This is my contribution to the ongoing symposium on genetics,race, and sociological theory as well as its twin on that other blog. A quick disclaimer: I was in graduate school with J. Shiao, lead author of the paper being discussed, and we talk occasionally at conferences.
My view of the original paper is that its contribution is real but quite modest in the scheme of theory. The best way to read it is as a social-constructionist “friendly amendment” to constructivism’s tacit, yet stubborn, insistence that there is no biological basis for racial categorization. Genetic information can be used “to distinguish race/ethnicity from the existence of genetic clusters” (emphasis mine). Shiao et al. suggest that constructivist approaches to race need not cling to a strong no-genetic-clustering claim in order to maintain most of the findings of constructivism (“In sum, relatively little of the empirical explanations made by sociologists of race/ethnicity require the claim of biological nonreality traditionally associated with racial constructionism.”). In short, race is a
social reality that is historical, processual, stratified, and analytically multilevel but that is also entangled with biological inputs inherited from the geographic distribution of humans in genetic watersheds over the past 50,000 years.
While I’m no fan of genetic essentialism, I don’t think that’s what’s actually going on in the Shiao et al. article, and overall I find the critiques in the special issue quite disappointing because by and large they respond reflexively to something else instead of engaging the article’s actual contents. I actually think the most important criticism of Shiao et al. is that it’s not really all that important of a finding: the idea that minor, generally meaningless, and ancient genetic variations produce phenotypes that then become inputs to the social construction of race and ethnicity is a minor correction to social constructionism. It becomes important enough for an article in ST because of the sheer symbolic importance of race and the reflexive anti-geneticism in the field. And the character of much of the responses provide further evidence that the objections are to the symbolic affront of the article instead of to its content.
The most serious critique is the one by Ann Morning. Morning does acknowledge the modesty of Shiao’s contribution at one point:
To the extent that racial assignments incorporate information (alongside beliefs) about individuals’ phenotypic characteristics or geographic origins, they can be said to be informed by (or correlated with) biology. If that is the “biological reality” of race that Shiao et al. (2012) wish to emphasize, it does not represent a novel contribution—much less a challenge—to the constructionist account.
…but from the text of the original, it’s clear that that is the contribution offered: that “clinal classes simply summarize a potential input to the social definition of racial/ethnic categories.” And from the content of all three responses, it’s clear that even this very modest claim evokes a very strong reaction-formation.
Dan praises Morning’s critique for bringing to bear insights from STS on the scientific work deployed in Shiao et al. I find that element much less convincing. While Morning is certainly right that “approximations of the natural world are always shot through with the social,” the implication of this use of STS is that scientific evidence therefore has no evidentiary value whatsoever. Another way of saying this: is there any way to falsify Morning’s position? Could we imagine any evidence that would satisfactorily overcome the “shot through with the social” concern and thereby demonstrate that, indeed, racial categorization is “entangled with biological inputs inherited from… genetic watersheds”? Morning’s theoretical critique strongly implies that no such evidence is even possible, although her later technical critique of STRUCTURE and PCA seem to entail the opposite (you can’t claim something is a “statistical artifact” unless you can imagine something else out there that’s not an artifact!).
Morning writes: “at issue is not the mapping but the construction of human clusters.” To translate that claim into genetic concepts, it seems to me, we would expect no discernible genetic clustering among human populations, or at least that any discernible clusters should be completely uncorrelated with commonly-understood racial and ethnic divisions. In a sense, that is the “null hypothesis” Shiao et al. are arguing that we should reject. But there’s no way to assess that hypothesis if we accept Morning’s strong interpretation of STS.
Turning to Fujimura et al., consider this shocking paragraph:
Despite their claims to the contrary, Shiao et al. (2012) are using recent human genetic research to reintroduce the age-old construct of biological races, just renamed as “clinal classes.” Shiao et al.’s interpretation of genetic clustering research opens the possibility that readers will misunderstand the clustering technologies and the genetic evidence.
This is the epitome of a straw-man argument. The original article doesn’t say something Fujimura et al. want to argue against, so they just state that it must say that, even though it doesn’t. Then the criticism they offer is that somebody else might misunderstand the evidence–not that the actual article is wrong, but that it might be misinterpreted! They proceed to offer a summary of genetic variation that is absolutely consistent with the claims in Shiao et al.:
our species is not subdivided into discrete, genetically distinct, and biologically homogeneous racial or continental groups. Instead, human populations tend to be most genetically similar to others who live nearby, irrespective of continental boundaries, and the degree of genetic similarity between populations is inversely correlated with geographic distance
Clearly, the thing they’re arguing against isn’t actually Shiao et al.’s article. It’s a tradition of bad scholarship with an ugly heritage and morally bankrupt motivations that Shiao et al. reminds them of. But the fact that Shiao et al. reminds them of this tradition is not sufficient for them to hang that whole tradition on Shiao et al.
The disappointment grows as we consider HoSang’s article, which is a morass of ad hominem arguments, claims to authority, and guilt-by-association. There’s very little even of relevance in this response. Shiao’s response strikes me as entirely apt:
his critique of our bounded nature approach bears little resemblance to its actual content or theoretical purpose…. Frankly, I do not recognize our article in his list of its supposed claims.
I don’t either. And while HoSang’s is the most guilty of these, all three of the responses go well beyond the actual material to criticize the effort, the tradition, or the (predicted) interpretation. That’s too bad. In my view the implied null hypothesis is almost certainly wrong (as Fujimura et al. tacitly agree), but the fact of clinal clustering has rather little actual importance to the historical, social, and lived experience of race and racial classification.
 Interestingly this contradiction–a variant of the “borrowed kettle” problem–directly mirrors Rick Biernacki’s simultaneous theoretical rejection and methodological criticism of coding.
 Which is all the more curious given that HoSang is a colleague of Shiao’s at Oregon. I wonder what encounters in the faculty lounge are like.