"... [W]e may be justified in explaining the emerging coalition as the result of the interaction between entire communities if an explanation of the micro-details is unnecessary because several such micro-causes would have led to a similar outcome. In the same way, a large organization may be said to be the relevant actor in the explanation of an interorganizational process if a substitution of the people occupying the specific roles in its authority structure leaves the organizational policies and its daily routines intact." (37)Isn't this the kind of reduction of agency that we've been trying to avoid - treating people and things as if they merely replicate the input without translation (that is, treating them as intermediaries rather than mediators)? Maybe my earlier claim that we could never test this was unjustified, but it's still an unfair reduction of agency - unless I'm completely misreading the paragraph.
I'm quite willing to accept non-mechanistic or even acausal explanations when the causes aren't clearly known and when those explanations add to our understanding of the world - sometimes magic is the only or best explanation we have. In this case, though, I feel that explanations that lack a mechanism do not add to our understanding, and actually subtract from the world by removing various types of agency that end up getting subsumed by the Social cause.
The mechanism I'm looking for is essentially Latour's associations. In fact, Alex's explanation for why students focus on proofreading in their peer comments despite having been instructed not to is exactly what I'm talking about. I don't see what benefit we get ascribing this to topology, redundant causation, or downward causation. Not that it's wrong to use those terms, but, please, explain what you mean rather than simply assuming that something is going on.