Bourdieu wrote several times about the emerging ethnomethodology and "sociology of science" literature. His main complaint: it would lead to "a kind of subjectivist nihilism" (1994: 94)
Against Garfinkel (refering to his papers in Studies in Ethnomethodology):
Our approach is thus radically opposed, on two essential points, to the interactionism which reduces the constructions of social science to "constructs of the second degree, that is, constructs of the constructs made by the actors on the social scene ”, as Schutz does, or, like Garfinkel [1967], to accounts of the accounts which agents produce and through which they produce the meaning of their world. One is entitled to undertake to give an "account of accounts ”, so long as one does not put forward one ’s contribution to the science of pre-scientific representation of the social world as if it were a science of the social world. But this is still too generous, because the prerequisite for a science of commonsense representations which seeks to be more than a complicitous description is a science of the structures which govern both practices and the concomitant representations, the latter being the principal obstacle to the construction of such a science. ... Moreover, the constitutive power which is granted to ordinary language lies not in the language itself but in the group which authorizes it and invests it with authority. Official language, particularly the system of concepts by means of which the members of a given group provide themselves with a representation of their social relations (e.g. the lineage model or the vocabulary of honour), sanctions and imposes what it states, tacitly laying down the dividing line between the thinkable and the unthinkable, thereby contributing towards the maintenance of the symbolic order from which it draws its authority. (1972] 1977: 21-22)
Against Latour:
Against other ethnomethodologists
So, while we may accept, with the ethnomethodologists, that the family is a principle of construction of social reality, it also has to be pointed out, in opposition to ethnomethodology, that this principle of construction is itself socially constructed and that it is common to all agents socialized in a particular way. In other words, it is a common principle of vision and division, a nomos, that we all have in ourheads because it has been inculcated in us through a process of socialization performed in a world that was itself organized according to the division into families. This principle of construction is one of the constituent elements of our habitus, a mental structure which, having been inculcated into all minds socialized in a particular way, is both individual and collective. It is a tacit law (nomos) of perception and practice that is at the basis of the consensus on the sense of the social world (and of the word "family" in particular), the basis of common sense. Thus the prenotions of common sense and the folk categories of spontaneous sociology which, methodologically speaking, have to be called into question, may, as here, be well founded, because they help to make the reality that they describe. In the social world, words make things, because they make the consensus on the existence and the meaning of things, the common sense, the doxa accepted by all as self-evident. (p. 66-67)
Note that there is now a large literature debating Bourdieu vs. Latour