Roughly speaking, I think there are three kinds of social explanation. I mean “explanation” in a very thick sense; an explanation is an account of why some phenomenon is the way it is, grounded in some kind of theory that could be used to explain other phenomena as well. To say there are three kinds of social explanation is roughly equivalent to saying there are three ways to model social processes.
The first of these kind of social explanation is functionalism. This explains some social phenomenon in terms of the purpose that it serves. Generally speaking, fulfilling this purpose is seen as necessary for the survival or continuation of the phenomenon. Maybe it simply is the continued survival of the social organism that is its purpose. A kind of agency, though probably very limited, is ascribed to the entire social process. The activity internal to the process is then explained by the purpose that it serves.
The second kind of social explanation is politics. Political explanations focus on the agencies of the participants within the social system and reject the unifying agency of the whole. Explanations based on class conflict or personal ambition are political explanations. Political explanations of social organization make it out to be the result of a complex of incentives and activity. Where there is social regularity, it is because of the political interests of some of its participants in the continuation of the organization.
The third kind of social explanation is hardly an explanation at all. It is explanation by chaos. This sort of explanation is quite rare, as it does not provide much of the psychological satisfaction we like from explanations. I mention it here because I think it is an underutilized mode of explanation. In large populations, much of the activity that happens will do so by chance. Even large organizations may form according to stochastic principles that do not depend on any real kind of coordinated or purposeful effort.
It is important to consider chaotic explanation of social processes when we consider the limits of political expertise. If we have a low opinion of any particular person’s ability to understand their social environment and act strategically, then we must accept that much of their “politically” motivated actions will be based on misconceptions and therefore be, in an objective sense, random. At this point political explanations become facile, and social regularity has to be explained either in terms of the ability of social organizations qua organizations to survive, or the organization must be explained in a deflationary way: i.e., that the organization is not really there, but just in the eye of the beholder.