Response to Birney, Raff, Rutherford & Scally

Responds to arguments in a recent blog post by Ewan Birney, Jennifer Raff, Adam Rutherford and Aylwyn Scally

In a recent “explainer” on “Race, genetics and pseudoscience”, Birney et al. argue that a small number of researchers “have seized upon some of the new findings and methods in human genetics” to make “misleading” arguments about “racial groupings” and the “alleged genetically-based intelligence differences between them”. Their blog post “aims to provide an accessible guide for scientists, journalists, and the general public for understanding, criticising and pushing back against these arguments.”

After the publication of Birney et al.’s “explainer”, there followed a lively debate on Twitter, which included commentary from some individuals who one might assume are among the “small number of researchers” to whom the authors refer. My arguments here were informed partly by this commentary and partly by my own prior understanding of the relevant phenomena. Note that James Thompson has already penned a response to Birney et al.’s “explainer” on his own blog. In addition, several other commentators have responded via ‘tweet threads’, which readers may find instructive.

My response takes the form of a series of quotations from Birney et al.’s blog post, each followed by comments from me:

A small number of researchers, mostly well outside of the scientific mainstream, have seized upon some of the new findings and methods in human genetics, and are part of a social-media cottage-industry that disseminates and amplifies low-quality or distorted science, sometimes in the form of scientific papers, sometimes as internet memes — under the guise of euphemisms such as ‘race realism’ or ‘human biodiversity’.

Birney et al. might have been so courteous as to identify the specific “scientific papers” and “internet memes” which they regard as “low-quality or distorted science”. However, they chose not to provide any links. One might assume they were referring to articles such as: this one by Davide Piffer; this one by Curtis Dunkel et al.; this one by Emil Kirkegaard et al.; this one by Jordan Lasker et al.; and this one by Meng Hu et al.

More broadly, one might assume they were referring to all individuals who have indicated that they do not believe something called the ‘hereditarian hypothesis’ (the hypothesis that genes contribute to group differences in IQ and other psychological traits) is totally implausible. Note that this group would comprise some academics, some independent researchers, some bloggers, and some individuals running pseudonymous Twitter accounts.

If Birney et al. were referring to articles such as the ones listed above, the charge of “low-quality or distorted science” does not seem to be justified. The articles in question are not necessarily correct––they may be proven wrong by subsequent research––but they do not strike me as any less scientific than, say, articles claiming to show that group differences in IQ are purely environmental in origin. Indeed, some of the critiques that Birney et al. put forward are already anticipated by the authors of those articles.

Research in the 20th century found that the crude categorisations used colloquially (black, white, East Asian etc.) were not reflected in actual patterns of genetic variation, meaning that differences and similarities in DNA between people did not perfectly match the traditional racial terms. The conclusion drawn from this observation is that race is therefore a socially constructed system […]

As Timofey Pnin pointed out on Twitter, Birney et al. fail to provide a convincing argument for why patterns of genetic variation should have to “perfectly” match traditional racial terms in order for those terms to have validity. Moreover, it would appear that the authors have conflated two different senses of the word ‘race’ here.

The first sense of ‘race’ refers to categories that people may identify with, or that may have been assigned to them, in multi-racial countries like the US, Brazil or South Africa (e.g., ‘black’, ‘white’, ‘Hispanic’, ‘Asian’, ‘mulatto’, ‘coloured’). These terms certainly correlate with ancestral populations (e.g., ‘blacks’ have a lot of Sub-Saharan ancestry; ‘whites’ have a lot of European ancestry, etc.) but they do not do so closely enough to justify regarding them as ‘biological races’. For example, ‘Hispanic’ and ‘mulatto’ are obviously not ‘biological races’, but rather populations of mixed ancestry.

The second sense of ‘race’ refers to the actual ancestral populations from which individuals’ ancestries may be derived. To say that a certain individual is ‘half-black’ and ‘half-white’ is to say that half his ancestry comes from the Sub-Sarahan ‘ancestral popluation’ and half comes from the European ‘ancestral population’. In this case then, ‘race’ is roughly synonymous with ‘ancestral population’. And ancestral populations are not socially constructed (or if they are, then so is almost every concept in science). Note that the leader author Ewan Birney even admits that he uses the (unlovely) term ‘European ancestries’.

Furthermore, as the geneticist Neil Risch pointed out in a 2005 interview with PLOS Genetics, other classifications like ‘sex’ and ‘age’ do not perfectly line up with individuals’ biology, but that does not mean they lack all validity or are merely ‘social constructs’. As Risch notes, “Perhaps just using someone’s actual birth year is not a very good way of measuring age. Does that mean we should throw it out? No.”

Even though geography has been an important influence on human evolution, and geographical landmasses broadly align with the folk taxonomies of race, patterns of human genetic variation are much more complex, and reflect the long demographic history of humankind.

Despite earlier claiming that race is “a socially constructed system”, here Birney et al. admit that “geographical landmasses broadly align with the folk taxonomies of race”. This is noteworthy because it hints at a reason why one might regard continental-scale populations (e.g., ‘Melanesians’, ‘Europeans’, ‘East Asians’ etc.) as important. (Note, I did not say “more important than any other division”; I just said “important”.) Specifically, these populations were separated by significant natural barriers during the course of human evolution (e.g., the Sahara, the Himalayas, the Atlantic Ocean), which impeded gene flow between them for substantial periods of time. For more detail on this point, please see a previous blog post of mine, ‘Are Racial Classifications Arbitrary?’

If an alien, arriving on Earth with no knowledge of our social history, wished to categorise human ancestry purely on the basis of genetic data, they would find that any consistent scheme must include many distinct groups within Africa that are just as different from each other as Africans are to non-Africans. And they would find it difficult to identify any natural or obvious subdivision of people into groups which accurately partitions human genetic variation due to the constant migrations of people across the world.

As Steve Sailer pointed out, numerous biological classifications are subject to lumper-splitter debates; there is nothing particularly special about human ‘races’ in this regard. Indeed, scientists disagree about the number of subspecies in many other animals, such as wolves, brown bears, giraffes, chimpanzees, lions, killer whales and leopards.

A similar point was made by the evolutionary biologist Jerry Coyne in a 2012 blog post: “races of animals (also called “subspecies” or “ecotypes”) are morphologically distinguishable populations that live in allopatry (i.e. are geographically separated). There is no firm criterion on how much morphological difference it takes to delimit a race. Races of mice, for example, are described solely on the basis of difference in coat color, which could involve only one or two genes.”

DNA sample collection typically follows existing cultural, anthropological or political groupings. If samples are collected based on pre-defined groupings, it’s entirely unsurprising that the analyses of these samples will return results that identify such groupings. This does not tell us that such taxonomies are inherent in human biology.

It is not clear what “inherent in human biology” means, and Birney et al. already admitted that “geographical landmasses broadly align with the folk taxonomies of race”, so it is hard to know exactly what point they are making here. However, they seem to be saying that continental-scale populations are an artefact of the sampling scheme, which is a point Rosenberg et al. (who — it should be noted — reject the term ‘race’) addressed in their 2005 paper.

Specifically, they computed the genetic and geographic distances between pairs of sub-populations sampled from different genetic clusters, and then plotted them on a chart (see Figure 6). They found that pairs of sub-populations separated by a given geographic distance tend to be more genetically similar if they are from the same cluster than if they are from different clusters. This indicates that although human genetic variation is mostly clinal, it is partly discontinuous.

As Rosenberg et al. note, “the clusters arise not as an artifact of the sampling scheme, but from small discontinuous jumps in genetic distance for most population pairs on opposite sides of geographic barriers, in comparison with genetic distance for pairs on the same side”. They are of course not the only authors to have found this result. Xing et al. conclude: “Patterns of human genetic variation are influenced by mating patterns, and the latter are in turn influenced by geographic and cultural factors […] Consequently, it is not surprising that human genetic variation, while correlated with geographic location, is not perfectly clinal.”

I would also note that, even if human genetic variation were perfectly clinal, it would not matter for the hereditarian hypothesis. On average, nations and self-identified races would still differ from one another genetically, and in principle those genetic differences could still contribute to observed differences in their average IQs.

Some ‘human biodiversity’ proponents concede that traditional notions of race are refuted by genetic data, but argue that the complex patterns of ancestry we do find should in effect be regarded as an updated form of ‘race’. However, for geneticists, other biologists and anthropologists who study this complexity, ‘race’ is simply not a useful or accurate term, given its clear and long-established implication of natural subdivisions.

I am not well-acquainted with the work of early anthropologists and biologists who studied ‘race’, but I am told by colleagues that they had a much more sophisticated understanding than critics — such as Birney et al. — typically give them credit for. Since Birney et al. do not provide any evidence for their assertion that ‘race’ has a “clear and long-established implication of natural subdivisions”, I have no particular reason to accept it. Note that in 2015, Richard Dawkins — who has a highly informative discussion of ‘race’ in his book The Ancestors Tale — wrote: ““Social construct”? Forget it. Race is biologically real.”

It is often suggested that geneticists who emphasise the biological invalidity of race are under the thumb of political correctness, forced to suppress their real opinions in order to maintain their positions in the academy. Such accusations are unfounded and betray a lack of understanding of what motivates science […] The charge that thousands of scientists across the world are covering up a real discovery for fear of personal or wider social consequences is absurd.

Contrary to what Birney et al. claim, it is highly plausible that scientists deny the existence of ‘race’ for reasons of political correctness. Indeed, in 2013 Steven Pinker wrote, “Every geneticist knows that the “Race doesn’t exist” dogma is a convenient PC 1/4-truth.” In a 2005 interview, the researcher David Goldstein noted, “Until recently, most human geneticists almost … disallowed discussion about genetic differences among racial and ethnic groups”. And As Nathan Cofnas showed in a 2016 paper, various scientists and philosophers have openly advocated lying about race differences in intelligence.

After the geneticist David Reich penned a careful and heavily-caveated article in the New York Times, entitled ‘How Genetics is Changing Our Understanding of ‘Race’’, he became the subject of an ‘open letter’ signed by 67 academics, which claimed that his “understanding of “race” […] is seriously flawed”. James Flynn’s forthcoming book, In Defense of Free Speech: The University as Censor, was rejected by his publisher on the grounds that it “could be seen to incite racial hatred”. And this was despite the fact that the author has “no intention of promoting racism” because “intent can be irrelevant”.

According to NPR’s write-up of their interview with David Epstein (author of The Sports Gene), “Academics told him they had evidence of genetic advantage but wouldn’t share their research with him for fear they’d lose their jobs.” (Note: Epstein has since clarified that the academics he spoke to were not scared of losing their jobs, but only of being “exposed to criticism”, and that their work did not concern “genetic advantage”, but rather “esoteric medical stuff”. However, he also says that the academics “were afraid that their work — which had absolutely zero to do with intelligence — would somehow be construed as if they were commenting on intelligence.”)

The fear expressed by these academics is of course quite rational, given how many witch-hunts there have been against scholars interested in the genetics and evolution of human behaviour. Examples include: E.O. Wilson; James Neel; Napoleon Chagnon; Thomas Bouchard; Sandra Scarr; and Arthur Jensen. (For a longer discussion of this issue, I would recommend reading Chapter 6, ‘Political Scientists’, in Steven Pinker’s The Blank Slate.)

There are countless traits one can measure in humans, but none more controversial than those associated with intelligence, such as IQ. ‘Human biodiversity’ proponents tend to fixate on IQ, and one can speculate about why this is and what conclusions they wish to draw […]

Here Birney et al. imply, without stating it explicitly, that hereditarians might have nefarious motivations. (As James Thompson noted, they do so despite earlier referring to “accusations [that] betray a lack of understanding of what motivates science”.) Contrary to Birney et al.’s characterisation, I would suggest the main reason hereditarians “tend to fixate on IQ” is that IQ is a highly predictive variable. For example, Stuart Ritchie has described it as one “of the most reliable and valid instruments in all of psychological science”, while Linda Gottfredson has called it “psychology’s greatest single achievement”. Moreover, many socio-economic gaps between self-identified races are greatly reduced or eliminated when controlling for IQ.

Interestingly, Birney et al. actually acknowledge the predictive validity of IQ, noting that it is “a valid and measurable trait”, and that it “does an excellent job of correlating with, and predicting, many educational, occupational, and health-related outcomes”. However, they apparently cannot understand why it would be interesting to study group differences in a variable with those properties. The following quote from James Flynn seems relevant here: “I tell US academics I can only assume that they believe that racial IQ differences have a genetic component, and fear what they might find. They never admit that the politics of race affects their research priorities. It is always just far more important to establish whether squirrels enjoy The Magic Flute.”

‘Human biodiversity’ proponents sometimes assert that alleged differences in the mean value of IQ when measured in different populations — such as the claim that IQ in some sub-Saharan African countries is measurably lower than in European countries — are caused by genetic variation, and thus are inherent.

It is revealing that Birney et al. refer to “alleged differences in the mean value of IQ” when these differences are extremely well-documented. Of course, there is some debate about the magnitude of differences in mean IQ between nations and self-identified races. But there is essentially no dispute about whether they exist. Later in the blog post, the authors again refer to “apparent population differences in IQ scores”. (This is despite the fact that they acknowledge IQ is “a valid and measurable trait”.)

Such tales, and the claims about the genetic basis for population differences, are not scientifically supported. In reality for most traits, including IQ, it is not only unclear that genetic variation explains differences between populations, it is also unlikely.

As Bo Winegard pointed out, it is noteworthy that Birney et al. dismiss evolutionary explanations for group differences as mere “tales”, even though similar explanations have been put forward in the non-human animal literature. (Unlike some commentators, they do not employ phrases like “scientific racism” or “intellectual racism” — a gesture for which they should presumably be given credit.) The authors proceed to explain why they regard the hereditarian hypothesis as “unlikely”.

However, as several people pointed out, Birney et al. give short shrift to the hereditarian hypothesis. They focus on the limitations of GWAS studies (and some of their points in this regard are well-taken), but fail to discuss all the other evidence that has been adduced in support of a genetic contribution to group differences, including recent admixture studies. (Note, I am not claiming that this evidence is irrefragable, but simply that the authors do not even discuss it.)

As Holtz pointed out, the psychologist James Lee mentioned admixture studies as a potential way to disentangle the hereditarian and environment-only hypotheses in his 2013 review of Richard Nisbett’s book Intelligence and How to Get it. Specifically, he wrote, “If Nisbett is truly confident that degree of European ancestry shows no association whatsoever with IQ, he should call for studies employing superior ancestry estimates”. Like many others, I would be genuinely interested to hear the authors’ critiques of recent attempts to utilise admixture analysis.

At this point, it is worth quoting from Angela Saini’s account of her interview with David Reich, which appears in Chapter 7 of Superior: The Return of Race Science. According to the author, “[David Reich] suggests that there may be more than superficial average differences between black and white Americans, possibly even cognitive and psychological ones, because before they arrived in the United States, these population groups had this seventy thousand years apart during which they adapted to their own different environments.”

The genetic variants that are most strongly associated with IQ in Europeans are no more population-specific than any other trait. To put it bluntly, the same genetic variants associated with purportedly higher IQ in Europeans are also present in Africans, and have not emerged, or been obviously selected for, in recent evolutionary history outside Africa.

As Gregory Cochran noted, whether any particular allele is “present” in one population and not another is largely irrelevant. What matters is whether the frequencies of alleles affecting IQ vary across populations. To quote Cochran, “Many of the alleles that influence height or behavior in dogs are found in most or all breeds, but that does not mean all breeds of dogs are the same. The frequencies of those alleles differ.”

[…] the genes responsible for determining skin pigmentation cannot be meaningfully associated with the genes currently known to be linked to IQ. These observations alone rule out some of the cruder racial narratives about the genetics of intelligence: it is virtually inconceivable that the primary determinant of racial categories — that is skin colour — is strongly associated with the genetic architecture that relates to intelligence.

As several peopled pointed out, almost nobody with a serious interest in the causes of group differences in IQ believes the “cruder racial narratives” to which Birney et al. refer here. The fact that the authors decided to mention them suggests they did not give very much consideration to the arguments of those with whom they disagree.

Many socioeconomic and cultural factors are entangled with ancestry in the countries where these studies are often performed — particularly in the USA, where structural racism has historically and continues to hugely contribute to economic and social disparities

It is of course possible that “structural racism” explains differences in mean IQ between self-identified races. However, there are several observations that militate against this interpretation. For example, Japanese Americans were widely stigmatised and forcibly incarcerated during World War II, yet they typically achieve higher IQ scores than white Americans. Likewise, Ashkenazi Jews were subjected to centuries of persecution followed by one of the worst genocides in history, yet they typically achieve higher IQ scores than any other group. In addition, some black immigrants to the US (e.g., West Indians and Nigerians) earn much higher incomes than native-born blacks, most likely because they are positively selected for education.

The political scientist James Flynn observed that IQ was rising in test groups on average by around three points per decade from the 1930s onwards. Factors that account for this include improved health, nutrition, standard of living and education, but changes in genes can be ruled out.

While the Flynn effect is not yet well-understood, there is already a certain amount of evidence that whatever causes it is different from whatever causes the black-white IQ gap (see Section 2 of this blog post by Noam.) Even James Flynn himself has stated, “the magnitude of white/black IQ differences on Wechsler subtests at any given time is correlated with the g loadings of the subtests; the magnitude of IQ gains over time on subtests is not usually so correlated; the causes of the two phenomena are not the same”.

Overall, while Birney et al. make some reasonable points about the limitations of GWAS studies, their arguments for why race is “socially constructed”, and for why the hereditarian hypothesis is “unlikely”, are distinctly unconvincing.