depth first search

The blog Computational Complexity usually focuses on issues of, well, computational complexity, but sometimes a post veers off in another direction, like this particular post on the relative strength of French engineering:

In my last post one of the comments asserted without justification that The French will never be as good at Engineering as the Germans. Rather than ponder if this is racist or bigoted, I would challenge the poster to either give HARD EVIDENCE that this has been true in the past, and a REASON to think it will continue. I still stand by the notion that globalization will make all of these local factors go away. Mainly because locality is not longer as important as it once was.

Then there’s this comment later on:

In retrospect, I do not have hard evidence of why french are not as good engineers as germans, and perhaps this is not correct (perhaps it is though). Because I cannot defend it, it was a bad example.

But I would defend myself if such a statement is called bigoted or racist. Whether it is depends on how people interpret it. I interpret it as follows:

We can measure every french person at how good they are at X, and plot the distribution (f). We can do the same for Germans (g). When I make a statement that “the french are not as good as germans in X” I am saying that the expected value of f is significantly lower than g. I am not saying about why this is (it could be connected to culture, or it could be connected to the weather, or it could be connected to the fact that the french never had any need to do X). I am also not saying anything about a particular french person and a particular german person. There is a prior on whether the german is a better engineer, but this prior is usually insignificant given that I would probably have much more information about this people besides that they are french and german.

I don’t like the term racist or bigoted in this context. The term racist does not seem to apply because France, as a country, consists of people of many different races. For the statement to be bigoted, it would have to reflect some notion of intolerance. Saying the French are bad at engineering does not strike me as intolerant as, for example, saying the French are rude (a seemingly more common opinion among Americans). It does strike me as prejudicial, which is the term I am going to use to denote this particular pattern of American/French animosity.

What I find interesting about the comment is the appeal to formalism, and the assertion that true prejudicial remarks need to explicitly cite a cultural cause or single out an individual. The appeal to formalism is both fascinating and baffling. Consider these statements:

White men are stupid.

The mean intelligence of white men is considerably less than all other races and genders.

If we plot the distribution for intelligence of white men (wm) against the any distribution of a different race or gender (rg), then we will see that the expected value of wm is significantly lower than rg.

Adding progressively more formalism turns the original and obviously prejudicial statement into something that is still obviously prejudiced. Also consider that none of the example statements above make causal claims or single out any individual, e.g. why white men are stupid or I am stupid (since I am a white man). Yet they are still prejudicial.

I’m perfectly happy to give this particular commenter the benefit of the doubt. We all carry around biases that, when exposed to even simple skepticism, turn out to be unsubstantiated, prejudicial, or otherwise flawed. What interests me are the various ways that people (and I am definitely guilty) try to rationalize these biases after they have been exposed. Part of the problem, I think, is that in this culture prejudice, racism, or bigotry have become things for which apologies are never enough, and so there is always some utility in trying to argue your way out of acquiring that kind of label.