Wednesday, July 20, 2005

Metaphors (analogies) we don't live by....   posted by Razib @ 7/20/2005 04:11:00 PM

In many discussions where genes and sociology intersect there is often a group of individuals who will deny that one can truly make non-trivial assertions about genetic effects over the generations, that interactional influences, whether they be gene-to-gene, gene-to-environment, or even more complex feedback loops, make talk of "heritability" null and void. I've talked about epistasis, how R.A. Fisher rejected its relevance as an evolutionary force and its role in his dispute with Sewall Right over the character of the adaptive landscape. By the mid-20th century the orthinologist and systematist Ernst Mayr expressed skepticism toward "Bean Bag Genetics," a paradigm which held that it was acceptable to treat loci as independent agents which each injected their own quanta or fraction to the variance of a particular phenotype (an additive approach). J.B.S. Haldane, an avowed Marxist, came to the defense of the English tradition of mathematical genetics that he and Fisher pioneered in a famous essay where he rebutted Mayr. My point though is not to recap the history of genetics in the 20th century, but to note that a certain stream of thinkers have a habitual tendency when engaging in public policy debates where the issue of heredity comes up to reflexively appeal to interaction and other non-additive factors to subborn the arguments of their opponents in the genetical-social realm, drawing on the same talking points made prominent by Mayr. By emphasizing the non-linear and contextual/contingent factors in the equation they attack the utility of provisional models as a guide toward making decisions. A genetic-environmental produced phenotype is a complex, almost mysterious system, which simply does not brook analysis and decomposition.

Now, let us move to the broader canvas of society as a whole. The law of unintended consequences suggests to us that social systems are filled with unknown contingencies and variables which are simply not transparent to us on first, or second, or third analysis. Some people would argue that social systems are simply irreducible, that model building and attempting to "rationalize" society is a futile endeavour, that actions have wildly unpredictable consequences. On the other hand, there are those who seem to posit a "Bean Bag Sociology," where social good A has implication B, rather than being enmeshed in a nest of interlocking relations which might be disrupted if social good A's state is changed. Let us move this to the realm of specifics. Women were given the right to vote, the South was desegregated and abortion was legalized. I point out policies which progressives tended to favor and conservatives tended to reject to illustrate that in this case the social Left is engaged in a kind of sociological Bean Baggery. In contrast, the more traditional conservatives appeal to custom and tradition because they believe that the organically developed social systems of the past are not simply the reducible sum of their parts (additive), and that "progress" on one issue may have a drastic, non-linear, effect on the society as a whole.

So you have a situation where in two domains of knowledge the parties who demand absolute certainty switch polemics. Of course, I would admit that the analogy is imperfect, genes are a rock-hard theoretical basis for evolutionary biology and the disciplines which draw from it (though I am not offering that the theories themelves are rock-hard, simply asserting that the accused jelly sits upon bedrock), while memes have yet to be properly characterized as anything beyond a metaphor. I will lay my cards on the table and say that I am cautiously optimistic about model building in both situations, though I am not a strict Fisherian in that I think there may be multiple equilibria or peaks on the landscape of genes and society, but even if I was a Fisherian, recall that his theory of evolutionary change via microevolution posited sequential fixations of loci over time, one at a time. Change occurs, simply not through dynamic gene-to-gene interactions, let alone anything like a "genome reorganization." Make of that what you will in the domain of sociology.