When thugs menace someone because he looks Arabic, that’s racism. When airport security officials single out Arabic-looking men for a more intrusive inspection, that’s something else. What is the difference? The difference is that the airport security folks have a rational reason for what they do. An Arab-looking man heading toward a plane is statistically more likely to be a terrorist. That likelihood is infinitesimal, but the whole airport rigmarole is based on infinitesimal chances. If trying to catch terrorists this way makes sense at all, then Willie-Sutton logic says you should pay more attention to people who look like Arabs than to people who don’t. This is true even if you are free of all ethnic prejudices. It’s not racism.
But that doesn’t make it OK. Much of the discrimination that is outlawed in this country—correctly outlawed, we (almost) all agree—could be justified, often sincerely, by reasons other than racial prejudice. Without the civil rights laws, employers with nothing personal against blacks might well decide that hiring whites is more cost-efficient than judging each jobseeker on his or her individual merits. Universities could base their admissions policies on the valid assumption that whites, on average, are better-prepared for college. Even though this white advantage is the result of past and present racism, these decisions themselves might be rational and not racially motivated.
All decisions about whom to hire, whom to admit, whose suitcase to ransack as he’s rushing to catch a plane are based on generalizations from observable characteristics to unobservable ones. But even statistically valid generalizations are wrong in particular instances. (Many blacks are better prepared for college than many whites. Virtually every Arab hassled at an airport is not a terrorist.) Because even rational discrimination has victims, and because certain generalizations are especially poisonous, America has decided that these generalizations (about race, gender, religion, and so on) are morally wrong. They are wrong even if they are statistically valid, and even if not acting on them imposes a real cost.
Until recently, the term “racial profiling” referred to the police practice of pulling over black male drivers disproportionately, on the statistically valid but morally offensive assumption that black male drivers are more likely to be involved in crime. Now the term has become virtually a synonym for racial discrimination. But if “racial profiling” means anything specific at all, it means rational discrimination: racial discrimination with a non-racist rationale. The question is: When is that OK?
The tempting answer is never: Racial discrimination is wrong no matter what the rationale. Period. But today we’re at war with a terror network that just killed 6,000 innocents and has anonymous agents in our country planning more slaughter. Are we really supposed to ignore the one identifiable fact we know about them? That may be asking too much.
And there is another complication in the purist view: affirmative action. You can believe (as I do) that affirmative action is often a justifiable form of discrimination, but you cannot sensibly believe that it isn’t discrimination at all. Racial profiling and affirmative action are analytically the same thing. When the cops stop black drivers or companies make extra efforts to hire black employees, they are both giving certain individuals special treatment based on racial generalizations. The only difference is that in one case the special treatment is something bad and in the other it’s something good. Yet defenders of affirmative action tend to deplore racial profiling and vice versa.
The truth is that racial profiling and affirmative action are both dangerous medicines that are sometimes appropriate. So when is “sometimes”? It seems obvious to me, though not to many others, that discrimination in favor of historically oppressed groups is less offensive than discrimination against them. Other than that, the considerations are practical. How much is at stake in forbidding a particular act of discrimination? How much is at stake in allowing it?
A generalization from stereotypes may be statistically rational, but is it necessary? When you’re storming a plane looking for the person who has planted a bomb somewhere, there isn’t time to avoid valid generalizations and treat each person as an individual. At less urgent moments, like airport check-in, the need to use ethnic identity as a shortcut is less obvious. And then there are those passengers in Minneapolis last week who insisted that three Arab men (who had cleared security) be removed from the plane. These people were making a cost, benefit, and probability analysis so skewed that it amounts to simple racism. (And Northwest Airlines’ acquiescence was shameful.)
So what about singling out Arabs at airport security checkpoints? I am skeptical of the value of these check-in rituals in general, which leads me to suspect that the imposition on a minority is not worth it. But assuming these procedures do work, it’s hard to argue that helping to avoid another Sept. 11 is not worth the imposition, which is pretty small: inconvenience and embarrassment, as opposed to losing a job or getting lynched.
A colleague says that people singled out at airport security should be consoled with frequent flier miles. They’re already getting an even better consolation: the huge increase in public sensitivity to anti-Muslim and anti-Arab prejudice, which President Bush—to his enormous credit—has made such a focal point of his response to Sept. 11. And many victims of racial profiling at the airport may not need any consolation. After all, they don’t want to be hijacked and blown up either.