This is my sixth post in a series on the metaphysics of race. If you want to start at the beginning, go to the introduction to the series. Each post after that has a link at the bottom of the previous post.

Biological race realism was rejected in the mid-20th century for good reasons. The biological race view current at the time held that intellectual and moral natures were true of each race and different for each race, and once we discovered DNA that all went out the window. In the second half of the 20th century we learned that the surface-level traits we associate with each race (primarily skin color, hair type, bone structure, etc.) are really all the major variation that you find from group to group, and there is actually more genetic variation within each race than there is between races. Nearly all of the traits that are particular to certain racial groups are not all that important biologically. So at best it seemed as if the traits that might serve as biologically distinctive for each racial group are biologically arbitrary. It certainly wasn’t enough to justify thinking of such a thing as a racial essence that each member of each race supposedly had.

That led to most people who had a view about these either being anti-realists (denying that races exist) or social kind theorists (thinking races exist but are social, not biological entities). A handful of philosophers tried to revive biological notions of race in the mid-late 1990s, most notably philosophers of science Philip Kitcher and Robin Andreasen, but they didn’t win a lot of support, mostly because the groups they found that they could categorize biologically didn’t line up as closely with the groups we call races and refer to in ordinary conversations about race. (This objection was called the mismatch objection.) Kitcher himself even changed his mind in favor of a much more complex view.

Meanwhile, some actual scientists had been doing some innovative work in population genetics. I noticed this around 2006, but the work had begun to be published a few years before that. I am only aware of three philosophers who have spent much time looking at this work: Quayshawn Spencer, Michael Hardimon, and me. All of us have published reflections on its significance, and we don't all say precisely the same things, but we all think this research shows that there are biological races of a very different sort than the old biological race view believed in.

The population genetics discipline within biology has given reason to think that there are biological races. What this new research shows is that you can design a computer program to look at genealogical populations, which are populations that share more of their ancestry with each other than with people outside the group. The computer program Structure and other similar programs start with the data about which genetic material is found in which people, with no information about which groups those people belong to. Then it tries to use that data to figure out the best sub-divisions within the entire sample, arranging people into groups that are likely to have the most amount of common ancestry. The resulting groups should be genealogical populations that represent all of humanity, as long as you have a large enough sample from enough people from enough places around the world.

The results were very surprising to a scientific community who had rejected the idea of biological races. The result of these studies is that you can take a large sample of the complete human genetic code of a large enough and diverse enough sample of human beings and form them into groups, and they came up with five racial divisions – African (sub-Sahara), Eurasian (Europe plus Asia west of the Himalayas), East Asian (Asia east of the Himalayas), American (the native populations of North and South America), and Oceanian (native populations of Australia, New Zealand, Malaysia, Polynesia, Micronesia, Hawaii, and other Pacific islands).

What is so interesting about this result is that those are almost the same groups that at least one common way of dividing human beings into races has done. It’s not the only one, but it’s how the United States government has been doing it for decades, long before this kind of research had been done. Quayshawn Spencer has pointed to the U.S. government categories (from the Office of Management and Budget, because they were the first agency to do this, and everyone else followed them). He observes that the two sets of categories are pretty much the same ones, and they come from trying to identify the main continental areas that people come from, using geographical divisions as boundaries where they separate the categories. So the creators of these U.S. categories used the Sahara, because the cultures south of the Sahara are distinct from those north of it, who had been more interacting with Europeans and had mixed with them. They gave reasons for putting the divisions where they did. It was a conscious choice. That makes the OMB races socially constructed in a very clear way. It was a deliberate choice of the US government to draw those lines where they did for reasons that they explained. Yet we now have a biological method using pure science to end up with categories very close to those socially constructed categories.
​
Quayshawn Spencer concludes that human continental populations are both biological and socially constructed. We can arrive at them just using biological science, but we also have arrived at them by classifying people for the social reasons we have done.

Now Spencer also acknowledges that there are different social constructions than the OMB one, and those don’t have anywhere near the kind of biological support to count as biological races. So what if some practice takes Pacific Islanders or aboriginal Australians (who have black skin and tight, curly hair, just as Africans do) to count as black because they have the same appearance as Africans and African Americans? The only way to get such a group is to focus on biologically insignificant categories in comparison to what Structure does. But some people think about race that way, and one social construction of race does put those three genetically distinct populations into one race. Skin color and hair type are such a tiny percentage of anyone’s genome that picking those out doesn’t have any biological reason to support it. The reason you would lump those diverse genetic groups together is because people treat them similarly because of their appearance. The practice of putting such groups into a racial categoy has social reasons, because our society makes those features significant when people discriminate or make assumptions about people because of their skin color. That allows you to have socially constructed races with different boundaries than the ones the OMB uses. Those socially constructed races are not biological, he says.

But the OMB socially-constructed races are also legitimate biological categories, in his view. Those categories are socially constructed but also refer to legitimate biological entities that make sense to use in scientific research as biological classifications. That's an odd combination, but if you think back to what makes something social construction it makes more sense. A social construction is something that we might not have had to choose to draw attention to and delineate with the lines that we do.  There is a biological category that we refer to with this set of categories, but we didn't have to draw any attention to that (it actually took computer programs developed as recently as 20 years ago even to discover these groups had a biological underpinning), and it was because of choices we made as a society that we did draw attention to it.

It's also important to see that this is a much more minimal biological classification than the classic biological race view, which took there to be these racial essences that define a lot about who we are and that in their view have to do with race. Nothing in this view relies on racial essences or connections between race and intelligence or race and any particular moral behavior. The genetic differences among these populations are surface level, and yet the categories being identified are being identified because of biological reasons. They are biologically distinct, but that distinction does not amount to much of any significance. Michael Hardimon thus calls this a deflationary account of race. Races are not much of any interest, but there is a purpose in biology to talk about them. As biological categories, they do not have much significance beyond tracing out ancestry and looking at history of migrations or DNA transference across generations.

So this new biological view would argue that biology alone does not explain why certain racial groups do better or worse on standardized tests or have cultural traits that make them more or less likely to do certain things. In the worst of the classic biological race view they assigned criminality and moral traits to certain races, blaming higher amounts of certain behavior on biology. No biological view of that sort stands up to close examination, and Spencer’s view is saying something much weaker – biology allows us to determine from someone’s DNA some facts about which other people are very likely to have similar ancestry, and thus we can group people in ancestral populations who have a large amount of ancestry in common, which happens to coincide with populations whose ancestry comes from certain geographical regions because of historically isolated breeding populations that are not so isolated anymore.

However, just as I said there was a mismatch problem with the 1990s biological race views, there is one here as well. One very important part of all this is that someone can belong to multiple populations. For example, African Americans in the U.S. today usually have a significant amount of ancestry from sub-Saharan Africa and from the Eurasian population, and many have ancestry from the native American population. By this view, that puts them in 2-3 ancestral populations. And that does not line up well with how we actually classify people most of the time using socially constructed means of drawing lines.

For example, Barack Obama has an African father and a white American mother, and he considers himself black. Someone else with similar ancestry might consider themselves mixed race or half-black, half-white. The things we use to answer questions like that are not always biology or ancestry, although those play a role, but sometimes it’s just self-identification. He does consider himself black, not white, and not really even mixed. But Dwayne Johnson (once also known as The Rock), who has a lot of mixed ancestry from three of the populations (African, Eurasian, and Oceanian), does not like to classify himself racially at all. Nor does Tiger Woods, who jokingly called himself cablinasian (he has black, white, native American, and Asian ancestry, placing him in four of the five OMB groups and genomic ancestral populations).

And when Spencer says the OMB races and the results of Structure line up with 99% overlap, he’s a bit exaggerating, because the study that lined up that well left out one important and very large group, which is South Asians. There are a lot of ethnic groups in India and Pakistan alone (remember that they were once one country, much as North Korea and South Korea were). In fact, if you set the number of populations to discover in Structure at six instead of five, the sixth group that separates itself out is a tiny ethnic population in northwest Pakistan called the Kalash, who don't interact much with other populations in the vicinity. But how do we fit South Asians into our racial categories?

Are people whose ancestry is from India or Pakistan going to be lumped in with East Asians under the general category of Asian? Sometimes Americans do exactly that in how we socially construct our racial groups, and there are some cultural similarities between East Asians and South Asians. For purposes of affirmative action, some people think that makes sense, with the number of people coming to the U.S. from those countries and the number of people going into certain fields of study (especially the sciences), there is a lot of similarity. There are also lots of differences culturally. But ancestrally, what would Structure say? It would surprise many people, I think, that it puts them in the Eurasian group. That part of the world is west of the Himalayas. And technically speaking the very term “Caucasian” originally came from a population from India, and a couple of the languages present in India (including the now-dead language of Sanskrit) are Indo-European, the same language family that most European languages are from. But many Americans would not consider most people from India or most Indian Americans to be white.

You get a pretty ambiguous situation with people from the Middle East, who are also in the Eurasian group. Some people would be reluctant to consider Arabs white, although a lot of Arabs do self-identify as white. A lot people consider Jews white, and a lot of Jews consider themselves white, but a lot of Jews don’t consider themselves white. I had a good friend in college who was Jewish, very American, and culturally probably not hugely different from non-Jewish white Americans, but he didn’t think he was white, because he was Jewish. And I know plenty of Jews who think they very much are white.
​
But Spencer has admitted that there are plenty of social constructions that draw race boundaries differently than the genetic populations from Structure do. Finding a social construction that is different from what Structure discovers does not disprove his claim. His claim is that there is at least one social construction that lines up with what Structure produces, and that social construction is OMB races. Is that correct? Not for people with Indian or Pakistani ancestry, at least. Structure puts them with Eurasians, and the OMB puts them with Asians. The OMB and Structure agree on Middle-Easterners. They are Eurasian by both Structure and the OMB. But South Asians are the major place where  Structure and the OMB disagree, and that's a fairly large group to have the two systems putting in different places if we want to treat the two sets of classifications as 99% the same. That number is not correct.

So there is a mismatch problem, and it does count against Spencer's thesis that the OMB socially-constructed races are the same groups as the biological races that Structure has discovered. It seems obvious to me that both groups are real and that they have a considerably amount of overlap, but it also seems obvious to me that they are not the same groups as each other. If I am right about that, then there are biological races that don’t quite correspond to the social ones we use in everyday life. This debate is ongoing, and this is pretty new research in the grand scheme of things, so we’re dealing with cutting edge thinking about this right now.

If you find it confusing, join the club, but I think we can agree on the data, and some of it is data that shows where we were on this in the 1990s was wrong. The arguments that everyone found convincing in their dismissal of classic biological race  views were good arguments, and they are still good. We should not hold such a view. It is wrong. Even so, that doesn't mean there are no biological races, and I think this new research shows that there are in fact groups that have some level of biological significance that it makes sense to call biological races. We have to be careful in saying that, though, because we need to be clear in rejecting classic biological races, and we need to distinguish these biological races from the socially constructed ones that we ordinarily refer to outside the biology discussion. Those groups are the ones we usually mean when talking about races, and those are not biological categories, even if there are biological categories that overlap considerably with them.

Next time I will begin wrapping up this series with a look at some of the moral issues that come up with how we use race terms and engage in racial thinking, an issue that I like to call the ethics of the metaphysics of race.