The most recent issue of Sociological Theory contains a four part symposium on the genetics of race. More specifically, three of the pieces are responses to a 2012 article in ST by Shiao et al., The Genomic Challenge to the Social Construction of Race, and the last is a reply by Shiao. The debate is an important one for both sociology and the broader public. Every advance in biology and genetics seems to trigger a new round of scientists, and social scientists, trying to justify widely-held beliefs about race in essentialist terms. Shiao et al. offer a new, sophisticated argument in a very long tradition.
Fortunately, sociology – especially sociology informed by the science studies (STS) tradition – is well-equipped to engage with the guts of the genetics underlying Shiao et al’s claims. Ann Morning’s response walks through both the logic of constructionist arguments – which do not, in fact, ignore biology – and then turns to the research in genetics that Shiao et al. mobilize to make their claims. In the tradition of the best STS work, Morning carefully documents how the evidence showing that humans cluster into a small number of genetic groupings is as much an artifact of scientists’ beliefs about variation as it is a finding of some objective research process. Those beliefs are become “self-vindicating” in Hacking’s (1992) terms through issues of sample selection (size and geography), the kinds of genetic data collected, assumptions about human evolution, and even statistical assumptions about the number of clusters to be identified using various factor analysis-like techniques. Here’s a nice example of that last issue:
One of the most powerful ways in which geneticists’ racial preconceptions can shape their analyses of human population structure lies in their assumptions about which clusters and/or how many characterize our species. … For one thing, it is common for scientists to test the model fit of only a few possible options for the number of clusters (K), ranging in the single digits (e.g., K = 2 through 6 in Rosenberg et al. 2005), which is consistent with contemporary notions of the number of races in our species; the U.S. federal classification system, for example, includes five races at present (i.e., white, black, American Indian or Alaska Native, Asian, Native Hawaiian or other Pacific Islander; see U.S. Office of Management and Budget 1997). Paschou et al. (2007) manually set the number of clusters to four, although they base their choice on estimates of how much more an additional component would explain (p. 1675).
In a striking example, Pritchard et al.’s (2000) analysis suggests that their genotype data from Africans and Europeans would be better described as forming three, four, or five clusters than merely two. In other words, a racial “black/white” binary structure is not supported by their method. Yet they go on to conclude that “our methods find it quite easy to separate the two continental groups into the correct clusters” (Pritchard et al. 2000:952), reverting to a racial binary that is not indicated by their statistical analysis. (Morning 2014: 199-200)
Morning also shows (in line with her earlier work) how most geneticists themselves reject the argument that their methods justify a biological conception of race. Highly recommended for the cutting edge of both essentialist arguments about race, and sociological refutations of those arguments.