The evolution of racism

I've long believed that a proper scientific understanding of racism can help us engineer societies which are relatively free from racial tensions. However I haven't spent much energy writing about the topic. That is largely because of the heat the whole topic often attracts. However, I don't think that is a very good reason for avoiding discussing the issue in public. If everyone did that it would become impossible to have any sensible policy discussions. Anyway, here are some of my views on the topic:

Some worry that treating racism as a natural phenomenon, will be used to justify it as "natural" and therefore excusable. I don't really share that concern. Science explains "rape" as a natural phenomenon, but "my genes made me do it" is not treated as a valid excuse in court. It's much the same with explaining racism scientifically.

Xenophobia is likely to have had survival value for our distant ancestors. Back when outgroup members were uniformly likely to bash your head in with a rock, knowing who is part of your tribe would have been pretty important.

Humans are tribal creatures. We know empirically that humans like to form groups which exaggerate their similarities with ingroup members while exaggerating their differences from outgroup members. The result is tribal markers which are culturally transmitted, subject to rapid cultural evolution and likely to diverge quickly and easily.

It has long been observed that kin selection and homophily are associated with racism. Humans are nice to their relatives, and part of the clues of relatedness involve physical appearance. The implementation of kin recognition involves some similarity detection. Based on this, people of a different race would represent a super-stimulus of unrelatedness. The idea has been dubbed "ethnic nepotism". However, kin selection based on similarity between DNA genes probably doesn't explain racism very well. It mostly preducts outgroup indifference, not outgroup hostility. There is the phenomenon of Hamiltonian spite - which predicts active hostility to non-relatives, but this is widely thought to be a minor phenomenon.

I think that the simple intuition that kin selection is involved in racism is correct, but I also think that cultural kin selection needs to be invoked, along with cultural hijacking of kin selection mechanisms.

Other thinkers have also invoked culture in the explanation for racism. For example, here is what Richard Dawkins had to say on the topic in his 2004 article "Race and creation".

We are indeed a very uniform species if you count the totality of genes, or if you take a truly random sample of genes, but perhaps there are special reasons for a disproportionate amount of variation in those very genes that make it easy for us to notice variation, and to distinguish our own kind from others. These would include the genes responsible for externally visible "labels" like skin colour. I want to suggest that this heightened discriminability has evolved by sexual selection, specifically in humans because we are such a culture-bound species. Because our mating decisions are so heavily influenced by cultural tradition, and because our cultures, and sometimes our religions, encourage us to discriminate against outsiders, especially in choosing mates, those superficial differences that helped our ancestors to prefer insiders over outsiders have been enhanced out of all proportion to the real genetic differences between us.

However, Dawkins has also stated that he doesn't think that kin selection is involved. Here are comments from "Darwin's dangerous disciple":

The National Front was saying something like this, "kin selection provides the basis for favoring your own race as distinct from other races, as a kind of generalization of favoring your own close family as opposed to other individuals." Kin selection doesn`t do that! Kin selection favors nepotism towards your own immediate close family. It does not favor a generalization of nepotism towards millions of other people who happen to be the same color as you.

Dawkins goes on to mention another theory involving divergent selection:

I could imagine that racist feeling could be a misfiring, not of kin selection but of reproductive isolation mechanisms. At some point in our history there may have been two species of humans who were capable of mating together but who might have produced sterile hybrids (such as mules). If that were true, then there could have been selection in favor of a "horror" of mating with the other species. Now that could misfire in the same sort of way that the cuckoo host's parental impulse misfires. The rule of thumb for that hypothetical avoiding of miscegenation could be "Avoid mating with anybody of a different color (or appearance) from you."

The "divergent selection" theory predicts racism would be most pronounced between individuals of the opposite sex. That might be true, but I don't think any such effect is very big. Divergent selection might explain some racism, but I don't think it is a very good or complete explanation.

I don't think Dawkins's rationale for rejecting kin selection makes sense. For one thing, he is apparently only thinking of genetic kin selection. If you take cultural kin selection into account, it becomes immediately obvious how kin selection can be applied to large groups of people who are not closely related in terms of their DNA genes. Although they might not share genes they do share memes. Patriotism memes can convince soldiers to fall on grenades to save their unrelated soldier "brothers in arms". Kin selection is not just a theory about DNA genes - it also applies to shared cultural phenomena.

I think it is important to note that culture often exaggerates and amplifies tribal signals. These work by acting as superstimulii of relatedness. Memes involving uniforms for example dabble in kin selected psychology, and try and convince people that they are surrounded by their super-brothers and super-sisters. They frequently do this in order to encourage altruistic behavior associated with kin altruism. Memes often need their hosts to be nice and sociable to promote their own spread. Surrounded by culturally strengthened kin signals - and attempts at manipulation involving artificially strengthened kinship signals - it would probably have benefitted our ancestors to pay close attention to such signals and detect when the signals are being manipulated.

Hijacking and manipulation of kinship signals by culture gives a different dynamic to the process. Hijacking of kinship signals can be done without culture - as when a long-lost relative turns up and claims their inheritance. However cultural kin selection makes hijacking and manipulation much more common.

What about the objection to kin selection that I raised earlier? That it predicts indifference towards outgroup members, rather than hostility?

Indifference can still lead to very bad behaviour. Kin selection siggests that humans are indifferent to the fate of chickens, but that doesn't prevent humans from killing chickens in huge numbers in slaughterhouses. In the absence of kin selected altruism and altruism based on reciprocity, and reputations, it seems reasonably plausible that the default standard of behavior towards others among our ancestors involved bashing their heads in with rocks.

Anyway, this is my proposed explanation for racism. In a nutshell, kin selection, especially cultural kin selection plus some meme-gene coevolution.

Some would prefer to reframe this in terms of group selection and cultural group selection. That should make no technical difference, due to the equivalence of modern kin selection and group selection frameworks.

What are the policy recommendations associated with this idea? The obvious suggestion is to increase sharing. This could involve shared genes, or shared memes. If either promotes cooperation, then more shared genes and memes seems as though it would have a positive effect. Creating genetic uniformity could be done, for example, by promoting inter-racial marriage, or even just international travel. However, even these sorts of intervention could prove controversial. Rather than creating genetic uniformity, the most obvious policy would be to aim at creating shared memes. Get people speaking the same language, using the same money, the same software and following the same religion. That should help to make them nice to each other.

There are some downsides to this sort of proposal. Historically, globalization has led to more widely shared memes, but it has simultaneously led to larger and more powerful groups. It does look as though the general trend lines are pretty positive, but it is at least worth noting that large powerful groups can cause a lot of damage if they come into conflict with one another. The path to humans all being of one tribe is likely to lead through a state where there are two or three tribes - and that stage could potentially have some associated dangers.

Another problem is more abstract. More shared memes is likely to lead to less memetic diversity that could lead to a less effective search of meme-space and slower memetic evolution. That could have several consequences. It could result in monopolies and stagnation. It could mean that the quest for shared memes is self-limiting - as those societies that pursue it are out-competed by those with greater memetic diversity. Or it could result in some of the down-sides of monocultures - for example catastrophic parasite attacks. Diversity is, amongst other things, a defense against parasitism.

Since it is the appearance of similarity that is most important, it might be worth focusing on methods that superficially hide race-related signals. A man wearing a suit has obliterated 90% of the signals related to the colour of his skin. As technology improves the options here may also improve. Michael Jackson's racial transformation may become more widely accessible, and less surgery-intensive options involving drugs or gene therapy may become available.

Another related policy area involves what I call virtualization. It appears that armed conflicts and sports share some traits and that indulgence in sporting events substitutes and displaces armed conflict - at least to some extent. A similar approach could be used in an attempt to defuse racial tensions - again using sports or other areas where uniformed teams compete with each other. This approach seems a bit dangerous and it is easy to imagine ways in which it could backfire. However, it should at least be explored and studied.

Lastly, it seems as though a lot of progress has been made via memetic evolution of anti-racism memes. One marker for this evolution is the rapid rise of the words "racism" and "racist". Before the 1960s these words were virtually unknown. Then, between 1960 and 2000, their use rapidly skyrocketed. See this chart for more details:

The rapid rise of the terms "racism" and "racist" probably does not indicate an increase in these phenomena. Instead, it probably marks a rise in anti-racist memes. A number of modern policy efforts focus on direct suppression of racism, by spreading anti-racism memes around and persecuting percieved racists.

This seems reasonable to me, though my favoured explanation of racism doesn't really throw much light on what policies in this area would be most effective.

One thing in this area which I am concerned about the "James Watson" effect, where social justice warriors ruin the careers of otherwise respectable scientists over the issue.

The area of science covered by this blog is all about differences between humans (and groups of humans). The differences being examined are primarily cultural differences, but that topic still touches on genetic differences, since these must often be controlled for. I would hate for my preferred area of science to become a hotbed of racial controversy because of this.

One problem involves the quest to minimise percieved racial differences. Rather obviously there will be less racial discrimination if lots of people believe racial differences are small or insignificant. That results in an advocacy effort to minimize percieved racial differences. The problem is that this domain is largely a matter of fact, accessible to scientific enquiry. The advocates are naturally inclined to distort the facts in favor of their position - perhaps hoping for a self-fulfilling prophesy effect. Unfortunately, this leads them into conflict with those seeking the truth.

I think that the quest to suppress racism should take care not to run too roughshod over the facts. If you base your moral position on false facts, then it is likely that the truth will out in the end, taking the basis of your moral position with it. In particular, there's no urgent need to deny the existence of heritable differences in ability with a geographic basis when trying to help people with different ethnic backgrounds get along. There are profound differences if you consider age or sex as your control variable - instead of race. With age, discrimination based on many of those differences is enshrined in the laws concerning marriage, drinking voting, driving, etc. Differences between people that affect their abilities are OK, society can cope with them. Positions like "race is a social construct" and general denial of race related differences are not really scientifically credible positions.

The current situation is that many of the "race denialists" are fighting with scientists and with the facts. It doesn't seem as though they occupy the moral high ground. I don't think they are actually helping their own cause. They are just making themselves look scientifically illiterate.