What’s an -ism?

If you’ve read one of my previous rants, you may recall my thoughts on what I think are pervasive “-ism”s and prejudices in America. (If you haven’t, and you’re bored, you might want to check it out now, since I’m going to be discussing some of the same ideas below.)

So, yeah, like most people, I think isms (e.g. racism, sexism, orientationism, etc.) are bad. The problem is, I never really bothered to define what, exactly, an ism is. What constitutes X-ist behavior? I thought about it for a while, and I was surprised at how difficult it is to pin down. For instance,

• If you are meeting a new client for lunch, one whom you’ve never seen before, and his name is (say) Yao Ming, is it racist if you expect him to look Chinese?
• If you see a man and a woman in a footrace, is it sexist if you think the man is going to win?
• If you’re getting in line at the grocery store, is it “looks-ist” if you choose to stand in the checkout line with the more attractive employee?
• What if you’re a bank employee and several potential clients are waiting for service, and you pick a more attractive one?

A preliminary definition
The crux of the matter is as follows: humans are really good at making generalizations. It’s a very useful skill. Imagine having to decide where to eat if you couldn’t generalize about the quality of a restaurant’s food. Or trying to marry someone without being able to generalize about his or her long-term behavior. (Maybe people do this already :). Or trying to budget your time between different activities without being able to generalize about your past experiences with them, and which ones you liked and didn’t like.

One problem, though, is that people are so good at generalization and inference that sometimes we infer too far. The result is superstition. Many superstitions are harmless, but some clearly aren’t, and when you start acting on one, you’re exhibiting prejudice: you have a preconceived expectation that isn’t based in fact.

So here is a baseline definition for isms: You are X-ist if you exhibit behavior in accordance with a belief about X that is not corroborated by empirical evidence.

For instance, a KKK member denying the Holocaust is exhibiting anti-Semitism. (Exactly why is an exercise for the reader.)

But it’s not that simple…
However, isms are far more devious than that. Often a belief is grounded in statistical fact. (Consider the example scenarios I gave above.) In what situations is it still okay to act upon such a belief?

Consider one more hypothetical situation: you’re in charge of hiring new employees at a company. You have a considerable amount of information about each candidate: college attended, GPA, SAT score, recommendations, economic background, race, etc. On what should you base your decision? Each bit of information may have some predictive power as to the candidate’s eventual success at the company (say, college GPA) or not (say, whether or not he likes Cinnamon Toast Crunch). Clearly some bits are more predictive than others. This leads us to a more comprehensive definition:

If you have a decision to make, and a number of predictive factors exist, you are X-ist if you base your decision on X even though a more predictive factor either a) is at your disposal or b) can be obtained with effort commensurate with its increase in predictive power.

Haha, this sounds pretty opaque! What I mean is this: If you’re hiring someone, and you make the decision based on race, chances are you’re being racist, since you have more predictive factors at your disposal, like college, GPA, and recommendations.

It’s important, though, that even if the factor is not immediately at your disposal, it is your obligation to try to obtain it, if doing so can be done with a reasonable amount of effort, say proportional to how predictive it is and perhaps other factors relative to your situation. No doubt this is a foggy notion, one that could probably be debated endlessly. (But hey, already many laws in this country are foggy, and we leave it to the courts to decide.) Here is a clear example:

To the dismay of gay-rights activists, the Food and Drug Administration is about to implement new rules recommending that any man who has engaged in homosexual sex in the previous five years be barred from serving as an anonymous sperm donor.

The FDA has rejected calls to scrap the provision, insisting that gay men collectively pose a higher-than-average risk of carrying the AIDS virus. Critics accuse the FDA of stigmatizing all gay men rather than adopting a screening process that focuses on high-risk sexual behavior by any would-be donor, gay or straight.

“Under these rules, a heterosexual man who had unprotected sex with HIV-positive prostitutes would be OK as a donor one year later, but a gay man in a monogamous, safe-sex relationship is not OK unless he’s been celibate for five years,” said Leland Traiman, director of a clinic in Alameda, California, that seeks gay sperm donors.

In this case, the more predictive information (high-risk behavior, actually has HIV) is not immediately available, but can be obtained via interviews or tests — and yet the FDA is going by a much poorer predictor, sexual preference. So this is, in my book, an ism.

On the other hand, if your new client is named Yao Ming, you could try to find a picture of him, since that would be a better predictor… but not much better, since the vast majority of people named Yao Ming look Chinese. So unless the picture is very easy to obtain, I think it would be all right to go ahead and expect the client to look Chinese.

Real life
Unfortunately, real life is very complicated, so applying this rule is not always easy in practice. In particular, it is possible for a generalization to be right (accurate) for the wrong reasons, namely: a) due to past wrongs or b) due to the fact that the generalization is widespread and self-fulfilling.

For instance, if there is a widespread belief that members of a certain demographic are not good workers, then companies will be less willing to hire them, thus riddling their resumes with periods of unemployment — and so future employers may feel that, based on their resumes, they have less job experience than their white counterparts and so are not as good workers…. and then not hire them as a result.

This is nefarious and pervasive. Of course, you’d like to nip this stuff in the bud (the original round of companies not hiring would be racist), but you can’t always, especially when there are historical factors in play (past wrongs) that can set up all kinds of crazy scenarios. So some generalizations should be fought, even if they are currently true. I guess I could talk about this more, but this entry is far too long as it is…

The scenarios
As for the the scenarios I listed at the beginning of this entry, well, I think some of them are isms and some are not. I’ll let you draw your own conclusions.