Mon May 18, 2009

Shake-n-Bake Social Theory redux

I think I’m broadly on Fabio’s side when it comes to the question of the vagueness of concepts in the social sciences. I think my main caveat is that, based on the evidence, successful social science requires precisely specified concepts coupled with a willingness—perhaps elevated to a principle—to strategically ignore any amount of empirical evidence accumulated against them.

But enough trolling. Beyond the problem of vague concepts lies the question of vague argument. On the plane home this Sunday I read Jon Elster’s new book on De Tocqueville. It’s typical Elster: incisive, clever, restless, and weirdly dissatisfying. At one point he remarks that too many writers are not clear enough to be wrong. And then, in passing,

Tocqueville here relies on what I called the first law of pseudo-science, “Everything is a little bit like everything else.”

He immediately follows this up with the second law of pseudo-science: “Everything is causally related to everything else”. The latter law is, I think, a besetting vice in much of sociology, where one reaction to monocausal explanation (on the one hand) or empirical murkiness (on the other) is to emphasize the irreducible complexity of the explanation, to the point where there is really no prospect of finding any disconfirming evidence. But while Elster’s second law leads directly to disaster, the first law seems more ambiguous in its effects. For one thing, any number of innovations in social theory (or I daresay the sciences more broadly) spring from applying this law as a heuristic. Drawing on a post from almost six∗ years ago, the strategy is as follows:

Take a few basic kinds of institutions, structures or practices that can be identified across many different social contexts. There are markets, say, and there is politics. There is ritual. There are hierarchies. There are networks. There is culture. And so on. (Not all of these are the same sort of thing; that doesn’t matter at the moment.) Identify the basic features of each. Now, pick one of these and show it underpinning a setting usually taken as governed by one the others.

For example, you can say Politics is really Markets. This is Public Choice Theory. Because the market form is such a dominant feature of contemporary societies and of talk about them, applying the “x is really a market” trick to any given x is by now ubiquitous not just in theory but also often as a matter of common interpretation and even public policy, facts on the ground notwithstanding.

Markets are really Politics. This is one wing of economic sociology—Neil Fligstein’s work for instance—but also a line of left-wing economics.

Markets are really Culture. Viviana Zelizer’s The Social Meaning of Money takes the world of apparently cold-blooded economic exchange and shows how the ritual creation of social ties between people is fundamental to the nature of money. Alternatively, for a comparative approach read Frank Dobbin’s Forging Industrial Policy to see how 19th-century economic policy about Railroads took as its model different conceptions of the polity in each of France, Britain and the United States.

Organizations are really Ritual. Meyer and Rowan’s “Institutional organizations: Formal structure as myth and ceremony” makes the case.

Markets are really Hierarchies. Sounds like a tough one, but Art Stinchcombe has a go at the theory in “Contracts as Hierarchical Documents.”, and the AEA junior job “market” instantiates the practice.

Ritual is really Markets. Rational Choice theory of Religion.

Markets are really Networks. Mark Granovetter; Harrison White. Production markets are really self-reproducing interfaces created by the mutual monitoring of firms trying to find a sustainable niche in a production system.

And so on and so forth.

**

∗Jesus Christ.