Sunday, March 09, 2008

The shape of human variation   posted by Razib @ 3/09/2008 08:21:00 PM

Most of you probably know about the "species problem." The short of it is that even though the level of the species is probably the most justifiable one within the hierarchy of taxonomic systems (as opposed to say genus or order), it is not a cut & dried category. I tend to agree with evolgen that the nature of the species concept you use is going to be guided by instrumental concerns. If you are are focused upon taxonomy I have no doubt that the phylogenetic species concept is the bomb. On the other hand, if you are an evolutionary geneticist interested in speciation the details of the structure of the tree of life is less important. You would be more interested in how the branching occurs, in which case the biological species concept and its cousins are of more relevance. During the 18th and 19th century some taxonomists argued that their discipline had a claim to being the Queen of Sciences precisely because it was window into the mind and intentions of God's Creation. For these pre-evolutionary thinkers there was a religious and metaphysical significance in species; inappropriate classification would have distorted God's intent and obscured the grand beauty of his plan. Today we don't species aren't loaded with some metaphysical importance; our conception of the world does not hinge upon the battles between lumpers & splitters.

Which brings me to a lower taxonomical category; that of race. Or whatever you want to call it, you know what I mean! Obviously over the years on this weblog we've put the spotlight on specific issues such as Lewontin's Fallacy, which though seemingly abstruse can resolve confusions and misinformation being perpetuated in other domains. We've also pushed into more controversial territory like between group differences in behavioral tendencies & IQ, as well as less radioactive topics such as tissue matching problems across populations for transplantation. There's a lot of ground over the past 5+ years. But I believe there are 4 primary dimensions of variation, which though related (they aren't really orthogonal), get at different concerns.

1) First, there's the phylogenetic/total genome content angle. This is simply a form of Steve Sailer's race is an extended family argument. The tens of thousands of genes an individual has all exhibit their own distinct phylogenies; but they're not totally independent as a practical matter. We don't live in a perfectly panmictic world, population substructure is real. Just open up History and Geography of Human Genes; the classical autosomal markers exhibit correlations, which allow one to make assertions about population histories and relationships.

2) Second, there are the functional loci under selection. The story of lactase persistence in Europe & Asia is the best illustration of what I'm talking about here. From the North Sea to the Punjab a haplotype which likely arose around the Volga region has swept to high frequency within the last 10,000 years because of gene-culture coevolution. If you look at most other markers Middle Eastern populations will be closer to Europeans than peoples from the Punjab in northwest India (see the Fst values in History and Geography of Human Genes for example). But not on this locus. In fact, on this locus Danes are closer to Punjabis than they are to Sicilians, on average!1 Though most recently selected alleles will not break the cladograms you derive from neutral markers to ascertain phylogenetic relationships of populations, there will be plenty of peculiarities on the margins, and I have argued that these deviations are not trivial to our understanding of the history of the shape of human variation. Who can deny that the nature of genetic variation (or lack of) of indigenous New World groups was not of particular importance in relation to Old World people as a whole, as opposed to the undoubted phylogenetic reality that New World native peoples were a subset of East Asian Siberian populations.

3) This brings me to the dimension of salient phenotypic traits. No need for complex exposition of what I mean about this. Insofar as one looks at total genome content Melanesians are closer to the peoples of East Asia than they are to those of Africa. But if most people in the world were shown portraits of individuals from these three groups, and asked to generate an outgroup, I think most would assume that East Asians differed from the other two groups more than either did from each other. This is not to say that Melanesians and Africans look the same, or that Melanesians and Africans do not exhibit a great deal of within group variation in physical appearance, it is simply that the human brain strong affected by particular characteristics when generating classifications. Folk biology has a universal rationale, and is shaped by innate biases. Skin is our largest organ, and it is something that we as humans notice because there are strong adaptive reasons to scrutinize the skins of our conspecifics (fitness, disease, a rough & ready ascertainment of age). I agree with Steve Sailer that the monomanical focus on skin color which reigns in the American social discourse on race is a bit ridiculous, but I also think it is entirely expected and reasonable that skin color would loom large in any folk racial classification system. Many of the peoples which Europeans during the Age of Discovery encountered were referred to as "blacks." They were very different from each other, and Europeans recognized this. Today we know that all the non-African blacks are genetically closer to Europeans than they are to African blacks (who are characterized by a great deal of within group population substructure as well), but it is no surprise that a skin color terminology came to the fore. The West & Central Asian Muslims who ruled India during the medieval and early modern periods referred to themselves as white, and the natives as black. The Chinese would sometimes refer to the Khmer peoples as black as well, because that was a salient contrast. Of course other traits were recognized, and there is also a long tradition of Europeans suggesting that the peoples of South Asia were not truly black, but rather a very pigmented form of their own kind due to similarities of hair form and facial features. This sort of counter-argument was aided by the fact that a non-trivial proportion of South Asians even exhibit a brunette white complexion (usually along the northwest fringes).

4) This brings me to the last of the major ways in which we perceive human variation, and that is through the socially biased and constrained lens. Obviously this is affected by and contingent upon #3 to a great deal, but the boundary conditions are illustrative. In high school I was taking calculus with a friend whose mother was Scottish American and father was a Palestinian Arab. His name was an Arab one, but his physical appearance is by any definition "white." His skin is white, his hair is brown, his eyes hazel, and his features favored his Scottish as opposed to Levantine side (I had a 1/4 Lebanese friend who had a more recognizable Arab visage). Another friend, who was of vanilla Anglo origin as most of my classmates were, observed that we were the only two non-whites in the class, referring to myself and my aforementioned friend. Here's the irony: by any standard my Anglo friend was darker than my non-white friend, he had dark brown eyes, dark brown hair, and less of a pink pallor to his complexion. I recall kind of laughing at that assertion, and the teacher bitched me out about being disruptive and I laughed again. The point here is that the idea that taxonomical perceptions are colored by power hierarchies is not totally incorrect! In fact, it seems trivially obvious. I've been reading a fair amount on the Chinese & Japanese interaction with Western powers between 1500-1800. It is interesting that in the earlier commentaries on Asian peoples many Europeans observed that the Chinese and Japanese were white, unlike the peoples of South or Southeast Asia. After the last wave of Sinophilia receded after the mid-18th century there was a noted shift toward a perception of East Asians as being non-white. In other words, the whiteness of East Asian peoples (even if that whiteness was a different kind) inversely tracked the European sense of superiority to them.

I think #1 and #4 are pretty easy to understand. Thinking of race as an extended family simply co-opts our native intuitions about genealogical relations. In other words, it's an extension of preexistent software. As far as #4 goes, there's enough pointers from the current academic dispensation that we can comprehend the nuances. The main point is to map the dynamics more accurately upon reality and not elide the complexities inherent within socially constructed categories which are only partly informed by folk biology. A tendency to pretend as if whites and non-whites are the only two relevant categories prevents the acknowledgment of the realities of how various groups relate to each other (e.g., the social science data which suggests most American non-white groups prefer whites to other minorities).2 Since #3 is derived pretty directly from folk biology and some gross social cues it isn't too difficult. It's the practical definition of human variation which people in the United States have in their head.

#2 is a tricky one. I think it's really the most interesting one at this point, but I don't know how to really communicate it to people outside of a more technical framework. Alluding to selective sweeps and QTLs of large effect seems kind of important. #3 is actually a visible subset of this, and so I think when it comes to the general concept of how a functional locus may not reflect the evolutionary history of the whole genome that's a good place to start. LCT is also important, it's very well elucidated in all sorts of ways and so you can speak with confidence. Skin color is also a good case because people are generally aware of the trait and the general outline of inheritance patterns, and its putative adaptive significance. A lot of the work on recent human evolution is related to this dimension, so I think it's important for intelligent people to keep this angle in mind.

The main issue is not to conflate the various facets of human variation. There are ideological reasons that people will privilege one or the other, but in terms of utility that depends on where you stand. I'm obviously pretty interested in questions and issues where #2, to some extent #1 and #3, are very operationally the relevant background definitions. #4 is not something I deny, but it's not particularly interesting to me, nor does it speak to the questions I'm focused on. #4 has come to the fore with the ubiquity of international trade and travel, along with a host of social and political movements. That's all real, and it's all significant, but it isn't really normally part of my brief. It intersects with my interests only a specific case of psychological issues, or an indicator of social dynamics.

1 - By this I simply mean that if you draw a random Sicilian, Dane and Punjabi, and are asked to predict which group will be the outgroup when it comes to the most recent common ancestor on the locus LCT, it would be the Sicilian. On most of the other loci, it would be the Punjabi.

2 - I recall listening to a radio show about tensions between South Asians and people of black African origin. Several white callers were shocked and outraged, and asserted that such tension and racial antogonism between non-white groups was by definition ludicrous and incoherent. Both groups were non-white, ergo, they had common interests by definition. There's a little bit of the sense that some whites view non-whites as objects which only operate within the bounds of some deterministic framework, as opposed to conscious agents who act in response to specific conditions and their own perceived self-interest.