« Fandom to Obsession: Star Trek Decorating | Gene Expression Front Page | Fame & Victimization: A Toxic Brew »
February 12, 2005

"Gene expression" might matter

One of the points that seems to come up over & over again in the Larry Summers debate was that men and women are genetically the same, so the differences we observe are not genetic (males are hemizygous in that our Y doesn't have a copy of most genes and so the X has to kick in). This makes for a nice little quip, but one could point out that male and female elephant seals are the same genetically, but the former are 3 to 10 times heavier by adulthood. Such trivial observations could be made about humans in the realm of sexual dimorphism, after all, men tend to be taller on average.1 In other words, the promoters of these views assume a naive understanding of the mapping of the genome to the full development of an organism.

In fact the genome is not an exact blueprint for how the organism will develop, it is not a miniature rescaling or compressed analog record, rather it is more like a series of commands and operations that initiate the cascade toward the processes that result in the construction of the full phenotype. A standard explanation of male-female differences is that the sex hormones mediate the expression of various traits. This is the main argument in The Essential Difference, the level of testosterone looms large in almost every trait difference. The author even suggests that the level of testosterone in females is a crucial metric that might shed light on different behavorial and cognitive tendencies of women at the high end of the distribution tail. But one must remember that there is even evidence that differential expression precedes the release of hormones.

The anti-Summers camp often depicts those of us who are sympathetic to the consideration of a full range of factors as contributors in sex/gender differences as oversimplifying essentialists. Yet here we have a case where they take the simple count of the lack of genomic differences and let that serve to falsify those who argue for biological differences. Implicitly, they are taking an essentialist and static view of the genome, what you see is what you get. There is no cry and call which emphasizes the importance of epistasis and other dynamic aspects of the genome, because in this case it opens the doors up for many models in which a difference on one chromosome could result in a cascade of pathways which alters gene expression, and therefore the final phenotype.

So, either....


  • Biology is complex, we can't make heads or tails of sex differences so we shouldn't attempt any speculation.
  • There is no genetic difference, so it is obvious that there could be no great biological differences.

I happen to think that the first point is closer to the truth, though I think we can make heads or tails of sex differences, more or less. Additionally, many in the anti-Summers camp seem to prima facie assert that the sociological evidence for discrimination is simpler than the biology. I don't necessarily agree.2 Since when was sociology, at least in its subject of study, simpler than biology? Nailing the biology, from the molecular genetic level up onward to the complex phenotypic expressions we see in brain morphology is where we should start, and then we can make a fuller examination of the sociology. While biology influences sociology, sociology does not influence biology (in theory).

Addendum: In the comments to this post I suggest to John Emerson that the "Larry Summers controversy" has exposed a fissure on the Left rather than the standard Left-Right battle. That is why I tend to get annoyed when some who found that Summers' statements were false or over the top try to portray them as standard right-wing cant, after all, Summers worked for Bill Clinton and one of his major supporters, Steven Pinker, is politically a pretty standard Democrat. During my initial roundup, I also missed an interesting comment by Steve Berlin Johnson, educated at Brown in semiotics, later to pursue graduate work with an influential disciple of Jacques Derriba's, and a confirmed Democrat:


Finally, I think it's more accurate (and also more palatable) to describe the differences here in a slightly modified language: not that women are innately worse at math or engineering than men, but rather that they're innately less interested in math or engineering than men. That lack of interest starts early -- boys tend to engage more with inanimate objects, girls with people and animals -- and ultimately spirals into larger differences in skills and experiences, which no doubt get amplified by social bias. But it's not that women are challenged by these subjects; it's that they don't find them stimulating.

The tunnel vision on the dearth of women in the mathematical sciences gets tiresome after a while when considered in the context of the 15:1 ratio of males to females in prison, or the prominence of women in the humanities and social sciences and their rise in the biological sciences. If all the vectors pointed in one direction I would be concerned, but as it is, it is a lot more complex than that....

1 - There might be regions of the Y that code for sex differences.

2 - Overt discrimination, which is a function of explicit legal formulations, is obvious and easy to characterize and remedy (simply nullify the formulations), covert discrimination, which is an emergent property of various sociological variables, is not. Biology is nested within sociology because we are social & cultural animals, the first two modify and shape the latter, but the animal constraints and biases remain.

Posted by razib at 04:07 PM