A Copernican Revolution of social science – Race and IQ

Race is real, black IQ is likely low for genetic reasons, but there is nothing to fear

Fields of science sometimes get stuck because they are based on wrong models and assumptions. Most famously, medieval astronomy adopted the wrong Ptolemaic model when the planets revolve around a stationary Earth and it needed the Copernican Revolution to adopt the correct heliocentric view which gave birth to modern physics. I think social science also needs a kind of Copernican Revolution unless we want to spend untold amounts of time, money and energy on policies that won’t go anywhere.

Most people are not particularly offended by theories of planetary movement, so the Copernican Revolution was relatively smooth. In contrast, the correct theories about society often sound preposterous because, being about biological creatures, they have on them the cruel cold touch of nature. For a long time there has been both a niche market for loud, “based” truth-sayers about human nature and serious researchers mostly staying quite outside of academia, but not much penetration of this new model into polite society or policy.

I plan to publish a series of articles to document the most controversial and provocative views which are likely true and important for public policy.

Race is real, black IQ is likely low for genetic reasons

Race is real

Evolution – the change of allele frequencies over time – is real. Four great evolutionary forces act on the genomes of all species: mutation (the appearance of random changes in the genome), selection (fitter individuals have more surviving offspring so subsequent generations are more like them), gene flow (migration) and drift (randomness or luck). When members of a population mate relatively freely – like people within the same city – these forces act on them all. But when two populations are in what we call reproductive isolation – that is, for some reason they don’t get to meet and mate – these forces act on them separately unless and until gene flow mixes them again.

As it happens, humanity has traditionally consisted of groups in reproductive isolation. It’s hard to cross the Sahara, the Atlantic, Siberia or the Himalayas so people on one side tended to keep together and vice versa.

From this paper using modern genomes to estimate historical human gene flow (=migration) and the differentiation of humans into reproductive clusters. The world map shows that geographical barriers prevented migration and interbreeding, and nothing was quite as formidable as the Sahara. The PC plot shows genetic clusters. There sort of are people with every degree of European and Asian admixture, but of course Siberia and Central Asia are much less sparsely populated than Europe and East Asia where the centers of these clusters are. The separation of Europeans and Africans is much more complete.

We are talking about considerable time frames here. Modern humans and Neanderthals ended up largely (but not completely) in reproductive isolation about 600k years ago. The most exotic extant human populations like the Khoi-San (Bushmen) or pygmies did this up to 350k years ago. If you ever wondered about how exotic Neanderthals might have been just look at Bushmen and double that (some assumptions apply). If Neanderthals were alive today, I think there would be affirmative action in their favor and it would destroy your career if you implied that they are a different race or that they have a negative genetic predisposition to anything socially important.

With Africans and everybody else, the split happened about 100k years ago. A typical European and a typical East Asian had their last common great-grandparent about 30k years ago. Within Europe, there has been a lot of mixing but genetic differences still exist because people have always had children with those around them. This is why 23andMe and similar companies made money: using reference populations worldwide they could pinpoint pretty precisely where your ancestors came from, even if all of these places are in, say, Northwestern Europe.

Separation of human genetic clusters from this paper. The time estimates may change from source to source but they are always in the same ballpark.

Traditionally, members of these different population were said to belong to different “races”. Today, the existence of race is either a dated right wing conspiracy theory (when you point out genetic differences between them) or the most significant thing in the world (when you suggest that some races should get preferential treatment via DEI policies). In the field of genetics, you best use euphemisms like “ancestry” when you want to talk about race. This preference is not wrong: “ancestry” is a much more precise and descriptive term than “race” and reflects even better how these groups are real. Two people have different “ancestries” when they had different ancestors and this is exactly where race comes from: reproductive isolation.

There is a lot of stuff you can throw at the idea of “race” if you don’t like it, none of it undermining the core of the idea but not all of it wrong or stupid either.

First, race is not some ideal Platonic entity as some imagine. History can always split a “race” in two by creating reproductive isolation. Reproductive isolation also remains completely reversible for a long time and two “races” can always mix again, both ceasing to exist and creating a new one. We have seen the creation of at least one new race within recorded history: after the colonization of Southern America, Europeans and Native Americans, two “races” separated by about 15k years of reproductive isolation, met again. A lot of Africans were added in some countries too, resulting in highly admixed populations like Brazilians who are arguably a new “race”.  Such a population never existed before 1500. European Whites, who are now seen as a single race, are themselves the product of several mixing ancient populations, most notably the Western Hunter-Gatherers, Early European Farmers and invading steppe people from the East. You can imagine some Bronze Age Hitler warning his charioteers not to race-mix with the swarthy Sardinian-looking EEF ladies to keep their race pure but in our time you probably had to look up what these “races” even were because all you know is generic Europeans.

Second, races are not perfectly split, this mixing that I just said is possible is always happening somewhere on the fringes. There are lots of people who do not clearly belong to one or another “race”, such as people in the Sahel, Omanians, African Americans (various degrees of mixed European and African ancestry) or Central Asians (mixed European and East Asian ancestry). Quibbling about race with this “but no race is pure” argument is a special case of the Sorites paradox: how many grains of sand do you need to take away from a heap of sand before it stops being a sand heap? But fringe cases do not make categories useless. “Red” and “yellow” are real colors just because orange also exists, and the existence of admixed individuals says nothing about the fact that there are very large clusters of people on Earth who did not have common ancestors for tens of thousands of years so evolutionary forces could drift their phenotypes apart.

A fact I never saw used against this lumper-splitter critique of race is that the size of admixed and unadmixed populations between the major continental races is very different. Humanity is not like a smooth real rainbow but big color bands with tiny strips of mixed hues in-between. Ther are about 47 million African-Americans, the Sahel countries (Senegal, Mauritania, Mali, Niger, Burkina Faso, Chad and Sudan) have a combined population of 114 million people, and Yemen and Oman have a combined population of about 20 million. This 182 million is about the highest possible estimate of people with mixed European and African ancestries – an overestimate because many Sahel people are pure African and many Arabians do not have significant African ancestry. On the other hand, there are 1.138 billion Sub-Saharan Africans, 235 million White Americans, 445 million people in the EU, and 135 million Russians for just a very conservative estimate of unadmixed African and European populations. (I got all numbers from Wikipedia.) This contrast is even bigger in a place like the United States where none of the constituent “races” aside from Native Americans are indigenous. Sure, there are mixed Euro-Africans somewhere on the world map but people migrating to the US initially had either 100% European or 100% African ancestries so it made perfect sense for them talk about categorical “races”. Even today after some admixture, Americans are pretty good at knowing which continent their ancestors came from. Self-reported “race” on censuses matches with genetic ancestry clusters something like 99%+ of the time. For practical purposes, especially in the US context, race is a good enough and practically useful index of the ancestry and resulting genetic makeup of the people living there.

Black IQ is likely low for genetic reasons

For these millennia all the evolutionary forces acted separately on these populations. It certainly was possible for human genetic clusters to evolve to be different.

In some cases, this obviously happened, visibly to the naked eye. Most people would accurately tell apart typical Swedes from typical Southern Italians, still relatively similar populations. Every person on Earth could tell apart an African and an European. It’s not just the skin color either, facial features, skin texture and many other things are different, all the result of divergent evolution.

African children with albinism. Despite a light skin color, they don’t look European.

In other cases, evolution didn’t appear to favor a very different phenotype in isolated populations. For example, non-Hispanic White males in the US have an average height of 177.4 cm, while Black males are on average 175.5 cm, a small difference given how much divergent evolution worked on their ancestors and the fact that White immigration to the US tended to originate from the taller European countries like Britain, Germany or Sweden.

The question is of course if socially important traits like intelligence or impulsivity resemble skin color or height, and whether there are evolved race differences in these.

This question is relevant for the comparison of all ethnic groups, but it only really blew up when it was about American Blacks and Whites. I don’t think there is a topic comparable to the genetics of Black American intelligence in its importance, the number of smart people engaging with it and the quality of their results, and the degree of exclusion of the topic from mainstream discussion. I’m not even pretending to put together a serious reading list about this topic because of it’s too big: here are some links to writings and collections by better people. For me, the most convincing arguments for a genetic cause of low Black intelligence were the following:

-          Measurement invariance of the US Black-White IQ gap. This basically means that in batteries with multiple tests the better a test alone approximates the total intelligence score, the larger the BW gaps are on it. (Strictly speaking this is what we call a Jensen effect, but measurement invariance also implies this and it’s an easier way to explain it.) This is what you expect to see if Blacks and Whites differ on the very thing (general intelligence) the battery is designed to measure, not some minor test-specific things like speed or vocabulary. Unlike in some samples of South Africans and European immigrants, the tests measure the same construct in the two groups and any difference is a difference of this construct, not something related to the manifest test content itself like Whites being more familiar with certain words.

-          Constancy of the race gap. Despite the US having gone from an openly racist anti-Black country to arguably an openly racist anti-White country within just over half a century, Whites and Blacks have always differed in their cognitive test performance by about 1 SD, suggesting that racism or social disadvantage caused by official policies has little effect on performance.

-          Black-White differences in test scores and other socially important metrics are larger in more Northern, more liberal, arguably less racist parts of the country. Selective migration is probably a big cause of this, but it is a score against the racism hypothesis. If racism caused Black disadvantage, its effect should be dose-dependent and larger in the rural Deep South than in the liberal, anti-racist coastal areas. Yet, the opposite pattern is seen.

-          The variance argument. We know how much genetics matters for intelligence within-race (a lot), and we also know how much the shared environment matters (maybe a little). If differences in the shared environment – poverty, violence in the neighborhood, intergenerational trauma – caused Black-White differences, they should be enormous to balloon the small effect of the shared environment into a full 1 SD gap. Maybe you do say they are enormous, but there are a lot of poor Whites from troubled families and the numbers don’t work out. (See Warne’s paper I linked on details. While reading it note that Warne underestimates how big the environmental gaps need to be because he talks about all environment, the effect of which are substantial, while racial differences would most likely be caused by shared environmental effects, which are small at best.) People tried to get around it by proposing things called “factor X”, bad shared environmental effects only affecting Blacks. Racism is a strong candidate for a factor X – some Whites were born poor and traumatized, but none of them were discriminated against by other White for being Black – but see the arguments above about why racism is a bad candidate mechanism.

-          Admixture regression studies. These are the ones that really convinced me. Barack Obama and his dad are both called “Black” in the US but one of them is 50% African and the other 100% African. (It is correct that race is to some extent a social construct.) The Obamas are extreme cases, but American Blacks are almost never 100% African anymore and there is a lot of difference in how African they are, which we can ascertain using molecular genetic data. If Africans have lower genetic intelligence and African-European mating was relatively random with respect to intelligence (which is a reasonable assumption) then you would expect that among admixed Blacks or biracials those with more European ancestry will be smarter. You could also test mostly Europeans with some African admixture, but American Whites tend to be close to 100% European, unlike, for instance, Brancos in Brazil. This hypothesis is correct, even within the same family. (Siblings can have slightly different ancestries due to what genetic variants they randomly inherited.) Even anti-hereditarian researchers acknowledged that admixture regression would be a good test of the hereditarian hypothesis back when it seemed impossible, and now they are pulling out desperate counter-arguments along the line of how, maybe, the dosage of racism you experience precisely tracks your granular African ancestry, not your binary race. This is unlikely: skin color, the most obvious sign of race, correlates quite weakly with African ancestry and controlling for it makes no difference to admixture regression results. Other than from skin color, how would people know exactly how African you are to adequately calibrate their racism?

Admixture regression can make or break your career depending on which phenotype you use. Showing that Black ancestry contributes to diabetes risk was a big thing in David Reich’s career who later wrote this amazing book. Showing that Black ancestry contributes to low intelligence got Brian Pesta fired from his tenured university job.

I think the results are pretty much in and they show that because of different evolutionary forces operating for the last 100k years these groups spent in reproductive isolation, Africans indeed have a genetic predisposition for lower intelligence compared to Europeans, this puts them at a disadvantage in life, and much of the real-life disadvantage American Blacks have comes from here. In my view, this should at this point should at least be the working hypothesis of every sociologist and politician, maybe not as universally accepted as, say, the viral theory of AIDS but close to it. Experts largely agree with me. Yet, assuming even a partial genetic cause for American Black societal underperformance remains a job-stopper as certain as a swastika tattoo on your forehead.

Maybe you think that admitting that Black IQ is low for genetic reason would justify overt discrimination or some form of genocide against them. This certainly is the argument academics use to support censorship. People wrote a lot about how this fear is unwarranted, see for example this paper or the relevant sections of the Bell Curve. Basically, we already know and accept that certain groups of people have below average cognitive ability for biological reasons and we know what happens to them – nothing too bad. At the extreme, people with neurodevelopmental conditions like Down’s syndrome, fragile X syndrome or Williams syndrome certainly have a genetic problem that causes them to have low intelligence, and society doesn’t put them in death camps but helps them live their lives the best they can. Hate against the intellectually deficient is rare and not socially accepted. Such a disorder is usually not even a legal liability: for example, in the US federal law explicitly forbids denying your right to vote just because you have an intellectual disability.

American Blacks are of course not intellectually deficient, but we have less extreme examples of genetically low IQ groups. We know, for example, that overlapping genetic variants predispose people for low intelligence and low educational attainment, in a way that works even between siblings so causality is not up for debate. We also know that intelligence tests are measurement invariant across education levels, just like between American racial groups. Less educated people really are less intelligent than more educated people for at least partially genetic reasons, a fact at least tacitly accepted by most, and I think what happens to them is a good model of what would happen to Black people in an unbiased race realist society. The example is not perfect because having a university degree sometimes carries licensing benefits in a way I hope being White never will, but in a such a society Blacks would still have their unavoidably worse average life outcomes, while being judged based on individual merit and having their fundamental rights intact.