Noted hacker Eric S. Raymond (ESR) recently posted on his blog about a development that he thinks is forthcoming from genetics: the fact that intelligence is largely determined by genetics, and that different racial groups have significant differences in average intelligence. He’s dismissive of the “secular piety” that has kept this idea down, but he’s also not sanguine about what it will mean when the truth of racial difference is acknowledged:

It’s not going to be easy. I saw this coming in the mid-1990s, and I’m expecting the readjustment to be among the most traumatic issues in 21st-century politics. The problem with repression, on both individual and cultural levels, is that when it breaks down it tends to produce explosions of poorly-controlled emotional energy; the release products are frequently ugly. It takes little imagination to visualize a future 15 or 20 years hence in which the results of behavioral genetics are seized on as effective propaganda by neo-Nazis and other racist demagogues, with the authority of science being bent towards truly appalling consequences.

He then proposes that we can overcome the racist implications of this discovery by insisting that people be evaluated as individuals, not members of a group, and that racial identity politics of both the white supremacist and black separatist kind as the true evils.

This post is not going to argue about the evidence for or against racial differences in intelligence. You have a Google; you can quickly find a lot of people presenting evidence on both sides and decide for yourself whether this is a legitimate scientific discovery or just a racist in a lab coat. Instead, I’m interested in critiquing Raymond’s proposed solution. Is individuality really a sufficient response to racism?

Three objections came immediately to mind when I read Raymond’s post. First, while it might be admirable to commit yourself to always evaluating others as individuals, I doubt that you can actually do it consistently. At the very least, I can’t do it consistently. The habit of interpreting people you meet through the lens of the categories they belong to is deeply ingrained, and not just by social custom. Humans are pattern-matching machines, and slotting others into social groups based on gender, race, education. etc., is an unavoidable, even necessary social tactic. A dedication to individuality may be a noble ideal, but it’s not an ideal which can be consistently practiced, and in the moments when you’re unaware you cannot help but allow your actions to be shaped by the beliefs you’ve acquired about group generalities.

The second objection is a stronger version of the first: even if you, personally, are able to achieve total segregation between your opinions of the group and opinions of the individual, you cannot count on others to do so. You can, in fact, be virtually guaranteed that others won’t distinguish between individual and collective characteristics. Most people–even highly intelligent people–are bad at statistics. Humans have a strong bias towards essentialist categorizations, and minor statistical variations between groups are often reinterpreted as absolute characterizations of all members of those groups. Some people may be able to overcome this bias with effort, but it’s not reasonable to expect everyone to do so. The ideal of individualism is further threatened by the fact that, in this case, the scientific discovery is congruent with a social belief about racial attributes that was very widespread until very recently, and which persists in mutated form in much of our cultural discourse. If there is a difference in mean intelligence between races, this information will justify and intensify existing covert racist beliefs. Even racist ideas unrelated to the question intelligence will be strengthened by this facade of scientific credibility.

The final objection is the strongest: it’s not always possible to evaluate people without recourse to group identity. You might vow to treat everyone you meet as an individual. You might even succeed in getting everyone else to do the same. But you cannot eliminate the need to make judgements in situations of low information, when the most salient fact that you may have about someone is their race. This is the problem of racial profiling: when pulling someone over for a traffic violation, the officer knows next to nothing about the other driver. But he can see the driver’s skin, and he knows which races have higher rates of violent crime. Is he allowed to use this information? Is he even able to not use the information? It’s there in his subconscious, whether he thinks he’s going to act on it or not. Now generalize this dilemma across the whole spectrum of situations in which intelligence is a deciding factor. Can this be anything other than disastrous?

This affects even highly intelligent members of the low-intelligence groups. Regardless of how intelligent you are individually, you will sometimes be in situations where you cannot demonstrate your intelligence. In those situations, those around you will fall back on their beliefs about people with your visible characteristics, and if those beliefs are largely negative you will suffer negative consequences. Lots of people will be able to succeed regardless, of course, but the effect of negative expectations will be visible across the whole group. And this drop in acheivement will further encourage racist expectations, creating a positive feedback loop that retards the potential of everyone in the low-intelligence group, even if they are individually exceptional.

What’s most striking is that this, in practice, is no different than the old racist hypothesis. It has always been possible for individual geniuses to get education, wealth, and a measure of prestige, though the racist expectations of society made it more difficult for them. But those of merely average talent and dedication (which is most of us, after all) get no such consideration, and the obstacles of racism prevented most of them from acheiving what they otherwise would have. If we resurrect the theory of important racial difference, we’ll get the same result. Even if this time we are supposedly backed up by science, and even if we temper statistics with a nod towards individualism.

The comment thread on ESR’s post was a dismal thing to read, as a big chunk of it was white guys congratulating themselves for being brave and honest enough to say that black people are less intelligent. This was every bit as tedious and cringe-worthy as it sounds. (I refrained from jumping into the fray, as by the time I got there the thread had over 100 comments, which is well past the point at which all comment threads devolve into shrieking and poop-flinging.) I was most interested in the argument that developed between one commenter, Jessica Boxer, and several others, which recapitulated several of the points that I’ve made here. (It also contained a typical Bob-and-Alice moment, as one commenter condescendingly explained to Jessica that she was “muddled about racism” because she tried to bring in some social considerations broader than raw prejudice.) Jessica posed the dilemma of judgement in low-information situations in a series of hypotheticals spread across several posts:

1. Running from a group of young black men dressed in what looks like gang paraphernalia on the belief that such men are more prone to violence (link).
2. Choosing a black dance partner over a white dance partner on the belief that the black man is more likely to be a good dancer (link).
3. Hiring a white man over a black man on the belief that the white man is more likely to be intelligent and a hard worker (link).

The reaction to these hypotheticals was interesting. Every commenter accepted the first as a rational response. Most of them accepted the second. And the third–which I thought was clearly parallel to the first two–mostly resulted in people denying the hypothetical. That is, most responders (including ESR himself) insisted that there has to be a better way to evaluate the job candidates to avoid the need to make race the determinant. The same objection could be raised to the first two hypotheticals, of course, so why did this only come up for the third? Probably because the last example is more obviously “racist” than the first two, and more harmful detrimental to the disfavored candidate.

But given the premises of ESR’s argument, the interviewer is just as justified in making his racist choice in this scenario as he is in the first two scenarios. And this is precisely the problem: now matter how thoroughly you distinguish between individual ability and the group, sometimes group identity is the tipping factor on the scale.

Which brings us back to my original, depressing point. If there really is good evidence for racial differences in intelligence, then the resurgence of ugly and harmful racism is inevitable. Racism is not a problem that can be solved by a commitment to individuality. The best solution may in fact be the one that we’ve been pursuing for the last forty years: make racism a social taboo, and suppress its expression in polite society. And for that reason, I hope that we never get the evidence that ESR thinks is forthcoming.