On generalizations, racial and otherwise

You cannot think clearly about racism, if you can't think clearly at all.

You cannot think clearly about racism, if you do not understand how human thought and language uses, misuses, and is ultimately made possible by, generalizations.

The only time I ever tried to write something for actual publication, it was a book called "The Stupid Switch," on how we manage to lie to ourselves; it sprang from a twelve-week class I offered at our church. Life got in the way (as you can tell from the blog, I'm sort of busy) and I never finished the book. But I did write the chapter on the use of generalizations, which happened to use discussions of race as a teaching example. So, since Alexandra has gotten me involved in a discussion about whether it's racist to call an African-American "articulate," I thought I'd copy that chapter into the blog, even though it will be a VERY long post.

The chapter naturally assumes that you already know (from previous chapters) what I mean by "the Stupid Switch," namely that perfectly intelligent people suddenly become capable of believing the dumbest things once emotions, unquestioned assumptions ("self-evident beliefs") and prejudice get themselves involved. This chapter concerns prejudice, and its original title was:

Prejudice: When Bad Stupid Switches Happen to Good Generalizations
...continue reading...
...We’ve now covered emotions and self-evident beliefs. It’s therefore time to turn our attention to prejudice, which is, fundamentally, the abuse of generalizations.

Now, generalizations in themselves are not a bad thing. In fact they are – generally speaking, of course – a very good thing, because they make it possible for us to think and function at all. Practically all language is based on generalizations. For example, if I say, “I’m going to read a book,” you assume that the “book” is going to have sheets of paper bound together with words written on them. Most of the time you’ll be right, although it is possible that I could be planning to download a book from the internet and read it on my computer. Again, when you decide to sit down in a chair, you do so because you believe that the chair will hold your weight, because that’s what chairs are designed to do. And almost always it does.

Our words represent general classes of things – chairs, books, tables, elephants, senses of humor. Or general classes of actions – running, swimming, thinking, cheating, giggling. Or general descriptions – pretty, wretched, obnoxious, sordid, monotonous. Each such generality is an abstraction from all sorts of individual things and experiences. Once we have assembled all these generalizations into a vocabulary, then we can talk and think about the world by taking each new experience and slotting it into the general class.

So we can neither live nor think without generalizing. But of course generalizations can be abused, and this chapter addresses the major ways in which we abuse them.

First of all, of course, a generalization must be valid. That doesn’t mean it can’t have exceptions; of course it will have exceptions. But it should be true more often than not. If it’s false more often than not, then it’s invalid.

Secondly, we must know how reliable the generalization is. The generalization, “When I go into a friend’s house, I can sit down on the nearest convenient chair without worrying about whether it will break,” is a highly reliable generalization; there are very few exceptions. The generalization, “People from New Jersey are jerks,” may or may not be valid; but even if it is valid, it isn’t terribly reliable, because there are lots and lots of exceptions.

Thirdly, if we have a choice of more than one generalization, we should choose the most reliable one. Let’s say the generalization, “People from New Jersey are generally trustworthy,” is valid; and let’s say that the generalization, “People who have multiple convictions for fraud and larceny are not generally trustworthy,” is also valid. Now an ex-convict from New Jersey, who happens to have served multiple sentences for running con games, invites us to take part in a sure-fire, can’t-lose business opportunity. Do we trust him because he’s from New Jersey, or do we distrust him because of his criminal record? I think for most of us, the answer would be obvious.

In other words, the rules for using generalizations are pretty obvious once you think about it. The trouble, of course, is that the Stupid Switch makes us blind to things that are obvious. So we need to examine just exactly what happens, when the Stupid Switch kicks in, to our ability to use valid generalizations wisely.

I am going to take a major risk here: I am going to discuss the way we think, and talk, about race in America – for, like it or not, we constantly use race-based generalizations in the United States. This may mean that some violent emotions are stirred up as we go along. On the other hand, it should make it clear how important it is to understand what emotional forces can disrupt our use of generalizations.

I am White, 36 years old, five feet eleven, and a not-exactly-intimidating 155 pounds. As a teenager I was a passionate and skilled basketball player, though one would never have guessed it to look at me; and until very recently I continued to play pickup basketball on the playground from time to time. And one of the things I loved about playground hoops is that the playground court is perhaps the only place in America where race is treated rationally and honestly and (at least frequently) without malice.

Let me explain. Imagine that I go to a new playground where nobody has seen me play, and when I walk on the court I am one of fifteen available players, of whom the other fourteen are Black teenaged boys. Now, I can tell you right now that I won’t get picked for the first round of five-on-five, and I don’t resent it at all – if I were them I wouldn’t pick me either. (The logical coherence is perhaps a bit suspect there, but you see what I mean.) They are applying a perfectly valid generalization: in general, if you have to choose between a 35-year-old White guy and a Black teenager, everybody knows which one is more likely to help you keep the court.

Now a whole bunch of you are all upset about the “racism” involved here. Let me urge you to set aside the moral question here and concentrate on the logical question. That is, don’t think about what a nice person would do; think about what a smart person would do. The salient points here are that: (1) They haven’t seen me play. (2) In order for them to see me play, there has to be a game. (3) Only ten of us can play; so in order for there to be a game five people are going to have to sit out the first one. (4) If you’re a captain and your team loses, you’re off the court and at the end of the line. (5) It really is true that the typical guy who looks like me isn’t going to be able to keep up with the typical Black teenager.

There’s nothing personal or malicious about it, you see. It’s just a rational maximization of your odds of being on the court (as the first game’s winners) for the second game.

But I’ll get into the second game, because I’ll “have next,” in playground terminology.[1] And after that game the situation will have completely changed. You’ll have seen me d-up, shoot the trey, rebound, run the break, do the no-look dish. I won’t be “the middle-aged[2] White guy” anymore. I’ll be me – which may be a good thing (it used to be) and may not (it probably wouldn’t be now). If I’m good enough, then two weeks from now I’ll have my own nickname. And tomorrow when I show up I won’t be judged on the color of my skin. I’ll be judged on the content of my game.

Now I love the court precisely because, in my experience, the players make no bones about the fact that everybody knows the Black guys as a whole can spank the White guys – but if you can ball, then as soon as you show that, your skin color doesn’t matter. The broad generalization is there, it’s real, it’s valid, it’s useful when you don’t know anything particular about somebody, and as soon as the particular information about your skills is available, the broad generalization is discarded.

It’s important to realize, though, that the information about my particular skills is itself a generalization. “Sweet’s got a good offensive game” may well be generally true (I used to answer to “Sweet” when at the city park in Elgin). But my offensive game (given my physique) all starts with the long jumper; I have to force you to come out and cover me away from the hoop in order to open up the court for everything else. Sometimes the touch isn’t there, and if it’s one of those days when there’s a lid on the basket, then as long as the jumper isn’t there for me, I have at best a mediocre offensive game. So when I stop being “the middle-aged White guy over there” and I become “Sweet,” the other players don’t stop using generalizations. They just stop using a relatively broad generalization about middle-aged White guys, and start using a relatively narrow generalization about “Sweet.”

In fact, the smarter defensive players will soon figure out how much my game depends on the jumper. Then on the days when my jumper is falling, they’ll play me tight and try to force me to drive, but on the days when I’m cold from twenty feet, they’ll sink back toward the goal and try to cut off my passing angles. When that happens, then they’ve moved from a relatively broad generalization about “Sweet” to a pair of relatively narrow generalizations about “Sweet when he’s hot” and “Sweet when he’s cold.” Thus Hall-of-Fame basketball player Bill Russell once explained how he blocked a critical Elgin Baylor shot by saying something like, “Well, I know that whenever the game’s on the line, Elgin usually goes to his left hand.” Which was still a generalization, but a bloody narrow one.

Let’s go back to the first time I show up on the court and have just started getting involved in a couple of games. It could happen, of course, that I could run into a player who had some sort of racial agenda and who had made up his mind that he wasn’t going to pick White guys no matter what. In that case, he might not pick me even if it became obvious that I was one of the two or three best players on the court. But such a player’s problem isn’t the fact that he believes healthy young Black kids are usually better than skinny middle-aged White guys – that’s actually true, and there’s nothing wrong with believing the truth. The problem is that he’s insisting on sticking to the generalization even though he knows that I’m an exception. When you don’t have specific information, you don’t have any choice but to use a generalization. But when you do have specific information that invalidates a particular generalization for a given case, there’s something very wrong if you insist on sticking to the generalization anyway.

Here’s another example from my personal experience. I was raised in southeastern Oklahoma, and young men always were expected to hold open doors for ladies. It was something so deeply ingrained in me that I wouldn’t even be consciously aware of doing it. Then I went to Princeton, New Jersey to go to school. About two weeks after I got there I was going to the local convenience store and arrived about two strides ahead of a lady I would guess to have been in her mid- to late-thirties. Automatically I held the door open – and I received a brief but blistering and indeed unblushingly profane lecture on how perfectly capable she was of completing the physically and mentally undemanding task of opening a door for herself, and how despicable I was for daring to insult her by implying the contrary.[3]

After four more years at Princeton, and then sixteen years (and counting) in Austin, Texas, I can say with confidence that there are dramatic cultural differences between New Jersey women and Oklahoma/Texas women (including the strong and personally confident Southern women who can hold their own amongst any group of men). These cultural differences do not reflect any moral superiority one way or another; they are merely differences in cultural interpretations of an action. In Texas, holding open the door is interpreted as a sign of respect and consideration. In New Jersey, the same action may well be interpreted as a sign of disrespect and chauvinism.

I went to Princeton operating under a simple and naïve generalization: when you reach a door at the same time as a lady, you should hold open the door and allow her to go through first. At Princeton, I refined the generalization: when you reach a door at the same time as a lady in Texas, she will generally prefer for you to hold the door open; she will occasionally glare at you if you don’t, and she will rarely actively resent it if you do. In New Jersey, she will generally not expect you to hold open the door; she will almost never glare at you if you don’t (and if she does, that’ll mean she’s not from New Jersey but is a visitor from the South); and she will occasionally actively resent it if you do.

So if you’re a polite man and you have any sense at all, you hold the door open for ladies when in Austin and you bust right on through ahead of ’em in Newark.

But note that when you reach the door at the same time as a woman does, you have to decide whether or not to open the door for her. If you are rude, then you’ll just consult your own convenience no matter where you are. If, on the other hand, you wish to be polite, then you’ll need to take her wishes into account, and since you usually don’t know her, all you can do is guess – and your guess will be based on a cultural generalization. If you’re a redneck hick fresh from the Southern backwoods, you’ll probably use a clumsy generalization and mess it up some at first. With experience (some of it unpleasant), you’ll refine your generalizations so that they have more and more sophistication, which is to say that you’ll move from an initial stage of having a few very broad rules of thumb to having more and more, and narrower and narrower, generalizations.[4]

This is the first rule of using generalizations: you should always be moving from relatively few, relatively broad generalizations toward relatively many, relatively narrow generalizations.[5]

Now there’s a second rule as well, and to illustrate it, let’s do another hypothetical example, which I hasten to say is not based on my actual personal experience. Let’s say that I’m a single guy, and that my experience with women from Texas leads me to think that, in general, women from Texas are very self-centered and materialistic. And (just so that I can have the fun of outraging most of the women I hang out with in Austin) let’s pretend that this is a completely valid generalization. Now, a friend comes to me and tells me that he’s got a blind date for me and she’s perfect, but as he’s describing her he makes the mistake of mentioning that she hails from Beaumont. And I – applying my valid generalization – immediately interrupt him: “Oh, hey, no way. I’m not going out with a Beaumont girl. I don’t want to get stuck married to a self-centered woman who just wants me for my money.”

We all know that’s a ridiculous attitude. What I’d like you to think about is why the attitude is ridiculous. My friend, of course, is going to say something like this: “Look, Kenny, this girl is different. She isn’t like that. Besides, I’m not asking you to marry her. I’m just asking you to have dinner with myself and my wife and this girl. If you don’t like her, then what have you lost? – I’m buying, after all. But if it turns out that she’s perfect for you and that she isn’t like most Texas girls, and you don’t even get to know her because of your stupid prejudice, you could miss out on the perfect girl.”

If you look back at the situation with my not getting picked on the basketball court, you’ll note that one of the important facts is that the captains had to go ahead and pick a team. When you have to make a decision right now, then you have to make that decision based on the best information you have.[6] But in this situation, I do not have to decide tonight whether I’m going to marry the girl. There’s absolutely nothing to keep me from saying, “Well, I think I’ll wait to make up my mind until I’ve met her.” Why leap to a conclusion when I could just as well wait and find out more? Why use a broad and clumsy generalization about ten million Texas women to make an unreliable decision right now, when I could go collect information about this specific girl and then make a decision based on her and her alone?

So the second principle is simply, don’t rush to judgment.[7]

Now, here’s something to think about. If you know somebody is an exception to a general rule, it’s obviously stupid to insist on clinging to the generalization. If you don’t have to make up your mind right now, it’s obviously stupid to rush to judgment. So why do people do it?

I’ve asked this question before in a classroom situation, and I get lots of answers, all of them valid, and all of them reducing to one of two simple reasons: (1) because it’s never occurred to them that the generalization might have an exception, and (2) because they want to. In other words, either you have a “self-evident” belief, or else your emotions have kicked in the Stupid Switch. It’s really that simple.

When you mix emotions and self-evident assumptions and generalizations (even valid ones), you get prejudice. Prejudice, you see, is not a belief in generalizations; generalizations in themselves are good and useful (as long as they are valid). Prejudice either involves using generalizations that are invalid in the first place, or else in using a generalization even after you ought to realize that you’re seeing an exception to the generalization, or else in rushing to judgment before you have to. And when you act on prejudice, it’s going to be because either your emotions have stacked the deck, or else because the prejudice stems from some of your self-evident beliefs.

It is important to understand this whole dynamic when we turn to topics of race. For example, ought persons of Arab descent be scrutinized more carefully at airport security checkpoints? Well, the first thing to establish is whether such a rule would result in more effective security. And this will depend on the validity of some generalizations.

Is it, for example, true that most Arab-looking men getting onto an airplane will be al Qaeda suicide bombers? No, not at all. That’s an invalid generalization. So you can’t act based on that one.

Is it true that the vast majority of al Qaeda suicide bombers are Arab-looking men? Well, in fact that one’s true.

Is it true that if you stop and search an Arab-looking man you will have more likelihood of intercepting an al Qaeda suicide bomber than if you stop and search a White guy? Unquestionably it is. But is it also true that if you stop and search an Arab-looking man you will have more likelihooding of pestering a nice guy who just has the misfortune to look like an Arab terrorist than of actually intercepting an al Qaeda suicide bomber? Unquestionably it is.

That’s what makes the “racial profiling” controversy such a difficult issue. From the perspective of saving lives and protecting innocent people from another 9/11, there’s no reason that racial profiling can’t be an entirely rational tool. But from the perspective of protecting innocent people from constant harassment on the basis of race, there’s no doubt that racial profiling is a bad thing, since most of the people who fit the racial profile will be innocent, but will suffer the consequences of being associated with a relatively few bad guys.

Or look at it this way (people who don’t like math can skip this paragraph). Let’s say, for the sake of argument, that 0% of White men are al Qaeda suicide bombers, and that 0.0001% — one in ten thousand — Arab men in America are. From one standpoint, an Arab man getting on a plane is infinitely more likely to be an al Qaeda suicide bomber than is a White man; if you look at it that way, racial profiling makes sense. On the other hand, an Arab man getting on a plane is only 0.0001% more likely to be an al Qaeda terrorist than is a White man. And from that perspective, racial profiling seems highly objectionable.

It’s important to understand that the biggest problem isn’t ridiculously invalid racial generalizations (like “Them there coloreds jus’ luv that there watahmelon”). The problem is what to do with the valid ones. Mainstream Americans really believe that it’s a good thing to be color-blind, and therefore mainstream America went to some trouble not to blame all Arabs or all Muslims for the behavior of the 9/11 terrorists. But that didn’t stop mainstream America from feeling nervous whenever they saw an Arab man getting on the plane with them, because mainstream America couldn’t help but understand that planes with no Arab guys on them were planes with no al Qaeda suicide bombers on them. And that generalization was true. The problem, of course, is that it is also generally true that that planes that did have Arab guys on them...were also generally planes with no al Qaeda suicide bombers on them. That generalization was also valid. It just wasn’t quite as valid as the first one. And with the image of the crumbling Twin Towers fresh in every passenger’s mind, we all wanted perfectly valid generalizations. We wanted to be on the planes that we knew had no al Qaeda lunatics on them, not the ones that we could be almost certain were lunatic-free. Our fear, in other words, worked to push us toward prejudice.

So what do we do with these valid, but offensive, generalizations? I don’t pretend to have the answer to racial profiling (other than to recognize that for a great many Americans, “Racial profiling is always Evil, Evil, Evil!!!” is a self-evident belief of great force). There are basic moral principles that have to be invoked, and have to be invoked appropriately, and this is not a book on fundamental ethics.

But I can tell you a few things, just on the basis of how generalizations work.

First, the sooner we devise better and more accurate profiles than “Arab,” the sooner we can stop using race. (For example, a Tunisian passport is rather less worrisome than a Saudi Arabian or Egyptian passport.)

Second, searching an Arab man before he gets on a plane is just moving from a very unreliable generalization about “Arab men” to a vastly more reliable generalization about “people who’ve been carefully searched at the gate.” I can see somebody trying to defend that. But kicking them off the plane and refusing to fly with them, even if they are willing to be searched, just because “they make me nervous” seems altogether too much like insisting on using an obviously unreliable generalization when a much more reliable one is available. I doubt that such an action could possibly be justified.[8]

Third, non-Arabs should be empathizing with all the innocent Arabs who have been associated with evil actions that they themselves abhor, and should be remembering that the vast majority of the Arabs in America detest the actions of al Qaeda and would never engage in terrorism. They should be remembering, in other words, that while it is true that the vast majority of al Qaeda members are Arab, it is equally true that the vast majority of Arabs are not al Qaeda members.

Fourth, Arabs who find themselves guilty of “Flying While Arab” – that is to say, Arabs who find themselves the target of suspicion just because they are Arab and on an airplane – should reflect on the fact that their predicament is not the fault of their fellow passengers, but of the terrorists who have chosen to do great evil in the name of a cause that is associated entirely with the Arab world. They should remember, in other words, that it is true, however inconvenient and regrettable they may personally find it, that the vast majority of al Qaeda members are in fact Arab men.

This whole dynamic underlies the Yiddish phrase shanda fur die goyim. A shanda is a “scandal,” and a shanda fur die goyim is literally a “scandal for the Gentiles.” But culturally, the phrase captures the Yiddish comprehension of the damage done whenever a person who was known to be Jewish behaved in a way that confirmed the negative stereotypes that European Gentiles held of Jews. Even the Jews who were not, say, miserly and usorious, were affected by the behavior of a usorious Jewish miser, for every such Jew – every shanda, in the phrase – rendered the stereotype that much more nearly valid. Accordingly, there was a great deal of social pressure within the Jewish community against being a shanda fur die goyim.

You see, a caricature with no truth behind it is a lie, and it has power only with the very ignorant or the very hateful. A stereotype that can be justified by an occasional shanda has more power, because it has more truth, even though it is not really valid. But a negative generalization that is actually valid (in the sense that it is more often true than not) is devastatingly powerful, for the “scandals” are the rule rather than the exception.

In such a case, if you’re a member of the affected group, you have a big problem, because there will be times when a decision must be made, and when the negative generalization based on race or religion or cultural allegiance is the most reliable generalization available. At that point you can try to convince people not to use that generalization anyway, on the grounds that it represents “racism” or “stereotyping” or “religious bigotry.” But you will have a lot of trouble doing that, for the simple reason that you are essentially trying to tell people that morality requires them to do something stupid; and that can be a very hard sell.

It is thus critically important for a vulnerable race or culture to put a great deal of pressure on its members to avoid being shandas. If you are a Black man and you want to be able to catch a cab whenever you want, you need to realize that the problem will persist until young Black men stop mugging cabbies more often than do young men of other races. If you are an Arab and you don’t want people to worry about you on airplanes, you need to realize that as long as al Qaeda poses a threat, your presence will make people who don’t know you nervous. If you are White and you don’t want Black people to distrust you at first sight, you need to know that as long as a lot of White people treat Black people differently than they treat White people, the problem will persist. (And of course that last remains equally valid if you switch the roles of White and Black in the sentence.)

And if you are a Christian, then here’s one final thing for you to consider. Most people base their opinions on emotion rather than fact and say very stupid things when trying to defend their beliefs. Therefore, since most agnostics are human, most agnostics base their opinions on emotion rather than fact and say very stupid things when trying to defend their beliefs, which is either amusing or frustrating to Christians, depending on temperament. Unfortunately, by the same token, most Christians are human, and unsurprisingly most people who call themselves Christians base their opinions on emotion rather than fact and say very stupid things when trying to defend their beliefs. Therefore the stereotype of Christians as basing their opinions on emotion rather than fact and stupid is actually a valid generalization, and thus has great power. Every time a Christian behaves in accordance with that generalization, he reinforces the validity of the generalization and increases its power – which is to become a shanda.

Of course, the fact that most Christians base their opinions on emotion rather than fact and say stupid things doesn’t at all mean that Christianity itself is a stupid lie. And of course, while it’s reasonable for an agnostic to see that most Christians say silly things, the same agnostic should also realize that most agnostics say equally silly things – but being an agnostic himself, he isn’t likely to recognize that one. Doesn’t matter. The fact remains that the truth is the most powerful lie, and every Christian shanda makes that particular lie more powerful, because he makes it more true.

We can sum up the chapter, then, in the following points.

· Generalizations are not a bad thing.

· You should not, however, rush to judgment, even though our emotions and self-evident beliefs try to get us to leap to conclusions. Rushing to judgment involves acting on prejudice.

· When you do have to make a decision, you should use the most reliable generalization, even though your emotions and self-evident beliefs may be pushing you toward a less reliable but more agreeable generalization.

· In general, the more narrow and specific a generalization is, the more reliable it is.

· Prejudice happens when self-evident beliefs or emotions push you into either rushing into judgment, or else choosing an unreliable generalization when you could have picked a better one.

· The more truth there is in a negative generalization about a particular group, the more trouble that group is in; and therefore it is important for any group to keep itself as free as possible from shandas.

[1] Or what was playground terminology three or four years ago in central Texas. The playground has its own patois that is in a constant state of rapid evolution, and I’ve been out of the loop a while.
[2] If you’re a sixteen-year-old boy, 38 is definitely middle-aged.
[3] I know there are people who consider this type of episode to be an urban myth, but it didn’t happen to “a friend of a friend.” It happened to me.
[4] As a matter of fact, on the door-holding issue, a sensible man also takes into account the age of the female in question. Thus he uses an even more complicated set of narrow generalizations than the one I’ve described.
[5] In language, the greater your vocabulary, the more precisely and accurately you can think and express yourself.
[6] Returning to the example of cab drivers’ discrimination against Black men: a cab driver has to decide right now whether to pick up the fare, and if he makes the wrong choice he can (and numerous cab drivers every year do) die.
[7] There is, of course, a point of diminishing returns. At a certain point you have to decide that you’re going to go ahead and decide just because you don’t think any more information is going to come your way to change your opinion. But I didn’t want to go off on that tangent, and besides, years ago Thomas Sowell did a much better job than I could do in his Knowledge and Decision.
[8] This example is, as Arab-Americans are all too aware, based on actual events.