“Race is not an accurate or productive way to describe human biological variation”

Update: For an attempt to describe some of the findings from “race reconciled” in an Intro-to-Anthropology class see:

Race Reconciled 2009

“Race Reconciled” was a 2009 special issue of the American Journal of Physical Anthropology with cutting-edge work by biological anthropologists. These researchers did not agree on everything, and they had pointed debates. They hailed from the number-crunching and bone-measuring side of anthropology. Some of the articles were dense and difficult reading, with enough numbers, statistical tables, and computer simulations to make it hardly like reading at all. These researchers had all read the critique of Richard Lewontin, and some had been in the forefront of re-examining Lewontin’s work (see previous section Attacking Anthropology and the Race Revival).

It is still important to plow through the findings of Race Reconciled, because it is what anthropology’s best bone measurers and number crunchers accomplished. They very clearly recognize human biological variation. They see variation and measure it every day, examining things people cannot even visibly discern, like tiny bone markers and genetic material. And with all the disagreements, number-crunching, and consideration of how much humans vary, they agree

Race is not an accurate or productive way to describe human biological variation (Heather Edgar & Keith Hunley, Race Reconciled, 2009, 2)Click To Tweet

Wait, What?

Race is a categorization at the sub-species level. Everyone has long agreed that human beings are a single inter-breeding species, and have been for thousands, perhaps hundreds of thousands, of years. To sort a species into sub-species, it is necessary to have biological variation AND a way to group that variation. We have biological variation. The problem comes in establishing the ways variation clumps, groups, or sorts into subsets. We can try this in terms of skin color or skull characteristics, bone measurements, and genetic variation.

Does Skin Color Cluster?

Most people in the United States think they use skin color to classify races. US categories relate to skin color, but not exactly. If it was actually about skin color, racial classifications would look more like Brazil, with lots of different terms and gradations. If it were about skin color, then people might change race classification over the years, or children from the same parents could be classified as different races.

In contrast, the traditional US system is known anthropologically as hypo-descent: children get the racial classification of the parent with the least socially desirable classification. Barack Obama, Halle Berry, and some of Thomas Jefferson’s descendants are considered black. The most extreme example is the “one drop rule,” that any black ancestry meant being classified as black. There have been recent shifts in these attitudes, and there have been regional and historical variations, but this system remains dominant.

Any racial terminology related to skin color, even in Brazil, must have some categories, or ways of marking off groups. However, what do these categories look like when compared to skin color around the world? In a discussion of Race and global patterns of phenotypic variation, John Relethford plots human skin color variation:

The result is a continuous straight line ranging from the darkest extremes to the lightest extremes in skin color. There are no identifiable clusters. . . . Researchers are of course free to subdivide this continuum into different groups, but such clustering would be arbitrary and subjective in terms of the number of groups and the cutoff points used to distinguish them. The lack of apparent clusters is a reflection of the fact that skin color shows a classic pattern of clinal variation. (2009, 17)

There are no clusters or clumps of black skin color or white skin color. Like many traits used to measure race, skin color exhibits clinal variation, along a cline or smooth gradient between the extremes. A walk from the African tropics to northern Europe reveals this gradual variation in skin color. Some people postulate one reason for extreme racial classifications is because Europeans were traveling by sea, and so would meet an extreme example at each stop. The categories used in the United States may in part be a result of a small initial sample, drawn from the extremes of skin color variation.

For an interesting 2017 example, see On the Evolution of Skin Tones In Africa. As noted, there is plenty of skin tone variation within Africa. But perhaps surprisingly: “The latest findings suggest that some particularly dark skin tones evolved relatively recently from paler genetic variants, and these people migrated out of Africa. SNPs in OCA2 and HERC2 that are associated with lighter skin are ancient, over 1 million years old and come from Africa.”

Race Reconciled: Do Skulls Cluster?

Unlike some textbooks and pronouncements which use information about skin color to declare that all physical variation is clinal, Relethford also considered craniometric or skull variation. Here the picture is different, as Relethford finds that crania are “geographically structured” (2009, 18). The differences cluster according to geographic region and reflect genetic relationships: “Global patterns of craniometric variation reflect largely underlying patterns of genetic relationship, which in turn reflect geographic structure” (2009, 19). However, even though there are recognizable clusters, “there are no abrupt breaks in the relationship between phenotypic and geographic distance . . . indicating that decisions for subdivision into clusters (or races) are going to be subjective” (2009, 19). Although it is possible to discern geographic ancestry by continent, the number of groups which could be classified and the geographic cutoffs would be “subjective decisions” (2009, 20).

Relethford considers racial labels as “a culturally constructed label that crudely and imprecisely describes real variation” (2009, 20). Variation is real, exists, and has been structured by geography and migration, but the labels we use are a “crude first-order approximation” (2009, 21). Relethford uses the example of how we see height as short, medium, and tall: “We tend to use crude labels in everyday life with the realization that they are fuzzy and subjective. I doubt anyone thinks that terms such as ‘short,’ ‘medium,’ and ‘tall’ refer to discrete groups, or that humanity only comes in three values of height!” (2009, 21). (See also Relethford’s textbook section in The Human Species: An Introduction to Biological Anthropology and the discussion in the blog-post Race is a Social Construction).

Scientific consensus in 2009 was that craniometrics yields clustered geographic groupings, but those groupings are subjective and arbitrary. However, this consensus could be changing, as 2010 research called into question the relationship between craniometrics and genetics, particularly these clustered groups:

Classificatory analyses achieve high levels of success because they depend on the a priori definition of group centroids. As a consequence, when a large number of variables is considered, the probability that this kind of analysis will find a dimension in the original data that differentiates among the a priori defined groups is high. Yet the precise biological significance of this kind of difference is hard to establish, especially when the high values of dissimilarity fractions reported here are considered. High rates of correct discrimination of groups can thus be misleading in understanding the structure of human biological diversity. . . .

Our results also have implications for the discussion about the existence of races in the human species from the phenotypic point of view because they support the notion of an absence of discrete biological groups. . . . The results presented here demonstrate that roughly one-third of the pairs of individuals within a population are more different than pairs of individuals between populations. This indicates that cranial morphology is less able to identify nonclinal variations among populations (which would be in accordance with the existence of biological races in the human species) than molecular data . . . (Strauss and Hubbe 2010, 326)

I quote at length to highlight the potential challenge to scientific consensus. The idea that craniometrics exhibits geographic and genetic clustering could be due for revision, further undermining what has sometimes been used as evidence for racial groupings. Moreover, as discussed in the section Human Skulls: Boas Head Shape Studies Revalidated, it would also be interesting to see more work on the possibility that infant head molding influences craniometric markers. Infant head molding may also be geographically structured by cultural area.

Race Reconciled & Measuring Bones

Skin color, like many other racial measures, is continuously variable. Crania may be structured geographically, but classifications based on geographic clusters would be arbitrary. But what about measuring all the bones? Television shows feature forensic anthropologists easily identifying race from skeletal remains. Does that mean race is real?

Forensic anthropologist Norman Sauer answered this question in a classic article titled Forensic Anthropology and the Concept of Race: If races don’t exist, why are forensic anthropologists so good at identifying them? (1992). Sauer explained that “the successful assignment of race to a skeletal specimen is not a vindication of the race concept, but rather a prediction that an individual, while alive was assigned to a particular socially constructed ‘racial’ category” (1992, 107). Forensic anthropologists have samples of bones from many geographic areas, and can classify bones according to what race society has assigned to people with ancestry in those geographic areas. However, examining the bones provides a probability estimate of likely race assignment: “In ascribing a race name to a set of skeletonized remains, the anthropologist is actually translating information about biological traits to a culturally constructed labeling system that was likely to have been applied to a missing person” (1992, 109).

Despite the provocative and sometimes misunderstood title, Sauer was pleading for forensic anthropologists to better explain what it means to make racial classifications from skeletal remains. He asked forensic anthropologists not to “sail on” without making an effort to expose people “to the notion that perceived races are not reflections of biological reality” (1992, 110). We should “not fall into the trap of accepting races as valid biologically discrete categories because we use them so often” (1992, 110).

Ah, Sauer writing in 1992 now seems so quaint. Since then, popular media has over and over trumpeted the notion of forensic anthropologists perform identification miracles (see Kristina Kilgrove on The Forensics of Temperance Brennan). Sauer’s plea does not seem to have resulted in an institutionalized move by forensic anthropologists to expose people to the difference between perceived race and biology. I have not systematically surveyed forensic anthropology syllabi, but my informal conversations lead me to believe Sauer’s article did not become an explicit part of many forensic anthropology courses.

A different question is whether Sauer’s stance remains valid, given the increased sophistication of measurement and quantification in forensic anthropology. Here two articles in the Race Reconciled volume were especially insightful–the increasing sophistication of measurement and quantification only reinforces Sauer’s claim that forensic anthropology does not confirm traditional race classifications, even when race-identification probabilities are reported from skeletal remains.

The first article, Understanding race and human variation: Why forensic anthropologists are good at identifying race, obviously took its title in reference to Sauer. The authors specifically tackled physical differences between US blacks and whites, noting that there have been historically low rates of interracial marriage, given legal restrictions persisting in some states until the 1960s, and that “unofficial social penalties for interracial relationships and marriage included violence and murder” (Ousley et al. 2009, 69). Different continental ancestry combined with institutional racism made it possible to discern a “clear craniometric separation of American blacks and whites” (Ousley et al. 2009, 72).

However, the authors support Sauer’s contention that craniometric separation does not confirm traditional racial categories. “Sauer’s additional suggestion that differences in American blacks and whites did not validate the traditional biological race concept is likewise supported by our results” (Ousley et al. 2009, 73).

Why? The authors highlighted just how many social differences could be discerned by forensic anthropologists. Given an original sample of bones classified into social groups, a forensic anthropologist can with high probability predict to which group another case of bones belong. They can separate Japanese from Chinese from Vietnamese, or northern Japanese from southern Japanese. Or, and perhaps most incredibly, “white males born between 1840 and 1890 can be separated from white males born 1930 to 1980 very well, and they are distinguished by time, and would appear to qualify as different races” (2009, 74). Group bones by birth-year, run the statistics, and then introduce a new sample: the sample can be accurately classified, and a new race born every fifty years!

Forensic anthropologists sort real physical variation into categories we have made socially relevant. “There are so many possible distinctive biological races that the concept is virtually meaningless. We can only concur with Howells’ modification of Livingstone’s 1962 quote: ‘There are no races, only populations’” (Ousley et al. 2009, 74). (Livingstone’s original quote [1962, 279] was “There are no races, there are only clines.”)

The second article, Estimation and evidence in forensic anthropology: Sex and race did not have a provocative title, but it remains an even more incredible piece. The authors began with sex identification, showing how sex is reliably estimated from a few craniometric variables, and how a prior identification of a roughly 1:1 sex ratio is unimportant for making the call.

Things change when it comes to racial identification. Here, the authors took a set of bones from “Mr. Johnson” and compared them to a world database: “The results from these analyses fairly unambiguously estimate Mr. Johnson’s origin as an Easter Islander” (Konigsberg et al. 2009, 81). However, since Mr. Johnson’s bones were found in Iowa, plugging in the Iowa probabilities allowed Mr. Johnson to be reliably predicted as white. Forensic anthropologists base their estimates on the known prior composition of the population. If the same bones from Mr. Johnson had been found in Hawaii, they would have estimated “Easter Islander” or if found in Gary, Indiana, they would have estimated “American Black”:

Using the Iowa priors, the highest posterior probability is for “American White” at 0.6976. The identification of “Easter Islander,” which had the highest posterior when we used an uninformative prior, now has a relatively low posterior probability (0.0449). In contrast, using the Hawaii priors the posterior probability that “Mr. Johnson” was an “Easter Islander” is 0.9068, whereas the posterior probability that he was an “American White” was 0.0188. Using the Gary, Indiana prior the highest posterior probability (0.5342) was for “American Black” with “American White” having the second highest posterior probability (0.2728). (Konigsberg et al. 2009, 82)

Wow. Forensic anthropologists do not determine race from bones.

What actually happens is forensic anthropologists match bones probabilistic frequencies against known existing assortments. Those assortments can be anything socially relevant. Changing the context of bone discovery could lead to different predictive classification–of the same bones. “The use of different priors also shows the importance of prior information, as ‘Mr. Johnson’ would have been classified as a Pacific Islander had his remains been found on Hawaii and as an ‘American Black’ had his remains been found in Gary, Indiana” (Konigsberg et al. 2009, 83).

Forensic anthropologists “often do bring prior information to their cases, though this information is typically implicit, unstated, and not quantified” (Konigsberg et al. 2009, 84). Estimates are always probabilities, and probabilities rely on pre-existing information, such as a self-reporting census. Rather incredibly, the authors conclude that “forensic anthropologists are not particularly adept at identifying races when they must deal with a very heterogeneous population at large, and this is the one setting in which a definitive racial identification would be useful” (Konigsberg et al. 2009, 86). Guessing Mr. Johnson’s bones, when found in Iowa, were “probably white” is a guess anyone could make based on the Iowa census.

A 2018 example of how forensic anthropologists must bring prior information to their analyses can be seen in Richard L. Jantz’s Amelia Earhart and the Nikumaroro Bones: A 1941 Analysis versus Modern Quantitative Techniques. Jantz demonstrates all the bone measurements, but also reveals that identification depends upon “prior odds”: “It is often impossible to assign specific numbers to the prior probability, because it depends on how the non-osteological evidence is evaluated, and different people will usually evaluate it differently” (2018, 14).

Race Reconciled on Genetic Variation

Even after proving the continuous variation of skin color, and even after showing how bones and skulls do not confirm traditional race classifications, there is still the sense that genetics offers real proof of race. Genetic testing companies amplify this misconception in a rush to market ancestry, while pharmaceutical companies sell race-targeted medications.

On the issue of genetic ancestry testing, there seems to be greater recognition that the recreational activity does not say that much about real ancestors. The 2007 New York Times article DNA Tests Find Branches but Few Roots, included in the 4th edition of Anthropology: What Does it Mean to be Human? discusses exactly how many possible branches there may be in these kinds of tests. The wonderfully-titled 2013 article, To claim someone has ‘Viking ancestors’ is no better than astrology is also helpful:

You don’t have to look very far back before you have more ancestors than sections of DNA, and that means you have ancestors from whom you have inherited no DNA. Added to this, humans have an undeniable fondness for moving and mating–in spite of ethnic, religious or national boundaries–so looking back through time your many ancestors will be spread out over an increasingly wide area. This means we don’t have to look back much more than around 3,500 years before somebody lived who is the common ancestor of everybody alive today. And perhaps most surprisingly, it has been reasonably estimated that around 5,000 years ago everybody who was alive was either the common ancestor of everybody alive today, or of nobody alive today; at this point in history we all share exactly the same set of ancestors.

Finally, listen to the August 2018 podcast Is Your DNA You? on Sapiens.

For a more detailed look at actual DNA sequences, the Race Reconciled article Human DNA sequences: More variation and less race, is difficult reading but extremely helpful. The authors used sophisticated techniques to show how human genetic variation occurs in sets and subsets. Sub-Saharan Africa has the greatest genetic diversity. This is not a surprise, since Sub-Saharan Africa is where almost all human evolution occurred. For most of human history, it was also the region with the largest human population. What may be more surprising is “that the diversity in non-Sub-Saharan African populations is essentially a subset of the diversity found in Sub-Saharan African populations” (Long et al. 2009, 23).

Genetic classifications of races outside of Sub-Saharan Africa are simply subsets of Sub-Saharan African diversity. Moreover, and perhaps most strangely, “a classification that takes into account evolutionary relationships and the nested pattern of diversity would require that Sub-Saharan Africans are not a race because the most exclusive group that includes all Sub-Saharan African populations also includes every non-Sub-Saharan African population” (Long et al. 2009, 32). In the end, the authors “agree entirely with Lewontin that classical race taxonomy is a poor reflection of human diversity” (Long et al. 2009, 32). They disagree with Lewontin over whether this is intrinsic to human genetics–rather, it is a product of evolutionary history and migration.

This evolutionary history is explained in the article The global pattern of gene identity variation reveals a history of long-range migrations, bottlenecks, and local mate exchange: Implications for biological race. Once again, sophisticated techniques reveal a “nested pattern of genetic structure that is inconsistent with the existence of independently evolving biological races” (Hunley et al. 2009, 35). The authors confirmed greater genetic variation within Sub-Saharan Africa, and all other humans are a sub-set of this variation. Taxonomic classifications of race cannot account for observed genetic diversity. The authors took this point further, challenging medical research that uses visible race-markers as a proxy for genetic structure:

Our findings confirm that broad ethnic categories employed in medical genetic research might not adequately take into account the complex geographic pattern of genetic structure in the species, but for the same reason, neither may continental ancestry. This is because our results also indicate that substantial, potentially medically important genetic differences may exist between populations within regions. (Hunley et al. 2009, 45)

Drug companies desiring to find treatments targeted at black patients probably underestimate the true degree of genetic diversity present within Africa and for people of African descent living in the United States. But these stories get simplistically bantered about for profit, patents, and political correctness allegations. For more on this tale, see Race in a Bottle: The Story of BiDil and Racialized Medicine in a Post-Genomic Age.

Race Reconciled remains a superb collection of articles, deserving of thorough reading and reflection. Race is simply much more biologically complicated than classifiers imagine. These are all researchers who have taken the critique of Lewontin seriously, know how to crunch their numbers and measure bones. They re-debunk race.

Still, the authors of Race Reconciled could have done better on three fronts:

1) They Needed Bigger Megaphones

At the end of their introduction to the Race Reconciled volume, Edgar and Hunley wrote that the best way to communicate with the public was in the classroom: “Biological anthropologists teach thousands of students every year and have a real opportunity to influence public views of human variation” (2009, 2-3). Their final sentence was about communicating “to our students” (2009, 3). But as important as classrooms are, how many students actually take a course devoted to biological anthropology? We are lucky if they take Introduction to Anthropology. The authors needed to have been on television, on YouTube, in the newspapers, in the blogosphere. With 10 million undergraduate students in the United States, teaching thousands was not enough.

The need for greater public exposure grew obvious in 2010, when New York Times reporter Nicholas Wade covered a so-called debate within anthropology. Wade wrote that people in physical anthropology, studying things like skulls, were “evidence-based researchers” and separated from “more interpretive subjects, like research on race and sex.” Ridiculous! But this was very convenient for Wade, who had long been claiming that genetic research validated traditional race-thinking. By classifying anthropologists who studied race as “interpretive,” he got to ignore the evidence-based researchers who had invalidated race as an accurate way to describe human variation. Maybe if the Race Reconciled scientists had already been on television, Wade would not have been able to get away with it.

Since then, evolutionary biologists Jerry Coyne and David Barash also revealed they had no idea about the scientific literature on race and human variation. But while these big platform science bloggers precipitated extensive blogosphere back-and-forths and a flurry of comments–see Race Redux–the authors of Race Reconciled seemed to have nary a word.

2. They Needed to Understand Politics

The editors of the Race Reconciled volume did step out of the arena of academic peer-reviewed journals to record a podcast about their research. This was definitely a welcome step toward television and the blogosphere. However, they here seemed to misunderstand the politics of their research:

It seems to me like people expect a single answer: “Yes, there’s race,” “No, there’s not,” “Yes it means something important,” “No it doesn’t.” This is a complex issue that involves all of human history and all of human variation and I think it’s not possible to provide a single answer like that and I also don’t think it’s necessary. I think that we can develop a teaching perspective that will allow us to explain: “Well some people think this; some people think that; here are some ideas that they share; here are some parts that aren’t quite worked out yet.” And students are able to take that in. We teach them that about all kinds of scientific debates. We teach them about the debate about whether or not Neandertals are part of our human ancestry and we expect them to understand that there are multiple viewpoints on that. Why don’t we expect them to do the same thing when it comes to contemporary human races? (Formerly on UNMLive as “Does Race Exist?” My transcription)

It is not the same thing! See section on Biological Anthropology and Racism. Race remains a highly charged issue, with political and economic consequences. Although the way we represent Neandertals or classify species can and does interact with this contemporary politics, it certainly does not have the same political potential.

It is important to make sure everyone understands the headline message from Race Reconciled: “race is not an accurate or productive way to describe human biological variation” (Edgar and Hunley 2009, 2). Then we can work on the details of why it is not accurate or productive. After that, outline the areas of scientific debate. The model is the same as teaching evolution. Evolution is the scientifically correct and factual account. There are debates within the community as to particular evolutionary processes, but that does not change the overall message.

Similarly, Relethford’s analogy to how people make crude classifications of short-medium-tall potentially misrepresents the politics and history of racial classification in the United States. People may make crude guesses based on skin color, but hypodescent has been dominant. It would be similar to a situation in which there really was no “medium” category. Anyone born to one short parent and one tall parent would be considered short. Barack Obama, Halle Berry, Colin Powell. Relethford obviously knows that people categorized as black have a “highly variable” level of African ancestry (2009, 21). Still, he could have been clearer on how these classifications are politically, economically, and socially enforced, making the short-medium-tall comparison potentially misleading (see also the blog-post Race Remixed?).

In general, the authors of Race Reconciled seemed to believe in what Michel-Rolph Trouillot termed the “Liberal Space of Enlightenment”:

The desire to occupy a privileged space of enlightenment is a frequent feature of both philosophical and political liberalism. . . . Liberalism wishes into existence a world of free willing individual subjects barely encumbered by the structural trappings of power. The dubious proposition follows that if enlightened individuals could indeed get together within their enlightened space, they could recast “culture” or “race” and, in turn, discharge other free willing individuals of their collective delusions. But is racism a delusion about race? Or is race made salient by racism? That is the crux of the matter. (Adieu Culture, 111; see Is Anthropology better than Starbucks? for a 2018 take on how anthropology prefers to believe in this “privileged space of enlightenment”)

3. They Needed to Protect Turf

After I assigned the Race Reconciled articles in my anthropological theory course, one student wanted to do further research on race and forensic anthropology. The problem, I suggested, was that whenever a forensic anthropologist used the term race–in any context–people assumed this must mean race exists. It goes back to Sauer’s article: Forensic anthropologists need to explicitly state they are estimating ancestry probabilities to best-guess what racial group a person would have been socially assigned.

As I searched relevant articles, I was shocked and disgusted to discover the materials from forensic anthropologists were not only suffering popular misinterpretation, but were also being misused in peer-reviewed academic journals. One of the only 2010 articles citing the forensic anthropology pieces from Race Reconciled was a 2010 diatribe by philosopher Neven Sesardic titled “Race: A social destruction of a biological concept.” Sesardic claimed fuzzy-headed social scientists and philosophers had misunderstood the proper biological underpinnings of the race concept. Sesardic then used the forensic anthropology articles to make his case. Sesardic called Sauer a “bewildered and exasperated scientist” (2010, 156) and went on to approvingly quote Sauer’s 1992 article as supporting the idea of racial assignment. He similarly cherry-picked quotes from the 2009 articles in Race Reconciled to portray forensic anthropology as confirming traditional biological races.

I contacted the authors of the Race Reconciled articles, alerting them to this misuse of their research. I received two replies, eventually leading to a productive e-mail exchange. However, no one felt it was worth writing a rebuttal. A journal titled Biology and Philosophy was not part of their academic engagement. They questioned the journal’s standards and peer-review process, since it published an article with such blatant misuse of data. Still, it would seem that if it is impossible to track and respond to what is published in academic journals, there is limited chance of wider public intervention. So, while I can understand not wanting to get muddy in the internet blogosphere, Race Reconciled could even have done better at protecting the scientific turf.

Since I began blogging about these issues in 2011, I’ve discovered how much mocking the idea of race as a social construction yields huge political dividends, or as I’ve put it: Social Construction of Race = Conservative Goldmine. So although it remains important to understand that race has been repeatedly debunked biologically, the political implications of this debunking were a dead-end.

Next: 1.7 – Race Becomes Biology, Inequality Embodied

Previous: 1.5 – Attacking Anthropology and the Race Revival

To cite: Antrosio, Jason, 2011. “Race Reconciled: Race isn’t skin color, biology, or genetics.” Living Anthropologically website, https://www.livinganthropologically.com/biological-anthropology/race-reconciled-debunks-race/. First posted 5 June 2011. Last updated 20 May 2020.