One of the more convincing arguments for the existence of social facts that lie above the level of individual actors is the social reality of rules and ascriptive identities. Bob and Alice are married by Reverend Green at 7 pm, July 1, 2015. The social fact that Bob and Alice are now married is not simply a concatenation of facts about their previous motions, beliefs, and utterances. Rather, it depends also on several trans-individual circumstances: first, that their behaviors and performances conform to a set of legal rules governing marriage (e.g., that neither was married at the time of their marriage to each other, or that they had secured a valid marriage license from the county clerk); and second, that various actors in the event possess a legal identity and qualification that transcend the psychological and observational properties they possess. (Reverend Green is in fact a legally qualified agent of a denomination that gives him the legal authority to perform the act of marriage between two qualified adults.) If Bob has permanently forgotten his earlier marriage in a moment of intoxication to Francine, or if Reverend Green is an imposter, then the correct performance of each of the actions of the ceremony nonetheless fails to secure the legal act of “marriage”. Bob and Alice are not married if these prior conditions are not satisfied. So the social fact that Bob and Alice are married does not depend exclusively on their performance of a specific set of actions and utterances.
Is this kind of example a compelling refutation of the thesis of ontological individualism (as Brian Epstein believes it is; link)? John Searle thinks that facts like these are fundamentally important in the social world; he refers to them as “status functions” (link). And Epstein’s central examples of supra-individual social facts have to do with membership and ascriptive status. However, several considerations suggest to me that the logical status of rules and ascriptions does not have a lot of importance for our understanding of the ontology of the social world.
First, ascriptive properties are ontologically peculiar. They are dependent upon presuppositions and implicatures that cannot be fully validated in the present. Consider the contrast between these two statements about Song Taizu, founder of the Song Dynasty: “Song was a military and political mastermind” and “Song was legitimate emperor of China.” The former statement is a factual statement about Song’s embodied characteristics and talents. The latter is a complex historical statement with debatable presuppositions. The truth of the statement turns on our interpretation of the legal status of the seven-year-old “Emperor” whom he replaced. It is an historical fact that Song ruled long and effectively as chief executive; it is a legal abstraction to assert that he was “legitimate emperor”.
Second, it is clear that systems of rules have microfoundations if they are causally influential. There are individuals and concrete institutions who convey and interpret the rules; there are prosecutors who take offenders to task; there are libraries of legal codes and supporting interpretations that constitute the ultimate standard of adjudication when rules and behavior come into conflict. And individuals have (imperfect) grasp of the systems of rules within which they live and act — including the rule that specifies that ignorance is no excuse for breach of law. So it is in fact feasible to sketch out the way that a system of law or a set of normative rules acquires social reality and becomes capable of affecting behavior.
Most fundamentally, I would like to argue that our interest is not in social facts simpliciter, but in facts that have causal and behavioral consequences. We want to know how social agglomerates behave, and in order to explain these kinds of facts, we need to know how the actors who make them up think, deliberate, and act. Whether Alice and Bob are really married is irrelevant to their behavior and that of the individuals who surround them. Instead, what matters is how they and others represent themselves. So the behaviorally relevant question is this: do Alice, Bob, Reverend Green, and the others with whom they interact believe that they are married? So the behaviorally relevant content of “x is married to y” is restricted to the beliefs and attitudes of the individuals involved — not the legalistic question of whether their marriage satisfied current marriage laws.
To be sure, if a reasonable doubt is raised about the legal validity of their marriage, then their beliefs (and those of others) will change. Assuming they understand marriage in the same way as we do — “two rationally competent individuals have undertaken legally specified commitments to each other, through a procedurally qualified enactment” — then doubts about the presuppositions will lead them to recalculate their current beliefs and status as well. They will now behave differently than they would have behaved absent the reasonable doubts. But what is causally active here is not the fact that they were not legally married after all; it is their knowledge that they were not legally married.
So is the fact that Bob and Alice are really married a social fact? Or is it sufficient to refer to the fact that they and their neighbors and family believe that they are married in order to explain their behavior? In other words, is it the logical fact or the epistemic fact that does the causal work? I think the latter is the case, and that the purely ascriptive and procedural fact is not itself causally powerful. So we might turn the tables on Epstein and Searle, and consider the idea that only those social properties that have appropriate foundations at the level of socially situated individuals should be counted as real social properties.