Monday, February 21, 2005

What Larry Summers didn't say

Dear Kieran Healy,

This is in response to your remarks about the Larry Summers affair (Pharyngula on Larry Summers and Minding the Kids, Again).

Both as a sociologist and as an ordinary concerned & curious citizen, I have been following the uproar over Larry Summers's remarks at the NBER conference with some perplexity. I don't know Summers, and I haven't been at Harvard for a while, so I don't have any personal stake in either defending or attacking him. And since I have always found the "rational actor" model pretty laughable, I am not shocked when such controversies (which belong, in large part, to the boundary-defining ritual processes that Durkheim summed up as "the positive function of crime" and that Kai Erikson later brought out so well) send people over the edge. But in this case, some of the things that have been said by people I know to be very bright and usually sensible have struck me as very puzzling (if not bizarre).

You and I don't know each other, but I generally find the stuff you write perceptive and level-headed. So I must confess I am baffled by the way you characterize the thrust of Summers's discussion.
The main point is the first step toward addressing what Matt properly calls “a set of social expectations that’s become very well entrenched over a very long period” is — contrary to what Summers did in his remarks — to stop treating it as a more-or-less simple result of the expression of individual preferences.
Yes, people do that a lot (including, quite notoriously, economists). But that's simply not what Summers said in his remarks. To pick just one relevant passage:
Now that begs entirely the normative questions-which I'll get to a little later-of, is our society right to expect that level of effort from people who hold the most prominent jobs? Is our society right to have familial arrangements in which women are asked to make that choice and asked more to make that choice than men? Is our society right to ask of anybody to have a prominent job at this level of intensity, and I think those are all questions that I want to come back to. But it seems to me that it is impossible to look at this pattern and look at its pervasiveness and not conclude that something of the sort that I am describing has to be of significant importance.
Maybe I'm just imagining things, but Summers appears to be doing the opposite of reducing "entrenched" social expectations to "a more-or-less simple result of the expression of individual preference." Summers asks here: "Is our society right to have familial arrangements in which women are asked to make that choice and asked more to make that choice than men?" This appears to suggest that the socially structured "familial arrangements" are the key factor. All through this passage, the way Summers begins each point is to ask "is our society right to expect .... Is our society right to have familial arrangements ... Is our society right to ask ... ?"

Now, it's also true that Summers didn't put all his suggestions, speculations, and reflections in precisely this form. At other times, he's asking whether factors other than social arrangements might also play some role, including marginal differences between categories of people due to socialization or (a tricky point) genetics. But that's really the point. Whatever might be said for or against his social graces in other situations, in this particular context Summers was not handing down hard-and-fast conclusions but trying to frame open questions and alternative possibilities in thought-provoking (and sometimes provocative) ways. One can certainly criticize a lot of his points on various grounds--and I agree that some of them provided reasons for people to get upset (though not hysterical)--but (in my humble opinion) there is just no way that the thrust of his remarks fits your characterization. One doesn't have to agree with everything Summers said to notice this.

I offer these comments in a friendly spirit ... but, at the same time, as I say, with a certain sense of bafflement ... and, yes, even some disappointment.

Jeff Weintraub

P.S. Unlike me, Brad DeLong does know Larry Summers, which could give him more insight into the matter or more potential for bias (or both). Be that as it may, his comments on this affair ("Academic Filters...") strike me as admirably calm, sensible, intelligent, and on target.

P.P.S. The arguments raised by Summers which suggest that different categories or groups of people might display significant statistical differences in specific aptitudes or inclinations at the high and low ends, even if the overall mean differences are slight, and that these differences could have noticeable consequences in some social contexts, may be right or wrong (and may or may not fit the cases to which he tried to apply it). But as long as it's not advanced as an Explanation of Everything, this is an intellectually quite respectable form of argument, which can't simply be dismissed a priori by a serious sociologist (or anyone else).

Leaving Larry Summers out of it for the moment, a user-friendly version of this argument, which has its shaky moments but is expressed in an engaging and perceptively thought-provoking way, was put forward in one of Malcolm Gladwell's New Yorker articles from the 1990s:

Malcolm Gladwell, “The Sports Taboo: Why Blacks Are Like Boys and Whites Are Like Girls,"
The New Yorker, pages 50-55 [May 19, 1997]

Whether you agree or disagree with Gladwell's argument in this article (or a little of both), would you be willing to assert that it's totally idiotic and sociologically unacceptable? If so, why?

The New Yorker
May 19, 1997


The Sports Taboo

Why blacks are like boys and whites are like girls.

Malcolm Gladwell


The education of any athlete begins, in part, with an education in the racial taxonomy of his chosen sport-in the subtle, unwritten rules about what whites are supposed to be good at and what blacks are supposed to be good at. In football, whites play quarterback and blacks play running back; in baseball whites pitch and blacks play the outfield. I grew up in Canada, where my brother Geoffrey and I ran high-school track, and in Canada the rule of running was that anything under the quarter-mile belonged to the West Indians. This didn't mean that white people didn't run the sprints. But the expectation was that they would never win, and, sure enough, they rarely did. There was just a handful of West Indian immigrants in Ontario at that point-clustered in and around Toronto-but they owned Canadian sprinting, setting up under the stands at every major championship, cranking up the reggae on their boom boxes, and then humiliating everyone else on the track. My brother and I weren't from Toronto, so we weren't part of that scene. But our West Indian heritage meant that we got to share in the swagger. Geoffrey was a magnificent runner, with powerful legs and a barrel chest, and when he was warming up he used to do that exaggerated, slow-motion jog that the white guys would try to do and never quite pull off. I was a miler, which was a little outside the West Indian range. But, the way I figured it, the rules meant that no one should ever outkick me over the final two hundred metres of any race. And in the golden summer of my fourteenth year, when my running career prematurely peaked, no one ever did.

When I started running, there was a quarter-miler just a few years older than I was by the name of Arnold Stotz. He was a bulldog of a runner, hugely talented, and each year that he moved through the sprinting ranks he invariably broke the existing four-hundred-metre record in his age class. Stotz was white, though, and every time I saw the results of a big track meet I'd keep an eye out for his name, because I was convinced that he could not keep winning. It was as if I saw his whiteness as a degenerative disease, which would eventually claim and cripple him. I never asked him whether he felt the same anxiety, but I can't imagine that he didn't. There was only so long that anyone could defy the rules. One day, at the provincial championships, I looked up at the results board and Stotz was gone.

Talking openly about the racial dimension of sports in this way, of course, is considered unseemly. It's all right to say that blacks dominate sports because they lack opportunities elsewhere. That's the "Hoop Dreams" line, which says whites are allowed to acknowledge black athletic success as long as they feel guilty about it. What you're not supposed to say is what we were saying in my track days-that we were better because we were black, because of something intrinsic to being black. Nobody said anything like that publicly last month when Tiger Woods won the Masters or when, a week later, African men claimed thirteen out of the top twenty places in the Boston Marathon. Nor is it likely to come up this month, when African-Americans will make up eighty per cent of the players on the floor for the N.B.A. playoffs. When the popular television sports commentator Jimmy (the Greek) Snyder did break this taboo, in 1988- infamously ruminating on the size and significance of black thighs-one prominent N.A.A.C.P. official said that his remarks "could set race relations back a hundred years." The assumption is that the whole project of trying to get us to treat each other the same will be undermined if we don't all agree that under the skin we actually are the same.

The point of this, presumably, is to put our discussion of sports on a par with legal notions of racial equality, which would be a fine idea except that civil-rights law governs matters like housing and employment and the sports taboo covers matters like what can be said about someone's jump shot. In his much heralded new book "Darwin's Athletes," the University of Texas scholar John Hoberman tries to argue that these two things are the same, that it's impossible to speak of black physical superiority without implying intellectual inferiority. But it isn't long before the argument starts to get ridiculous. "The spectacle of black athleticism," he writes, inevitably turns into "a highly public image of black retardation." Oh, really? What, exactly, about Tiger Woods's victory in the Masters resembled "a highly public image of black retardation"? Today's black athletes are multimillion- dollar corporate pitchmen, with talk shows and sneaker deals and publicity machines and almost daily media opportunities to share their thoughts with the world, and it's very hard to see how all this contrives to make them look stupid. Hoberman spends a lot of time trying to inflate the significance of sports, arguing that how we talk about events on the baseball diamond or the track has grave consequences for how we talk about race in general. Here he is, for example, on Jackie Robinson:

The sheer volume of sentimental and intellectual energy that has been invested in the mythic saga of Jackie Robinson has discouraged further thinking about what his career did and did not accomplish. . . . Black America has paid a high and largely unacknowledged price for the extraordinary prominence given the black athlete rather than other black men of action (such as military pilots and astronauts), who represent modern aptitudes in ways that athletes cannot.

Please. Black America has paid a high and largely unacknowledged price for a long list of things, and having great athletes is far from the top of the list. Sometimes a baseball player is just a baseball player, and sometimes an observation about racial difference is just an observation about racial difference. Few object when medical scientists talk about the significant epidemiological differences between blacks and whites-the fact that blacks have a higher incidence of hypertension than whites and twice as many black males die of diabetes and prostate cancer as white males, that breast tumors appear to grow faster in black women than in white women, that black girls show signs of puberty sooner than white girls. So why aren't we allowed to say that there might be athletically significant differences between blacks and whites?

According to the medical evidence, African-Americans seem to have, on the average, greater bone mass than do white Americans-a difference that suggests greater muscle mass. Black men have slightly higher circulating levels of testosterone and human-growth hormone than their white counterparts, and blacks over all tend to have proportionally slimmer hips, wider shoulders, and longer legs. In one study, the Swedish physiologist Bengt Saltin compared a group of Kenyan distance runners with a group of Swedish distance runners and found interesting differences in muscle composition: Saltin reported that the Africans appeared to have more blood-carrying capillaries and more mitochondria (the body's cellular power plant) in the fibres of their quadriceps. Another study found that, while black South African distance runners ran at the same speed as white South African runners, they were able to use more oxygen- eighty-nine per cent versus eighty-one per cent-over extended periods: somehow, they were able to exert themselves more. Such evidence suggested that there were physical differences in black athletes which have a bearing on activities like running and jumping, which should hardly come as a surprise to anyone who follows competitive sports.

To use track as an example-since track is probably the purest measure of athletic ability-Africans recorded fifteen out of the twenty fastest times last year in the men's ten-thousand- metre event. In the five thousand metres, eighteen out of the twenty fastest times were recorded by Africans. In the fifteen hundred metres, thirteen out of the twenty fastest times were African, and in the sprints, in the men's hundred metres, you have to go all the way down to the twenty-third place in the world rankings-to Geir Moen, of Norway-before you find a white face. There is a point at which it becomes foolish to deny the fact of black athletic prowess, and even more foolish to banish speculation on the topic. Clearly, something is going on. The question is what.


If we are to decide what to make of the differences between blacks and whites, we first have to decide what to make of the word "difference," which can mean any number of things. A useful case study is to compare the ability of men and women in math. If you give a large, representative sample of male and female students a standardized math test, their mean scores will come out pretty much the same. But if you look at the margins, at the very best and the very worst students, sharp differences emerge. In the math portion of an achievement test conducted by Project Talent-a nationwide survey of fifteen-year-olds-there were 1.3 boys for every girl in the top ten per cent, 1.5 boys for every girl in the top five per cent, and seven boys for every girl in the top one per cent. In the fifty-six-year history of the Putnam Mathematical Competition, which has been described as the Olympics of college math, all but one of the winners have been male. Conversely, if you look at people with the very lowest math ability, you'll find more boys than girls there, too. In other words, although the average math ability of boys and girls is the same, the distribution isn't: there are more males than females at the bottom of the pile, more males than females at the top of the pile, and fewer males than females in the middle. Statisticians refer to this as a difference in variability.

This pattern, as it turns out, is repeated in almost every conceivable area of gender difference. Boys are more variable than girls on the College Board entrance exam and in routine elementary-school spelling tests. Male mortality patterns are more variable than female patterns; that is, many more men die in early and middle age than women, who tend to die in more of a concentrated clump toward the end of life. The problem is that variability differences are regularly confused with average differences. If men had higher average math scores than women, you could say they were better at the subject. But because they are only more variable the word "better" seems inappropriate.

The same holds true for differences between the races. A racist stereotype is the assertion of average difference-it's the claim that the typical white is superior to the typical black. It allows a white man to assume that the black man he passes on the street is stupider than he is. By contrast, if what racists believed was that black intelligence was simply more variable than white intelligence, then it would be impossible for them to construct a stereotype about black intelligence at all. They wouldn't be able to generalize. If they wanted to believe that there were a lot of blacks dumber than whites, they would also have to believe that there were a lot of blacks smarter than they were. This distinction is critical to understanding the relation between race and athletic performance. What are we seeing when we remark black domination of élite sporting events-an average difference between the races or merely a difference in variability?

This question has been explored by geneticists and physical anthropologists, and some of the most notable work has been conducted over the past few years by Kenneth Kidd, at Yale. Kidd and his colleagues have been taking DNA samples from two African Pygmy tribes in Zaire and the Central African Republic and comparing them with DNA samples taken from populations all over the world. What they have been looking for is variants-subtle differences between the DNA of one person and another-and what they have found is fascinating. "I would say, without a doubt, that in almost any single African population-a tribe or however you want to define it-there is more genetic variation than in all the rest of the world put together," Kidd told me. In a sample of fifty Pygmies, for example, you might find nine variants in one stretch of DNA. In a sample of hundreds of people from around the rest of the world, you might find only a total of six variants in that same stretch of DNA-and probably every one of those six variants would also be found in the Pygmies. If everyone in the world was wiped out except Africans, in other words, almost all the human genetic diversity would be preserved.

The likelihood is that these results reflect Africa's status as the homeland of Homo sapiens: since every human population outside Africa is essentially a subset of the original African population, it makes sense that everyone in such a population would be a genetic subset of Africans, too. So you can expect groups of Africans to be more variable in respect to almost anything that has a genetic component. If, for example, your genes control how you react to aspirin, you'd expect to see more Africans than whites for whom one aspirin stops a bad headache, more for whom no amount of aspirin works, more who are allergic to aspirin, and more who need to take, say, four aspirin at a time to get any benefit-but far fewer Africans for whom the standard two-aspirin dose would work well. And to the extent that running is influenced by genetic factors you would expect to see more really fast blacks-and more really slow blacks-than whites but far fewer Africans of merely average speed. Blacks are like boys. Whites are like girls.

There is nothing particularly scary about this fact, and certainly nothing to warrant the kind of gag order on talk of racial differences which is now in place. What it means is that comparing élite athletes of different races tells you very little about the races themselves. A few years ago, for example, a prominent scientist argued for black athletic supremacy by pointing out that there had never been a white Michael Jordan. True. But, as the Yale anthropologist Jonathan Marks has noted, until recently there was no black Michael Jordan, either. Michael Jordan, like Tiger Woods or Wayne Gretzky or Cal Ripken, is one of the best players in his sport not because he's like the other members of his own ethnic group but precisely because he's not like them-or like anyone else, for that matter. Élite athletes are élite athletes because, in some sense, they are on the fringes of genetic variability. As it happens, African populations seem to create more of these genetic outliers than white populations do, and this is what underpins the claim that blacks are better athletes than whites. But that's all the claim amounts to. It doesn't say anything at all about the rest of us, of all races, muddling around in the genetic middle.


There is a second consideration to keep in mind when we compare blacks and whites. Take the men's hundred-metre final at the Atlanta Olympics. Every runner in that race was of either Western African or Southern African descent, as you would expect if Africans had some genetic affinity for sprinting. But suppose we forget about skin color and look just at country of origin. The eight-man final was made up of two African-Americans, two Africans (one from Namibia and one from Nigeria), a Trinidadian, a Canadian of Jamaican descent, an Englishman of Jamaican descent, and a Jamaican. The race was won by the Jamaican-Canadian, in world-record time, with the Namibian coming in second and the Trinidadian third. The sprint relay-the 4 x 100-was won by a team from Canada, consisting of the Jamaican-Canadian from the final, a Haitian-Canadian, a Trinidadian-Canadian, and another Jamaican-Canadian. Now it appears that African heritage is important as an initial determinant of sprinting ability, but also that the most important advantage of all is some kind of cultural or environmental factor associated with the Caribbean.

Or consider, in a completely different realm, the problem of hypertension. Black Americans have a higher incidence of hypertension than white Americans, even after you control for every conceivable variable, including income, diet, and weight, so it's tempting to conclude that there is something about being of African descent that makes blacks prone to hypertension. But it turns out that although some Caribbean countries have a problem with hypertension, others-Jamaica, St. Kitts, and the Bahamas-don't. It also turns out that people in Liberia and Nigeria-two countries where many New World slaves came from-have similar and perhaps even lower blood-pressure rates than white North Americans, while studies of Zulus, Indians, and whites in Durban, South Africa, showed that urban white males had the highest hypertension rates and urban white females had the lowest. So it's likely that the disease has nothing at all to do with Africanness.

The same is true for the distinctive muscle characteristic observed when Kenyans were compared with Swedes. Saltin, the Swedish physiologist, subsequently found many of the same characteristics in Nordic skiers who train at high altitudes and Nordic runners who train in very hilly regions-conditions, in other words, that resemble the mountainous regions of Kenya's Rift Valley, where so many of the country's distance runners come from. The key factor seems to be Kenya, not genes.

Lots of things that seem to be genetic in origin, then, actually aren't. Similarly, lots of things that we wouldn't normally think might affect athletic ability actually do. Once again, the social-science literature on male and female math achievement is instructive. Psychologists argue that when it comes to subjects like math, boys tend to engage in what's known as ability attribution. A boy who is doing well will attribute his success to the fact that he's good at math, and if he's doing badly he'll blame his teacher or his own lack of motivation-anything but his ability. That makes it easy for him to bounce back from failure or disappointment, and gives him a lot of confidence in the face of a tough new challenge. After all, if you think you do well in math because you're good at math, what's stopping you from being good at, say, algebra, or advanced calculus? On the other hand, if you ask a girl why she is doing well in math she will say, more often than not, that she succeeds because she works hard. If she's doing poorly, she'll say she isn't smart enough. This, as should be obvious, is a self-defeating attitude. Psychologists call it "learned helplessness"-the state in which failure is perceived as insurmountable. Girls who engage in effort attribution learn helplessness because in the face of a more difficult task like algebra or advanced calculus they can conceive of no solution. They're convinced that they can't work harder, because they think they're working as hard as they can, and that they can't rely on their intelligence, because they never thought they were that smart to begin with. In fact, one of the fascinating findings of attribution research is that the smarter girls are, the more likely they are to fall into this trap. High achievers are sometimes the most helpless. Here, surely, is part of the explanation for greater math variability among males. The female math whizzes, the ones who should be competing in the top one and two per cent with their male counterparts, are the ones most often paralyzed by a lack of confidence in their own aptitude. They think they belong only in the intellectual middle.

The striking thing about these descriptions of male and female stereotyping in math, though, is how similar they are to black and white stereotyping in athletics-to the unwritten rules holding that blacks achieve through natural ability and whites through effort. Here's how Sports Illustrated described, in a recent article, the white basketball player Steve Kerr, who plays alongside Michael Jordan for the Chicago Bulls. According to the magazine, Kerr is a "hard-working overachiever," distinguished by his "work ethic and heady play" and by a shooting style "born of a million practice shots." Bear in mind that Kerr is one of the best shooters in basketball today, and a key player on what is arguably one of the finest basketball teams in history. Bear in mind, too, that there is no evidence that Kerr works any harder than his teammates, least of all Jordan himself, whose work habits are legendary. But you'd never guess that from the article. It concludes, "All over America, whenever quicker, stronger gym rats see Kerr in action, they must wonder, How can that guy be out there instead of me?"

There are real consequences to this stereotyping. As the psychologists Carol Dweck and Barbara Licht write of high- achieving schoolgirls, "[They] may view themselves as so motivated and well disciplined that they cannot entertain the possibility that they did poorly on an academic task because of insufficient effort. Since blaming the teacher would also be out of character, blaming their abilities when they confront difficulty may seem like the most reasonable option." If you substitute the words "white athletes" for "girls" and "coach" for "teacher," I think you have part of the reason that so many white athletes are underrepresented at the highest levels of professional sports. Whites have been saddled with the athletic equivalent of learned helplessness-the idea that it is all but fruitless to try and compete at the highest levels, because they have only effort on their side. The causes of athletic and gender discrimination may be diverse, but its effects are not. Once again, blacks are like boys, and whites are like girls.


When I was in college, I once met an old acquaintance from my high-school running days. Both of us had long since quit track, and we talked about a recurrent fantasy we found we'd both had for getting back into shape. It was that we would go away somewhere remote for a year and do nothing but train, so that when the year was up we might finally know how good we were. Neither of us had any intention of doing this, though, which is why it was a fantasy. In adolescence, athletic excess has a certain appeal-during high school, I happily spent Sunday afternoons running up and down snow-covered sandhills-but with most of us that obsessiveness soon begins to fade. Athletic success depends on having the right genes and on a self-reinforcing belief in one's own ability. But it also depends on a rare form of tunnel vision. To be a great athlete, you have to care, and what was obvious to us both was that neither of us cared anymore. This is the last piece of the puzzle about what we mean when we say one group is better at something than another: sometimes different groups care about different things. Of the seven hundred men who play major-league baseball, for example, eighty-six come from either the Dominican Republic or Puerto Rico, even though those two islands have a combined population of only eleven million. But then baseball is something that Dominicans and Puerto Ricans care about-and you can say the same thing about African-Americans and basketball, West Indians and sprinting, Canadians and hockey, and Russians and chess. Desire is the great intangible in performance, and unlike genes or psychological affect we can't measure it and trace its implications. This is the problem, in the end, with the question of whether blacks are better at sports than whites. It's not that it's offensive, or that it leads to discrimination. It's that, in some sense, it's not a terribly interesting question; "better" promises a tidier explanation than can ever be provided.

I quit competitive running when I was sixteen-just after the summer I had qualified for the Ontario track team in my age class. Late that August, we had travelled to St. John's, Newfoundland, for the Canadian championships. In those days, I was whippet-thin, as milers often are, five feet six and not much more than a hundred pounds, and I could skim along the ground so lightly that I barely needed to catch my breath. I had two white friends on that team, both distance runners, too, and both, improbably, even smaller and lighter than I was. Every morning, the three of us would run through the streets of St. John's, charging up the hills and flying down the other side. One of these friends went on to have a distinguished college running career, the other became a world-class miler; that summer, I myself was the Canadian record holder in the fifteen hundred metres for my age class. We were almost terrifyingly competitive, without a shred of doubt in our ability, and as we raced along we never stopped talking and joking, just to prove how absurdly easy we found running to be. I thought of us all as equals. Then, on the last day of our stay in St. John's, we ran to the bottom of Signal Hill, which is the town's principal geographical landmark-an abrupt outcrop as steep as anything in San Francisco. We stopped at the base, and the two of them turned to me and announced that we were all going to run straight up Signal Hill backward. I don't know whether I had more running ability than those two or whether my Africanness gave me any genetic advantage over their whiteness. What I do know is that such questions were irrelevant, because, as I realized, they were willing to go to far greater lengths to develop their talent. They ran up the hill backward. I ran home.

Copyright 1997, Malcolm Gladwell