The Guardian being right about the genetics of IQ is unlikely
In the recent Guardian article Why genetic IQ differences between ‘races’ are unlikely,there is presented several arguments for why IQ differences between population groups are unlikely. They are all wrong.
[Addendum: The article is written by Kevin Mitchell, who is otherwise sensible on genetics, and I will applaud him for the polite tone in the piece. (Far from a given on this topic.) We also agree about the conclusion he writes at the end, that it is best to focus on things we can change to ensure that everyone can reach their potential. Additionally, there is currently a large amount of uncertainty regarding genetic architectures and population differences. However, none of these things change the fact that the article does not present a good argument.]
The main argument seems to be that intelligence has a highy polygenic architecture, and that most mutations decrease IQ. This is likely correct, but it doesn’t lead to the conclusion drawn in the article, that it’s highly unlikely that selection effects can cause differences between populations. There is a large amount of common variation in IQ, this variation explains a large part of the heritability, and many of these variants are IQ increasing variants. (source). Selection acts on mainly through soft sweeps affecting the trait through altering the frequencies of many common variants in the population, and there is plenty of varation for them to act on.
Some other arguments in the article:
[…] thousands of variants affect intelligence, they are constantly changing, and they affect other traits. It is not impossible for natural selection to produce populations with differences in intelligence, but these factors make it highly unlikely.
It is correct that IQ is affected by a large number of variants (e.g. it is highly polygenic), and also that many of the variants that affect intelligence also affect other traits (e.g. it is pleiotropic). But it doesn’t follow from this that population differences are highly unlikely. We can easily see this by looking at height. Height is also highly polygenic and pleiotropic. And nonetheless there is evidence of widespread selection on standing variation in Europe at height-associated SNPs, mirroring the differences in height between Northern Europeans and Southern Europeans. Additionally, other theoretical work shows that pleiotropy doesn’t decrease the ability to respond to selective challenges.
To end up with systematic genetic differences in intelligence between large, ancient populations, the selective forces driving those differences would need to have been enormous.
This is not correct. Selective forces do not need to be large just because the population is large. This is quite straightforward to convince yourself of. If parents with lower IQ tend to have fewer offspring, this would affect the proportion of low IQ descendents whether the total population is small or large. In a sense it is actually easier for selection to work on large populations, since the effect of drift is comparably lower. The effect of selection can be described by the breeders equation, which works for polygenic traits:
R = h² * S
R is the change in mean over one generation, h² is the narrow sense heritability, and S is the selection differential. The narrow-sense heritability of IQ is high, so a selection differential will lead to a proportional response. And even a small differential can easily compound to large differences between populations over generations.
What’s more, those forces would have to have acted across entire continents, with wildly different environments, and have been persistent over tens of thousands of years of tremendous cultural change. Such a scenario is not just speculative — I would argue it is inherently and deeply implausible.
This is also not correct. The selective forces don’t have to work over entire continent, just in specific population groups. It is quite easy to see that this is the case by considering pygmies. They are on the African continent, yet they are genetically shorter than other African populations. And the selection effects don’t need to be at a persistent level for tens of thousands of years. For example, many dog breeds have changed dramatically over a shorter period.
The result is that any population at any time will carry a varied bunch of mutations that affect intelligence. These will differ between populations, clans, families, and individuals. This constant churn of genetic variation works against any long-term rise or fall in intelligence.
This is a rather unclear statement. How exactly would a churn work against rise in intelligence? Also this quote seems to argue against the main claim in the article. If the mutations differ highly between populations, with different population sizes, selection and strength of genetic drift, it is unlikely that the populations would end up at the exact same sum of effects of detrimental variants.
Unlike height, where being ever taller had no benefit, strong evolutionary forces drove intelligence in one direction only in our ancient ancestors.
These claims may be correct, and the other arguments in the article would still be unfounded. However, additionally, these claims are also doubtful. Being taller as a man has the benefit of being more attractive to females, which is important for fecundity through sexual selection. This article describes selection for height in Britons over the last 2000 years. Regarding intelligence, there has clearly been a strong positive selection in humans in some periods, but it is not clear that this has been in one direction for all populations at all times. Here is an article describing Selection against variants in the genome associated with educational attainment, which is known to correlate strongly with IQ. And these are just two counter-examples that we happen to know of.
In any case, in a few years there will most likely be performed an analysis similar to the one performed for height, and we will have a more definite answer to the question adressed in the article. Until then it behooves us not to be convinced by weak arguments, and keep an open mind about what the results will show.