Home / HBD / The Existence of Race

The Existence of Race


When arguing about the existence race, there are three main things that people argue about:

  1. The majority view– what do “the experts” actually think about race?
    2. Heterozygosity– is there enough variation within homo sapiens for there to be subspecies?
    3. Clusteredness – if so, does this variation clump up into discrete subspecies?

And so in this post we will go over all three of these topics. More specific objections are dealt with in articles that can be found in the main site menu.

Majority View

Theoretically we shouldn’t be concerned with what the majority of experts think about a topic, we should simply be able to rely purely on our own evaluation of “the facts”.

But in reality nobody has enough time for that and they use expert opinion on various topics as a heuristic for the correct position on that topic. And one common argument against race is that the experts recognize that race doesn’t exist.

The problem is that there is no such consensus. There has not been a comprehensive survey on this, but we have found that get a rough idea of what biologists and anthropologists believe about race:

Morning 2008

Ann Morning looked at the usage of race in the 80 most commonly used high school biology textbooks from 1952-2002. While not as good as a survey of biologists, it gives us a rough indication of the trend in biology:

Year Non-medical description of race Medical description of race
1952-1962 90% 0%
1963-1972 60% 15%
1973-1982 59% 59%
1983-1992 21% 46%
1993-2002 42% 93%

And so we see a steep decline in the use of race, reaching a low point in 1983-1992, but an increase from 1992-2002, including an enormous increase in medical descriptions.

Lieberman 2004

In 2004 Leonard Lieberman reviewed several surveys of anthropologists in North America and Europe, and found that 31% of anthropologists in North America recognized race, while 43% in Europe recognized race:

Location Recognizes Race Denies Race
North America 31% 69%
Europe 43% 53%

Kaszycka 2009

In 2009 Katarzyna Kaszycka surveyed physical anthropologists in Eastern and Western Europe, and divided the results up by age group:

Age 21-35 Races exist Race don’t exist % Recognizes race
East 13 1 92.9%
West 5 10 33.3%
Age 36-55
East 11 11 50%
West 11 24 31.4%
East 18 6 75%
West 5 8 38.46%

Kaszycka shows an important split between Western Europe and Eastern Europe. This is an important clue as to how, where and why race denial emerged.

In 2001, Sun and Strkalj looked at 779 articles in “Acta Anthropologica Sinica”,China’s only biological anthropological journal. They were able to get 74 of the 78 issues that existed from 1982 to 2001. In it they found that 324 articles dealt with human variation. They described their results:

“When we applied Cartmill’s approach to the Chinese sample we found that all of the articles used the race concept and none of them questioned its value. Since these active researchers are also members of the teaching staffs at various educational institutions, it is very likely that this attitude will be transmitted to the next generation of Chinese scientists.”

Of course it is possible that the views of the Chinese have changed since then. But there’s no reason to think that they have. So the only place to have any evidence of a consensus on race at some point is in China, and it’s that race exists.

The best info on what “experts” believe on the existence of race comes from anthropologists. And all of these only deal with physical anthropologists. But what about biologists and geneticists?

Well I don’t know of any recent data on biologists’ views on race, however Lieberman 1992 looked at biology textbooks and surveyed college professors and graduates with BAs, MAs and AAs.

First we can look at the textbooks:

Biology Textbooks:

Position 1936-1964 1965-1974 1975-1983
Accepted Race 17 18 11
Neutral 0 11 8
Denied Race 0 2 2

Physical Anthropology Textbooks:

Position 1936-1964 1965-1974 1975-1983
Accepted Race 9 11 7
Neutral 1 8 13
Denied Race 3 4 13

And so the trend to race denial in the textbooks was much more pronounced in Physical Anthropology than in Biology.

Next we can look at Biologists and Physical Anthropologists with varying degrees of education:

Credential Position Biologist Physical Anthropologist
Ph.D Agree 73% (108) 50% (73)
Neutral 15% (22) 7% (11)
Disagree 22% (17) 42% (63)
B.A. / M.A. Agree 66% (73) 54% (66)
Neutral 18% (20) 8% (10)
Disagree 16% (18) 38% (47)
A.A. Agree 70% (108) 44% (42)
Neutral 10% (15) 15% (14)
Disagree 20% (31) 41% (39)
All Agree 70% (289) 49% (181)
Neutral 14% (57) 10% (35)
Disagree 16% (66) 41% (149)

And what we see is that, at all levels of education, Biologists are more likely to recognize race than Physical Anthropologists. At the PhD level, in 1983-1984, Biologists were 1.46 times more likely to accept race.

It’s always hazardous to extrapolate too much into the future toward the present, but there’s no reason to believe that Biologists are all on board the race denial train today when Physical Anthropologists aren’t even all on board today.

Regarding Geneticists, I don’t know of any scientific survey. But I do know of multiple genetics papers lamenting the return of “the race concept” in genetics:

2006 – “Straw Men and Their Followers: The return of biological race

“Until Armand Marie Leroi’s New York Times Op-Ed of March 14, 2005, it is unlikely that many Americans, even among the daily readers of the paper, knew that we are living in the midst of a raging debate over the existence of human races. This debate is occurring among and between a variety of researchers in genetics and social scientists from a range of disciplines. A number of evolutionary biologists, geneticists, biological anthropologists and medical researchers have recently challenged the view put forth by other scientists and social scientists that ‘Race is only social concept, not a scientific one.’”

2007 – “Genes, genomes and genealogies: the return of scientific racism?

2009 – “Return of the race myth?

2012 – “The Return of Biological Race? Regulating Innovations in Race and Genetics Through Administrative Agency Race Impact Assessments

2014 – “The Uncanny Return of the Race Concept

“The aim of this Hypothesis and Theory is to question the recently increasing use of the “race” concept in contemporary genetic, psychiatric, neuroscience as well as social studies.”

2015 – “Great Is Their Sin: Biological Determinism in the Age of Genomics

And there are certainly many more, this is just what I found with a google search and going to page 2.

Obviously it would be better to have proper surveys of geneticists, but the winds seem to be blowing in the direction of race recognition. More evidence of this will be shown in cluster analyses.

While the survey and textbook data is a bit spotty and / or dated, and for the geneticists we only have subjective analyses about trends in the field – it’s enough to where we can say that this “consensus” on race denial simply does not exist, or in the case for geneticists there’s certainly no evidence of a consensus or a trend against race.

The point here is not to appeal to a consensus to say that race exists, but to remove this appeal to authority as a line of argument so that we can move onto the evidence itself.

And the disagreement among the “experts” on race is evidence of another thing: these experts probably don’t spend a whole lot of time researching race. Remember in the Chinese journal – 59% of the biological anthropology articles were NOT about race. Biology and general anthropology it could be even higher.


One popular argument used against the possibility of races in humans is that “humans are 99.9% the same”. This claim was first made by Craig Venter in 2000. However, in 2007, Venter did another analysis that showed that only around 99.5% of human chromosomal DNA is the same between two random individuals. Despite this, it is still a popular talking point among race deniers than “humans are 99.9% the same”.

Keep in mind that humans and chimpanzees, on this scale, are 98.7% the same and 98.4% the same as gorillas.

Michael Woodley 2009 – Is Homo Sapiens Polytypic compared the heterozygosity in humans to other species with wide ranges. Heterozygosity is simply the probably that, at any given gene location, two organisms of that species will have a different gene variant (allele) at that specific location. A gene is a series of SNPs, so even though the similarly is 99.5% SNP by SNP, at any given gene they can be different more often than not.

Of course this doesn’t measure the magnitude or consequence of that difference, merely whether there is in fact a difference, and so it is a rough measure.

Species Heterozygosity Recognized Subspecies Source
Humans .776 ? Wise
Humans .73 ? Jorde
Humans .698 ? Bowcock
Chimpanzees .765 4 Reinartz
Chimpanzees .63 4 Wise
African Buffalo .729 5 Van Hooft
Leopards .58 13 Uphyrkina
Jaguar .739 9 Eizirik
Pumas .52 6 Culver
Canadian Lynx .66 3 Schwartz
NA Brown Bears .5275 19 Paetkau
Scan. Brown Bears .687 19 Waits
Coyotes .629 19 Garcia-Moreno
Gray Wolves .574 37 Garcia-Moreno
African Wild Dogs .643 5 Girman
NA Wolverines .55 2-3 Kyle and Strobeck
Scan. Wolverines .325 2-3 Walker
Elk .395 7-8 Polziehn
Bighorn Sheep .6235 3 Forbes
Bonobos .535 1 Reinartz
Polar Bears .68 1 Paetkau
Dingoes .445 1 Wilton
Domestic Dogs .5085 1 Garcia-Moreno

So by this measure, there appears to be enough total genetic variation in homo sapiens for there to be supspecies within homo sapiens.

In addition, there are several recognized subspecies which have diverged from each other in time spans similar to or more recent than humans have had to evolve subspecies:

Subspecies’ Time of divergence
Cyanoptera – Discors (birds) 65,000
Discors – Septentrionalium (birds) 70,000
Cyanoptera – Septentrionalium (birds) 95,000
North American Moose – European Moose 165,000
Proposed time for human subspecies to have evolved 100,000
8 subspecies of tigers 72,000
2 subspecies of lizards 12,000
Eastern and Western Wood Duck 34,000

And given the speed at which humans have spread around the world, it shouldn’t seem intuitively implausible that humans would evolve subspecies faster than most (but not all) other animals. One counter is that humans have longer generation times, and perhaps that offset the more rapid exposure to new environments. Or perhaps not, or perhaps partially.

We know that humans evolved regional differences in skin color and adult lactose tolerance, so we know for a fact that there has been some divergent evolution going on. It isn’t immediately obvious that there “isn’t enough time” for humans to form subspecies, especially since punctuated equilibrium (pioneered by our good friend Steven Jay Gould) and introgression provide mechanisms for fast evolution.

Genetic Clusteredness

Fst is the proportion of total variation that exists between two populations compared to the overall variation in both populations.

Fst distances between the subspecies of various species are similar to the distances between populations of humans:

Species Fst Distances Recognized Subspecies Source
Gray Wolves 0.168 37 Roy 1994
Pumas 0.167 6 Culver 2000
Humans (K=14) 0.155 ? Barbujani 1997
Asian Dogs 0.154 11 Kim 2001
Humans (K=44) 0.121 ? Jorde 2000
North American Coyotes 0.107 19 Roy 1994
North American Wolverines 0.067 2-3 Kyle and Strobeck 2001
Jaguars 0.065 9 Eizirik 2001
African Buffalo 0.059 5 Van Hooft 2000
Canadian Lynx 0.033 3 Schwartz 2002
Humpback Whales 0.12 3 Jackson 2014
Plains Zebra 0.11 5 Lorenzen 2008
European Wildcats 0.11 3 Pierpaoli 2003
Kob Antelope 0.11 2-3 Lorenzen 2007
SW European Cow 0.068 18 Jordana 2003
Red Winged Black Bird 0.01 5 Williams 2004

Fst values will vary depending on how many groups you divide a species into. The Fst value for each species is just an average for all subspecies-relations. “K=14” refers to the number of subdivisions in the species. The fewer the divisions that are recognized, the higher the average Fst value between the divisions will be. That’s why any species will have a higher Fst value when the number of populations is lower.

Another thing we can look at is how self-identification corresponds to best-fit genetic cluster. Cluster analyses are when researchers get a bunch of genetic data, and then tell the computer to sort it into clusters. For example, it may take 300 pieces of DNA, and then tell the computer to sort individuals based on those 300 pieces of DNA into five clusters.

The computer can sort it into however many clusters you tell it to.

So say there are five recognized races in an area. What researchers will do is take some genetic markers, give it to a computer, and tell the computer to sort it into five categories, and see how well those five categories correspond to the socially recognized races.

Bamshad 2003

Correspondence between genetic cluster and geography by number of loci used:

Location 20 Loci 60 Loci 100 Loci 160 Loci Number
Africa 91% 99% 100% 100%
Asia 82% 96% 99% 99%
Europe 79% 95% 99% 100%

These are the results when K=3, or the number of populations the computer is told to classify the data into is 3. The computer doesn’t know beforehand the race of the individuals, it just gets data, then is told to cluster them into X number of groups. Bamshad also looked at the correspondence between geography and cluster when the number of groups is 2 and 4:


This also deals with another criticism of race: that the categories are subjective. This is sort of true, in the same way that the lines between color are subjective. Where red ends and orange begins is a subjective thing, so are we going to use the subjectivity of boundaries to say that colors don’t exist?

In addition, the boundaries only go in one dimension. You don’t have, for example, a racial category where Japanese and Kenyans are in the same racial group.

The next picture provides a visual description of how well the races correspond to geography based on the number of loci.

Now if you were a researcher and wanted to show that race doesn’t exist, which triangle would you show? The one with the fewest loci.

This is often done in courses on race where the teacher gives the class a genetic test to see what race they are, and their test uses a small number of SNPs (around 20). The fewer SNPs you use, the less it will cluster, and thus the less genetic cluster will correspond to visibly recognized races. And so a dishonest or uninformed teacher will do this.

Alloco 2007 looked at RANDOM locations of SNPs and found that, using random SNPs, you still get very good correspondence between self-ID and best fit genetic cluster:
Using as few as 100 randomly selected SNPs, they found a roughly 97% correspondence between self-reported ancestry and best-fit genetic cluster:

In addition they found that coding and non-coding DNA had almost identical clusteredness, which is evidence against the idea that the races primarily differ on non-coding DNA due to genetic drift and differ less on coding DNA due to stabilizing selection.

Lao 2010 looked not at SNPs, but at short tandem repeats, and basically measured variation in the number of times a known sequence DNA repeated itself at various sites. In this case, they looked at 24 different sites, and compared self-reported ancestry with where they fit into STR clusters:

Self-reported race K1 K2 K3 K4 N
European 0 19.0% 80.6% 0.4% 245
Hispanic 2.4% 77.8% 15.7% 4.0% 127
African 0% 2.2% 1.0% 96.8% 246
Asian 99.9% 0.1% 0 0 46

This is a different method than the other studies, but yields similar results.

Guo 2015 is the most recent cluster analysis that compared self-identified race with best-fit cluster:

The ROOM study used 384 SNPs, and the ADD Health study used 1,536 SNPs.

ROOM Self ID “White” % (N) “Black” % (N) “Non-South Asian” (N) Total Number
White 99.5 (1399) 0.28 (4) 0.21 (3) 1406
Black 0.71 (2) 99.3 (279) 0 281
South Asian (India, Pakistan, Bangladesh) 100 (41) 0 0 41
East Asian 2.33 (2) 0 97.7 (84) 86
Amerindian 75 (3) 0 25 (1) 4
Others 80.7 (50) 12.9 (8) 6.45 (4) 62
Multiracial 52.9 (91) 36.1 (62) 11.1 (19) 172
ADD Health Wave I Self ID
White 99.4 (1429) 0.42 (6) 0.14 (2) 1437
Black 0 (0) 100 (381) 0 381
East Asian 6.29 (10) 0 93.71 (149) 159
Amerindian 100 (19) 0 0 19
Others 91.1 (163) 5.03 (9) 3.31 (7) 179
Multiracial 72.0 (67) 26.9 (25) 1.08 (1) 93

In the Guo study, Whites in the US matched the “White” genetic cluster 99.4% and 99.5% of the time, Blacks matched “Black” 99.3% and 100% of the time, and East Asians matched “Non-South Asian” 97.7% and 93.71% of the time.

More studies that replicate Bamshad and Guo are Rosenberg 2005Rische 2005Witherspoon 2007 and Porras-Hurtado 2013.

The History of Race Denial

As seen above, race denial appears to have come about following world war two.

The concluding statements of a Moscow conference in 1998 on race entitled “Race: Myth or Reality” included:

“(1) According to the old anthropological tradition big human morphological variations which are the result of polymorphism united by common origin in certain geographical areas had been given the name »races.« (2) Reality of the racial subdivisions of Homo sapiens are supported by the totality of the scientific data investigated on the different levels of human organism: morphological, physiological and genetical. Racial classification created with regard for morphological criteria clearly enough reflect the phylogeny of the separate populations and groups of populations. (3) Negativism to the race concept which became apparent during the last decades, in many respects might be explained by the psychological shock which all progressive humanity had felt in the epoch of Hitlerism.”

While certainly more polite, these conclusions could have come straight off of the stormfront.org message boards.

The survey data, as limited as it is, paints a picture of what happened: following world war 2 there was an academic movement to deny race. This began with the postwar popularization of the 1942 book by Ashley Montague entitled “Man’s Most Dangerous Myth: The Fallacy of Race” and then the UN 1951 statement on race.

This movement peaked somewhere between 1980 to 1990, and has since started to recede. But with the internet has come a curious thing: the race denial movement, though receding in the Universities, is gaining ground in public discourse.

Demanding a Hard Definition

One tactic of race denial is to get really hard up on some super-explicit definition of race. Races are obviously just populations of people geographically separated that interbred and thus are genetically and physically distinguishable. That’s obviously true, and all you need, and we should be able to just go forward and talk about the genetic differences between these races.

But the hard definition demander will demand some specific parameters for the sole purpose of deconstructing race. In the hopes that the race-recognizer will not have information on these parameters that he supposedly needs to show that races exist, or he will pre-select conditions which he knows in advance humans don’t meet. For example, an Fst of 0.2, or an inability to interbreed.

And it starts to get absurd when you point to specific alleles that correspond to variation in melanin, and how these alleles vary between the races. Because they will literally be at the point of saying that we cannot say that “black people have black skin” because “black people” is not a “scientific category”. Of course it’s perfectly scientific – we can point them out, their geographic ancestry, it corresponds to cluster groups using specific markers or random SNPs if you want.

And none of these games are ever done for any other subspecies or haplogroup of any other species. It is only with human subspecies, or breeds, that we get this bizarre, radical skepticism.


So first off, there is no obvious consensus on the existence of race, so we cannot use consensus as a heuristic one way or another.

Second off, there is definitely enough genetic variation within homo sapiens to have subspecies.

Third off, this genetic variation clusters into discrete groups, and to the extent Fst is a measure of clusteredness, this can be seen in Fst values and cluster analyses. And the cluster analyses include STR analysis, SNP analysis including ones which use random SNPs, and works just as well with coding and non-coding DNA.

Fourth off, we can trace race denial to a reaction to the Nazis primarily in Western Europe and to a lesser extent in Eastern Europe, which peaked from 1980-1990, and is now in decline among the experts if not among the public.