If you glanced around the room at a conference of geneticists, it would be easy to guess where in the world all the attendees' ancestors came from. Using skin color, hair, facial features, and other physical traits, you could distinguish the East Asians from the South Asians and the Africans from the Europeans. Our broad racial categories appear to be founded on genuine biological differences between people from different geographical regions. And these differences seem to define a set of natural human groups, the product of the last 70,000 years or so when modern humans emerged from Africa to colonize the other continents, acquiring distinct physical traits as they adapted to new environments.
The concept of human races appears to be solidly grounded in present-day biology and our evolutionary history. But if you asked that conference of geneticists to give you a genetic definition of race, they wouldn’t be able to do it. Human races are not natural genetic groups; they are socially constructed categories.
Without natural genetic boundaries to guide us, human racial categories remain a product of our choices.
Now, the truth of this claim is not obvious. The idea that humans fall naturally into racial groups is almost universally accepted in all societies. Sure, many people have mixed ancestry that crosses racial boundaries, but there are undeniable physical differences between people native to different parts of the world. Many of those physical differences reflect genetic differences, and over the past two decades, researchers have used those genetic differences to pinpoint the geographical origins of people's ancestry with ever-increasing precision. Just last month, one group reported that they could use DNA to place Sardinians within 30 miles of their native village. On a larger scale, geneticists will frequently talk about the “populations” of general geographical areas, making these broad populations sound very much like races. And there are clear examples of recently evolved adaptations in different human populations, such as the high-altitude physiology in Tibetans and Andeans.
Genes certainly reflect geography, but unlike geography, human genetic differences don't fall along obvious natural boundaries that might define races. As my Washington University colleague Alan Templeton has shown, by objective genetic definitions of race, human races don't exist. Writing in Studies in History and Philosophy of Biological and Biomedical Sciences, Templeton notes that "Human populations certainly show genetic differences across geographical space, but this does not necessarily mean that races exist in humans." For an objective, biological definition of race, this genetic differentiation has to occur "across sharp boundaries and not as gradual changes." Templeton examined two genetic definitions of race that are commonly applied by biologists to vertebrate species. In both cases, races clearly exist in chimpanzees, our nearest relatives, but not in humans.
One natural definition of race is a group whose members are genetically much more similar to each other than they are to other groups. Putting a number on what counts as "much more" is a somewhat arbitrary exercise, but Templeton found that the genetic differentiation between populations of chimpanzees is over seven times greater than the genetic differentiation between broad geographical populations of humans. Furthermore, the level of genetic differentiation between human populations falls well below the threshold that biologists typically use to define races in non-human species.
Races could also be defined by genetic branches on the family tree. For most of us, this is the most intuitive definition of race. It’s one that, at first glance, is consistent with recent human evolution: After originating in Africa, part of our species branched out first into Asia and Europe, and then to the rest of the world. We should thus expect different geographical populations to be distinct genetic limbs on our species' recent evolutionary tree.
But as it turns out, our species' family history is not so arboreal. Geneticists have methods for measuring the "treeness" of genetic relationships between populations. Templeton found that the genetic relationships between human populations don't have a very tree-like structure, while chimpanzee populations do. Rather than a family tree with distinct racial branches, humans have a family trellis that lacks clear genetic boundaries between different groups.
These findings reflect our unusual recent evolutionary history. Unlike the distinct populations of chimps, humans continued to exchange both goods and genes with each other even as they rapidly settled an enormous geographical range. Those ongoing contacts, plus the fact that we were a small, genetically homogeneous species to begin with, has resulted in relatively close genetic relationships, despite our worldwide presence. The DNA differences between humans increase with geographical distance, but boundaries between populations are, as geneticists Kenneth Weiss and Jeffrey Long put it, "multilayered, porous, ephemeral, and difficult to identify." Pure, geographically separated ancestral populations are an abstraction: "There is no reason to think that there ever were isolated, homogeneous parental populations at any time in our human past."
Without natural genetic boundaries to guide us, human racial categories remain a product of our choices. Those choices are not totally arbitrary, biologically meaningless, or without utility. But because they are choices, we have some leeway in how we define and apply racial categories. We shouldn't deceive ourselves; how we define race does not just reflect biology, it reflects culture, history, and politics as well.