Evolutionary Insights by Anthropology.net
Evolutionary Insights by Anthropology.net
Podcast Description
A podcast about anthropology. www.anthropology.net
Podcast Insights
Content Themes
Explores themes such as prehistoric technology development, human migration patterns, and cultural practices through archaeological evidence, with episodes covering topics like the evolution of pigmentation, the use of fire in the Upper Paleolithic, and the genetic backgrounds of Neanderthals and Homo sapiens.

A podcast about anthropology.
In April 2026, a team led by Ali Akbari and David Reich at Harvard analyzed ancient DNA from 15,836 people who lived across West Eurasia, 10,016 of them with newly generated data, and reported that natural selection had pushed hundreds of genetic variants up or down in frequency over the past ten thousand years. Most of the paper, which appeared in Nature, is about allele frequencies and statistical method. One sentence in the abstract is about something else: combinations of alleles that predict cognitive test performance in living people became more common over that span. When the team posted an earlier version as a preprint in September 2024, Emil Kirkegaard wrote in the online magazine Aporia nine days later that “the hunter-gatherers were apparently much duller than the farmers.”
Readers of this site will know the standard reply to race realism, and Jonathan R. Goodman of the University of Cambridge and Rebecca Sear of Brunel University of London made it well in an essay for The Conversation on 13 November, which Ars Technica republished this week. Richard Lewontin showed in 1972 that most human genetic variation lies within populations rather than between them; ancient DNA shows that every living population is a mixture of older ones; so the idea of genetically distinct races with innate differences in intelligence falls apart. I agree with their verdict. But the two genetic arguments at the center of that reply were built to answer an older kind of race science, the kind that argued from skulls and blood groups. Today’s hereditarians argue from polygenic scores (single numbers that add up a person’s many small genetic associations with a trait) and from selection scans of ancient DNA, which are the newest tools of population genetics itself. One of the companies at the center of the movement renamed itself Polygenic Scores LLC in April 2025.
“We believe it is unequivocal that race science is fundamentally pseudoscience.”
Jonathan R. Goodman and Rebecca Sear, The Conversation, 2026
So this essay picks up where theirs stops. The hereditarian case in 2026 has to be answered inside statistical genetics, on its own terms, and when it is, the answer is stronger than the one Lewontin gave. The controls that geneticists developed over the past decade (family-based studies, tests of how well scores travel between populations, and corrections for hidden population structure) cut directly against the claim that genes explain differences in average test scores between racial groups. Getting there requires conceding something race realists like to point out, which is that human genetic ancestry is real and can be measured.
Is race biological? What Lewontin’s 85 percent does and doesn’t show
Lewontin’s 1972 study took blood-group and protein data, sorted human populations into seven “races” as anthropologists of the period drew them, and asked how total diversity was divided. About 85.4 percent fell within populations, 8.3 percent between populations within the same race, and 6.3 percent between races. Studies with far larger datasets have since put the within-population share anywhere from roughly 80 to 95 percent, and in a 2022 review the geneticist John Novembre concluded that the qualitative result has held up for half a century. Lewontin drew a blunt conclusion from his numbers.
“…racial classification is now seen to be of virtually no genetic or taxonomic significance…”
Richard C. Lewontin, “The Apportionment of Human Diversity,” Evolutionary Biology, 1972
In 2003 the Cambridge statistician and geneticist A. W. F. Edwards called the reasoning that people drew from this “Lewontin’s fallacy.” His point was that the information separating populations is “hidden in the correlation structure of the data,” spread across many loci, so that small differences at each one add up. The arithmetic is on his side. When Noah Rosenberg and colleagues genotyped 1,056 people from 52 populations at 377 loci in 2002, they found that differences among individuals within populations accounted for 93 to 95 percent of genetic variation, and yet a clustering program given no geographic labels sorted those people into groups corresponding largely to major geographic regions. David Witherspoon and colleagues showed in 2007 why both findings hold at once. With a handful of loci, two people from different continents are more similar to each other than two people from the same population about a third of the time. With many thousands of loci, for populations sampled far apart, that almost never happens.
This is the part of the argument I think science writers handle worst. When an essay tells readers that human variation doesn’t fall into clusters, a race realist can pull up any commercial ancestry report and show a pie chart. Ancestry inference works because populations separated by distance and history drifted apart slightly in allele frequency, and enough slight differences let a computer estimate where someone’s recent ancestors lived. That is a statement about genealogy and geography, and it does not recover the races of the census form or the nineteenth-century textbook, as the 2002 paper itself explains.
Rosenberg’s team warned that because they sampled whole populations rather than people spread evenly across the map, their clusters were probably sharper than random worldwide sampling would produce, and that the differences they detected came mainly from gradual changes in allele frequency across regions. The number of clusters is also a setting the analyst picks. When the program was asked for a sixth group, it did not split Africa or Europe; it pulled out the Kalash, a single isolated community in northwestern Pakistan. Clusters track isolation and genetic drift, which is why a small mountain population can earn its own. A 2020 analysis of about 92 million variants by Arjun Biddanda, Daniel Rice, and John Novembre found that a variant common in one region is usually common in every region (63 percent of such variants are widespread, rising to 82 percent outside Africa), and that variants common in one non-African region and absent everywhere else are strikingly rare. When two people differ genetically, the variants responsible are most often ones found worldwide.
Africa makes the point sharper. The Human Heredity and Health in Africa (H3Africa) consortium sequenced 426 genomes from 50 ethnolinguistic groups in 13 countries and reported in 2020 more than 3 million variants never seen before, with discovery still not leveling off. The continent has roughly 2,000 ethnolinguistic groups, and only a small fraction have been studied. A racial category like “Black” groups the most genetically diverse people on Earth with one another and with descendants of the Atlantic slave trade, whose ancestry is itself mixed. In a 2015 study of 23andMe customers, African Americans averaged 73.2 percent African and 24.0 percent European ancestry, and most people with less than 28 percent African ancestry identified as European American. About 3.5 percent of European Americans carried at least 1 percent African ancestry. In Brazil, a 2003 study by Flávia Parra, Sérgio Pena, and colleagues found that color, as judged by appearance, was “a poor predictor of genomic African ancestry” at the individual level. The All of Us program reported in 2025, from 230,016 genomes, that people within each self-identified racial and ethnic group show gradients of ancestry rather than discrete clusters.
The scientific societies have reached the same conclusion. In 2019 the American Association of Biological Anthropologists stated that race does not accurately represent human biological variation, past or present. In 2023 a committee of the National Academies of Sciences, Engineering, and Medicine went further for researchers in genetics.
“Researchers should not use race as a proxy for human genetic variation.”
National Academies of Sciences, Engineering, and Medicine, Using Population Descriptors in Genetics and Genom…
This is a public episode. If you’d like to discuss this with other subscribers or get access to bonus episodes, visit www.anthropology.net/subscribe

Disclaimer
This podcast’s information is provided for general reference and was obtained from publicly accessible sources. The Podcast Collaborative neither produces nor verifies the content, accuracy, or suitability of this podcast. Views and opinions belong solely to the podcast creators and guests.
For a complete disclaimer, please see our Full Disclaimer on the archive page. The Podcast Collaborative bears no responsibility for the podcast’s themes, language, or overall content. Listener discretion is advised. Read our Terms of Use and Privacy Policy for more details.