Sunday, November 09, 2008
It is good news that the social scientists are finally starting to think seriously about genetic variation and social behavior. This latest Science paper by UCSD political scientists James Fowler and Darren Schreiber is a case in point. A key argument they make in the paper is that political scientists cannot ignore behavior genetics because genetic variation is so ubiquitous also in political "phenotypes". (Of course, it was Alford, Funk and Hibbing who got the ball rolling, though Nick Martin had reported pretty much the same findings some twenty years earlier, but not in a political science journal). The paper then goes on to argue that an evolutionarily and biologically informed political science is the way of the future.
Sympathetic though I am to the behavior (and molecular) genetic work currently being undertaken by political scientists, I still think that it is unfortunate that it has no serious critics within the political science community. Evan Charney at Duke University has made some attempts to fill the void with a pretty standard attack on twin studies, but let us be charitable and say that he has not been particularly successful. The best argument against heritability estimates is not that they are wrong, but that they are not as structurally informative as is sometimes supposed. This is an argument that needs to be taken seriously. We can quibble about functional form and independence assumptions, the equal environment assumption, and the blatantly false assumption that all gene action is additive. In the end, it will not alter the fact that there is moderate genetic variation in virtually every trait you study.
The deeper question is this: what does it mean that a complex trait such as voting, the strength of partisanship, or attittudes toward abortion, say, are heritable, beyond the obvious but important implication that we can't interpret parent-child associations in environmental terms? It seems like a stretch to expect that there are genes with a proximal relationship to voting behavior, though of course the indirect effects of the genome can be developed through genome-level influences on intelligence and personality which in turn, ultimately, affect voting. But what are these mechanisms? Unless we know what they are, can heritability estimates be used to discipline any of the political science models?
These are all fascinating questions and Fowler and colleagues are only beginning to uncover the answers. I anticipate that Fowler and his partners in crime will continue to leave a trail of evidence from which we can build an even stronger case for a political science which does not make assumptions that are at odds with stylized facts from behavior genetics. Or, for that matter, at odds with facts derived from any of the other scientific disciplines from which the "genopolitics" crowd draw inspiration.