical phenomenon, the speaking subject is characterized by a relative positionality and not by an individual psychology or perspective, and the concept is characterized by the procedures by which it comes to be accepted and revised and not by its content as a pure idea. We should note that the enunciative formation stands out from the other two 'formations' discussed so far in Foucault's method of individualizing discourses. While the object and the concept seem almost to disappear almost entirely into the relational regularities (and irregularities) of the discursive field, the idea of a style or point of view has a specific replacement.