Most philosophical work on race concentrates on two questions. The first is the question of whether race is real. Are there really kinds of people corresponding to racial categories like “white” and “black,” or is this merely an illusion? Suppose that races are real.

This takes us to the second question: if races are real, what makes them real? Races might be biologically real (analogous, perhaps, to breeds of dog). Alternatively, they might be real for purely social reasons. After all, money is real, and so are days of the week, but their ostensible reality depends on our social conventions. Perhaps racial categories like “black” and “white” are more like Mondays and Tuesdays than they are like beagles and corgis.

I’m going to concentrate on the biological question. I think that this is important because most people think that it’s just plain obvious that race is something that’s rooted in people’s biological make-up. I stress this because (in my view) the whole point of philosophizing about race is to contribute to the ongoing struggle for social justice. Getting clear about what race is (or, more accurately, what it isn’t) means getting rid false, oppressive, ideologically-infused conceptions and replacing them with a more accurate account of human diversity.

In particular, I want to tackle the issue of whether genetics can settle the question of whether races are real divisions of the human family. Rather than going over scientific objections to the claim that race is, in some sense, “in the genes,” I’m going to approach this issue from a philosophical perspective. I’m going to try to convince you that there’s something fundamentally confused about attempts to use genetics to prove the reality of race.

So, here goes….

Let’s begin with the ordinary, vernacular notion of race. The idea that there are human races is so stubbornly entrenched in our thinking that it’s difficult to recognize that we don’t see race. What we see is a profusion of biological diversity. We see that people come in different shapes, sizes, and colors. Some have very dark skin, and others are pale as ivory. Still others are every shade in between. Some have tightly curled hair, and others have straight hair. Some have eyelids with an epicanthic fold while others don’t.

Race isn’t the same thing as diversity. “Diversity” is a word for these kinds of variation in appearance that we observe. In contrast, “race” is a theoretical concept—a concept that we use to make sense of these variations. When I say that race a theoretical concept, I’m not saying that it’s a formal scientific theory (the sort that we learn about in scientific textbooks). Race is a folk-theory of diversity. Folk-theories are informal theories—theories that we imbibe just by being part of cultures that endorse them. We learn them at our mother’s knee, from TV and cinema, and from casual interactions with others. Because such theories are thoroughly ingrained in our ordinary ways of thinking—so taken for granted in everyday life—it’s hard to recognize that they are theories.  

The folk-theory of race has three basic components.  

First, it states that there are just a few pure human types (races) that originated in certain geographical regions. Focusing just on skin color for the moment, the folk theory has it that there are purely black people (from sub-Saharan Africa) and purely white people (from northern Europe). There are also brown people who are darker than pure white and lighter than pure black. These brown people came about through black people interbreeding with white people. The folk theory applies the same general pattern of explanation to every other racialized trait (such as hair texture or facial morphology).  

Second, it states that race isn’t just about how people look. It’s about the kind of people that they are. Even though each race is supposed to have a certain stereotypical appearance, the folk-theory allows that it’s possible for members of one race to “pass” as members of another race in virtue of being outwardly indistinguishable from typical members of that race. If race boiled down to having a certain sort of appearance, passing wouldn’t be possible. According to the folk-theory, a person’s appearance is supposed to indicate their race rather than constituting it (just as, for example, running a fever is a symptom of having the flu but isn’t what it is to have the flu).   

Third, the folk-theory states that being a member of a certain race means that one possesses a bunch of inclinations, preferences, abilities, virtues, and vices that are stereotypically associated with that race. These characteristics are imagined to be rooted in some “deep” inner property that’s shared by only and all members of the race. Philosophers refer to this as the idea of a racial essence—a kind of inner blackness, or whiteness, or Jewishness, or Asian-ness that’s the causal bedrock of all the other racial traits.

It used to be thought that the racial essence is located in the blood (hence, the notorious “one-drop rule” of racial purity). Nowadays, this silly idea has been largely replaced with the notion that race is located in one’s genes. People who don’t know the basic science, but who have the vague conviction that race in our DNA, often believe that there are genes “for” race—that is, genes that only and all members of a given race possess. This isn’t true and no scientifically literate person gives it any credence. However, there is another, more sophisticated strategy for to appealing to genetics to underwrite race. This second strategy focuses on analyzing gene frequencies in human populations. To explain how it works, I’ll first need to present a little background information.  

All human beings are practically identical at the genetic level. However, there are some places on our genome where two or more alternative “versions” of a gene can be present. These variants are called alleles, and they account for many of the differences that we can observe between people. For example, the fact that one person has blonde hair and another person has black hair is determined by a difference in alleles located at the sites in the genome that control hair pigmentation.

Alleles aren’t randomly distributed throughout the world. Some of them occur more frequently in certain populations than they do in others. When several alleles occur frequently in a population this is called a “cluster” of alleles. Genetic researchers use a computer program called STRUCTURE to identify these clusters. They input genetic data sampled from various populations and then tell STRUCTURE to find a specified number of clusters in the data (the exact number is up to the researcher). STRUCTURE then analyzes the data and sorts it into the specified number of clusters. It’s important to understand that STRUCTURE doesn’t identify alleles that are unique to the population. Rather, it paints a statistical picture of allelic frequencies in that population—a picture of what might (very loosely) be called the population’s genotype. Some individual members of the population will have a genotype that conforms precisely to this picture, but the clusters identified by STRUCTURE aren’t meant to represent the genotype of individual members of the population. They are statistical constructs representing allele frequencies in the population considered as a whole.

What does all this have to do with race?  

If STRUCTURE is provided with genetic data sampled from individuals worldwide and  told to partition this data into five clusters, it sorts the data into clusters whose distributions correspond to the five “continental races”—whites, blacks, Native Americans, East Asians, and Australasians.  

What does this show?  

It’s easy to leap to the conclusion that it proves that the everyday notion of race is vindicated by science. But that’s not warranted. To see why, we’ve first got to consider a more general philosophical issue: the distinction between characteristics that belong to individuals and those that belong to populations.

Baldness is an example of a characteristic that only individuals have. There are lots of people who are bald, but baldness isn’t a characteristic of any group of people. Speaking loosely, one might call the sum total of bald people the “bald population”, but it’s not really the population that’s bald—it’s the individuals that make up the population who are bald. To be bald, a thing has got to have to head, right? So anything that doesn’t have a head can’t be bald. Individuals have heads. Populations don’t. So, a population composed entirely of bald people isn’t itself bald. Or consider the population consisting of American philosophers. It would be ludicrous to say that this population is an American philosopher!

Other characteristics belong to populations but never to individuals. According to the United States Census Bureau, the median age for American women in the year 2012 was 38.1 years. It would obviously be ridiculous to ask what the median age of any individual American woman was in the year 2012. Individuals don’t have median ages: only populations do. Of course, an individual woman might be 38.1 years old. But the fact that her age conforms exactly to the statistical median doesn’t mean that she has a median age! Statistical constructs like “median age” give us information about whole populations but, very importantly, don’t tell us anything about individual members of that population.

To say that a population is bald, or that an individual person has a median age, would be to commit what philosophers call a category mistake. The British philosopher Gilbert Ryle invented the term “category mistake” to describe a particular sort of confusion that people sometimes fall into. A person commits a category mistake when she attributes a characteristic to a thing that things of that kind just can’t have. For I’d be committing a category mistake if I were to say that Thursdays are green, because days of the week aren’t the kind of thing that can be colored, or if I were to say that the number 37 was in my kitchen last night, because numbers aren’t the sorts of things that can occupy spatiotemporal locations.

Now let’s use these conceptual tools to examine the notion of race.  

First off, it’s clear that race is supposed to be a characteristic belonging to individuals, otherwise it wouldn’t make any sense to say of anyone that she is white, or black, or Asian, or whatever. Can race can also be a characteristic of populations? At first glance, the answer seems to be “yes.” There doesn’t seem to be anything conceptually wrong with expressions like “the white race.” But let’s look a little more closely. When people use the term “white race” they don’t literally mean to say that there’s a population that has fair skin. This can’t be what’s meant, because populations don’t have skin. It’s the individuals that comprise populations that have skin. It’s clear that “the white race” should be understood as a sort of shorthand expression meaning something like “the population consisting entirely of fair skinned people.”  

This simple point highlights a deep problem confronting any attempt to use facts about allele frequencies to justify race. If race is a characteristic of individuals then allele frequencies can underpin race only if allele frequencies are characteristics of individuals. But we’ve already seen that allele frequencies aren’t characteristics of individuals. They’re statistical features of populations. It follows that allele frequencies can’t be equated with race.  

To conclude, let’s recap. We’ve seen that a person’s race can’t be reduced to their bodily form (phenotype) because that’s inconsistent with the phenomenon of passing. We’ve also seen that race can’t be located in a person’s DNA (genotype) because there aren’t any genes that are specific to races and because equating a person’s race with allelic frequencies entangles one in a category mistake. Since there aren’t any other options for grounding race in biology (genotype and phenotype are all that there is, biologically speaking), we have to conclude that race isn’t biologically real.