There is a notion that has been circulating for as long as I can remember in every piece of writing in the humanities on race and racism to the effect that the concept of race has been scientifically discredited. The most recent example is a piece in last weekend’s New York Times, which contains the statement “Since the mid-20th century no mainstream scientist has considered race a biologically significant category.” In a sense this is true: biologists are not in the habit of using the word “race” in attention-attracting ways. The question is, in what sense has the concept of race been scientifically discredited? The answer proves to be more complicated than at first appears.
It is easy to see why regarding race as scientifically discredited would be useful morally. Racism involves negative value judgments of individuals based on their membership in a race. These judgments receive support from factual claims about racial characteristics. These factual claims, in turn, have to be about something that exists, otherwise the chain of reasoning falls apart. However, notice in passing that the chain might fall apart at any, or indeed at several, of the stages leading to the negative judgment about the individual. One might be correct in claiming that races exist, and be wrong about the factual claims about a race (for example, about average intelligence). One might be correct about a factual claim about an average trait while being wrong about an inference from the average to the trait in a particular case. Lastly, one might make the wrong moral judgment in response to a factual claim about an individual. The claim that the concept of race has been scientifically discredited seems to invoke error at all these stages except the last, since the notion that a moral judgment has been scientifically discredited (as opposed to the facts upon which it is based) is problematic. However, when people make claims like the aforementioned quote, they generally mean not that factual claims about races are false, but that racial predicates don’t map onto anything real (apart from social facts). If this were true, it would cut off racism at the knees quite effectively. The tagline that links to the aforementioned article expresses this with admirable clarity: “When we learned that witches did not exist, we threw out the category. So why do we cling to the discredited notion of race?” However, when the notion that race is scientifically discredited is presented, this is seldom explained.
Up until quite recently, the basic mechanisms of heredity were not understood at all. Two crucial discoveries, that of the rules of Mendelian genetics, and that of the structure of DNA, changed all that. Before that understanding emerged, one popular view, extending all the way back to antiquity, was that every species possessed an Aristotelian essence that existed within each individual, which determined the normal characteristics of members of the species, explained why our classification of individuals into that species was correct, and which was somehow involved in the transmission of heredity. These essences formed nested hierarchies upward (genera) and downward (subspecies). On this account, a possible view is that races are subspecies, determined by Aristotelian sub-essences.
One thing that “discredited” means here, apparently, is that there are no such sub-essences. If this means that there is no item present in each individual by virtue of which they are members of a race, that is both true and important. Rather, what there are are self-replicating systems (organisms) assembled from information coded in DNA sequences, and populations of such systems. What defines such populations are continua of genetic similarity which can be characterized statistically. However, it is important to note that this particular discrediting applies to species as well. There is no metaphysical item each member of the same species possesses by virtue of which it is a member either.
This point, however, tends to get confused with a different point, which is that, whereas classification by species is biologically meaningful, classification by race is not. This point is not about the existence or non-existence of racial essences, but about the character of the underlying continua, and the arguments that have been made to establish this claim are of a different sort. Crudely, the idea here is that each individual is a unique constellation of traits (phenotypes) and underlying genetic material that codes for them, and whether or not any two individuals are similar or dissimilar depends upon which traits are focused on. For example, if I compare a brown-eyed, dark-skinned person to a brown-eyed light-skinned person, they will be similar as to eye-color, more so than a brown-eyed light-skinned person and a blue-eyed light skinned person. However, if the comparison was by skin color rather than eye color, the similarities and dissimilarities fall out differently. Which group the individual is a member of depends on which trait is chosen as the basis. If several such traits are chosen and it turns out that there is more genetic variability within populations we folk-classify as a race than there is between then, racial classification is said to be not biologically meaningful. A sophisticated version of this argument was made by the famous biologist Richard Lewontin, in his 1972 paper “The Apportionment of Human Diversity.”
Lewontin’s analysis was subject to a powerful critique by A. W. F. Edwards in his 2003 paper “Human Genetic Diversity: Lewontin’s Fallacy.” The problem, in brief, is that not enough loci and alleles were being compared (say we compared eye color and skin color and hundreds of other traits, instead of a small number). There have been in the past decade or so several papers which, leveraging more recent data from DNA sequencing and a mathematical tool called “cluster analysis,” reveal that sets of DNA sequences do fall “naturally” into clumps based on degree of similarity and dissimilarity within and between populations, and that these DNA sequence clusters, arrived at independently of folk racial classification, happen to map on to them. In one cluster analysis paper, when the algorithm for clustering was set at K=6, the sets of DNA sequences naturally clumped into African, European, Asian, Melanesian, Native American, and an obscure mountain tribe in Pakistan called the Kalash (Noah Rosenberg, et. al., “Genetic Structure of Human Populations,” Science, vol. 298, Dec. 20, 2002.) In another study, ethnic self-identification deviated from independently determined cluster membership only 0.14% of the time (Hua Tang, et.al. “Genetic Structure, Self-Identified Race/Ethnicity, and Confounding Case-Control Association Studies,” American Journal of Human Genetics, vol. 76, 2005). In short, there are underlying biological facts (statistical frequencies) which correspond to the folk classification. To say that there is no such thing as race, rather, all there is are DNA sequences and statistical patterns populations of them fall into is not unlike saying there is no such thing as water, only underlying facts about protons, electrons and neutrons. We can choose to talk that way, if the concept of water is so freighted with bogus social significance that using the word “water” causes more harm than good. But absent such significance, we don’t say “there is no such thing as water, there is only H2O.” Rather, we say “water is H2O.”
There is of course a further question. That, however, is not a judgment about what is real, but a judgment about what similarities and dissimilarities are important, and how important. And though it is true that we have a confusing tendency to use the word “real” to mean “important” (using the phrase “The real question is not X but Y” to redirect attention from X to Y), there is another, more primary sense of the word “real.” In that sense of “real,” races (statistical clusters of DNA sequences) are real, and witches are not, whatever we choose to call them. The fact that historically, humanity has chosen to run straight to the lowest reaches of hell with that concept is, uncomfortably, a separate issue.
Postscript: I can’t really leave it there, can I? Well, here’s a proposal, and because it is a very slight modification of existing social practices, it is feasible. Let’s simply retire the word “race.” It is, to put it as I did above “freighted with bogus social significance.” However, the “scientifically discredited” meme not only is grossly misleading, but it encourages people to politicize biological inquiries into human geographical populations. But what to do with the “scientifically discrediting” business? I propose that we forge an exclusive link between the word “race” and the pre-Mendelian, pre-Watson and Crick notion of an Aristotelian racial subessence, and the large body of false empirical claims made in the 19th century and earlier about geographical populations. Then humanists can continue to say “science has discredited the concept of race” and for clarification refer, if necessary, back to Mendel, Watson and Crick, rather than referring back to Lewontin’s notion that race is not taxonomically real (since as far as we can now tell, it is). To make this even clearer, humanists should refer to the discrediting of “racial essentialism” and assert a strong analogy between this kind of discussion and gender versus sex discussions. Why? Because one can claim that gender is a social construction, that there is a kind of “essentialism” that promotes sexism, without for a minute denying the manifest fact that there are such things as biological males and females. Thus biologists can continue with their population genetics, and humanists can continue with their critique of essentialism, and not get in each other’s way. But let’s leave the Lewontin Thesis out of the discussion, because if we make the truth of Lewontin’s Thesis a central reason for eschewing racism, we thereby strengthen racism when we discover that Lewontin’s Thesis is better referred to as Lewontin’s Fallacy.