Emile Durkheim, The Rules of Sociological Method

The problem with Durkheim is that he knows his ontology and doesn't really care to take a straightforward stance. Halfway through, he declares that he wants to take a course midway between the nominalism of the historians and the realism of the philosophers (read Comte), but the result is hardly a position of moderation. Instead, it is of alternating extremes: the "social fact" defined in the first chapter is a perfect example of realist faith in abstraction, to the point that the society becomes more forceful, more solid, more real than any actual human. But then in the very next chapter he takes up a Baconian empiricism, leading the fight against universalisms (in particular the human universal).

Now, this isnt in itself so unreasonable as it first seems. Just because one is prepared to believe in universals does not mean that one has to accept all universals as real. But the switch in perspective is jarring and uncontextualized--it is hard to tell if Durkheim is aware of any seeming contradiction.