Is It Racism, Or Is It Something Else?

Crommunist just put up an interesting post over at Free Thought Blogs about conducting research into the cognitive bases of racism. In general I think such research is a worthy endeavor; knowing that particular bits of our brains are prone to making to making certain types of categorizations is probably conducive to reducing racism in the long run. However, I think there's a huge methodological hurdle that must be considered when evaluating the import of such research: There's no empirical test for the presence of racism.

That sounds like excuse-making, I know... bear with me for a few paragraphs. What got me thinking along these lines is the following bit from Crommunist's post:

I lay the blame firmly at the feet of our stupid mammal brains. We forge unwarranted connections between variables, weaving false causation from whole cloth. When we see women kept out of the boardrooms, for example, despite the line in our Human Resources policy manual that specifically says we won’t do that, our brains helpfully fill in the blanks for us – obviously women aren’t there because women aren’t supposed to be. I mean, once you remove the most obvious barrier, that’s the same as fixing things, right? If those lazy broads can’t even figure out how to walk in the door we so magnanimously opened for them, we can shift the blame right back to them, can’t we?

I agree with Crommunist on this to a large degree; my gut tells me that discrimination (on the basis of gender rather than race in this case) is a (perhaps major) contributing factor to the lack of women in the boardroom. But if we're going to claim to be "doing science" we need more than just gut feeling; the proposition in question must be testable and, more importantly, falsifiable. Before we can say "Aha! Sexism!" we must identify some set of facts that would lead us to say "Aha! Not sexism!".

I suspect that many people, upon reviewing Crommunist's example above, would say "But of course it must be sexism; what other explanation could there be for the gender imbalance?". To which I reply "I don't know, but the mere presence of an imbalance proves nothing one way or the other". Consider, as a counter-example, the NBA. African-Americans are overrepresented; they make up 76% of the players but only 12% of the US population. Clearly there is a racial bias in hiring within the NBA; there is a marked tendency to prefer African-American players over... say... Asians. It does not, however, follow from there that this is evidence of racism at work.

Racism is not merely differential treatment by race, nor is sexism merely differential treatment by sex; said treatment must also be unjustified. To continue the NBA example: Asians tend to be shorter, on average, than African-Americans. Since there's a premium on height in basketball this fact could account for the observed difference. Or, it could be the case that NBA coaches secretly hate Asians and are using height as a convenient subtext to avoid hiring them. Again, my gut tells me that one explanation is more plausible than the other, but I'm not sure how to subject that observation to scientific scrutiny.

Which brings me back to Crommunist's post and eir discussion of the suit/jumpsuit research. What does the default condition look like in this case? That is to say, what sort of outcome would we expect in the absence of some racist subsystem somewhere in our brain? One possible scenario is that ambiguous faces would split 50/50 regardless of clothing, reflecting an utterly unbiased categorization system. But it might also be the case that the brain takes into account the fact that blacks are more likely to have low-status jobs than whites when categorizing ambiguous faces, in which case we should expect the type of bias exhibited in the research. How do we determine the baseline against which the effects of racism are to be compared?

I think that's the heart of the matter; from where I'm sitting it doesn't look like its possible to establish such a baseline. Racism manifests itself indirectly in the physical world; racist thoughts and motivations may lead to specific actions, but its a category error to label an action as "racist". Methodological naturalism, on the other hand, deals only with what can be observed; it can verify that a particular set of circumstances occurred, but is unable to speak to the motivations which lie behind them. Since racism is ultimately about motivation this would seem to remove it from the domain of scientific observation. Which means that research such as that cited by Crommunist is interesting, but ultimately of less import than it seems at first blush.