September 2005

The Inequality TabooCharles Murray

WHEN THE late Richard Herrnstein and Ipublished The Bell Curve eleven years ago,

the furor over its discussion of ethnic differences inIQ was so intense that most people who have notread the book still think it was about race. Sincethen, I have deliberately not pubhshed anythingabout group differences in IQ, mostly to give thereal topic of The Bell Curve—the role of intelli-gence in reshaping America's class structure—achance to surface.

The Lawrence Summers affair last January mademe rethink my silence. The president of HarvardUniversity offered a few mild, speculative, off-the-record remarks about innate differences betweenmen and women in their aptitude for high-level sci-ence and mathematics, and was treated by Harvard'sfaculty as if he were a crank. The typical news storyportrayed the idea of innate sex differences as arenegade position that reputable scholars rejected.

It was depressingly familiar. In the autumn of1994,1 had watched with dismay as The Bell Curve'sscientifically unremarkable statements about blackIQ were successfully labeled as racist pseudo-science. At the opening of 2005,1 watched as somescientifically unremarkable statements about male-

CHARLES MURRAY is the W.H. Brady Scholar for Free-dom and Culture at the American Enterprise Institute.His previous contributions to COMMENTARY, including"The Bell Curve and Its Critics" (May 1995, with asubsequent exchange in the August 1995 issue), can be foundat www.com/mentarymagazine.com..

female differences were successfully labeled as sex-ist pseudoscience.

The Qrwellian disinformation about innategroup differences is not wholly the media's fault.Many academics who are familiar with the state ofknowledge are afraid to go on the record. Talkingpublicly can dry up research funding for seniorprofessors and can cost assistant professors theirjobs. But while the public's misconception is un-derstandable, it is also getting in the way of clearthinking about American social policy.

Good social policy can be based on premises thathave nothing to do with scientific truth. Thepremise that is supposed to undergird all of our so-cial poHcy, the founders' assertion of an unalienableright to liberty, is not a falsifiable hypothesis. Butspecific policies based on premises that conflici:with scientific truths about human beings tend notto work. Often they do harm.

One such premise is that the distribution of in-nate abilities and propensities is the same acrossdifferent groups. The statistical tests for uncover-ing job discrimination assume that men are not in-nately different from women, blacks from whites,older people from younger people, homosexualsfrom heterosexuals. Latinos from Anglos, in waysthat can legitimately affect employment decisions.Title IX ofthe Educational Aniendments of 1972assumes that women are no different from men intheir attraction to sports. Affirmative action in allits forms assumes there are no innate differences

[13]

COMMENTARY SEPTEMBER 2005

between any of the groups it seeks to help andeveryone else. The assumption of no innate differ-ences among groups suffuses American social poli-cy. That assumption is wrong.

When the outcomes that these policies are sup-posed to produce fail to occur, with one groupfalling short, the fault for the discrepancy has beenassigned to society. It continues to be assumed thatbetter programs, better regulations, or the rightcourt decisions can make the differences go away.That assumption is also wrong.

Hence this essay. Most of the following discus-sion describes reasons for believing that sorhegroup differences are intractable. I shift frdrii "in-nate" to "intractable" to acknowledge how complexis the interaction of genes, their expression in be-havior, and the environment. "Intractable" meansthat, whatever the precise partitioning of causationmay be (we seldom know), policy interventions canonly tweak the difference at the margins.

I will focus on two sorts of differences: betweenrnen and women and between blacks and whites.Here are three crucial points to keep iii mind as wego along:

1. The differences I discuss involve means anddistributions. In all cases, the variation withingroups is greater than the variation betweengroups. On psychological and cognitive dimen-sions, some members of both sexes and all races falleverywhere along the range. One implication ofthis is that genius does not coni^ in one color orsex, and neither does any other huirian ability. An-other is that a few minutes of conversation with in-dividuals you meet will tell you much more aboutthem than their group membership does.

2. Covering both sex differences and race differ-ences in a single, non-technical article, I have had toleave out much. I urge that readers with questionsconsult the fully annotated version of this essay,which includes extensive supplementary material; itis available at www.commentarymagazine.com andat wvvTv.aei.org.

3. The concepts of "inferiority" and "superiori-ty" are inappropriate to group comparisons. Onmost specific human attributes, it is possible tospecify a continuum running from "low" to"high," but the results cannot be combined into ascore running from "bad" to "good." What is thebest score on a continuum measuring aggressive-ness? WTiat is the relative importance of verbalskills versus, say, compassion? Of spatial skills ver-sus industriousness? Tbe aggregate excellencesand shortcomings of human groups do not lendthemselves to simple comparisons. That is why the

members of just about every group can so easilyconclude that they are God's chosen people. All ofus use the weighting system that favors our group'sstrengths.

II

THE TECHNICAL literature documenting sexdifferences and their biological basis grew sur-

reptitiously during feminism's heyday in the 197O'sand 198O's. By tbe 199O's, it had become so exten-sive that the bibliography in David Geary's pio-neering M^/e, Female (1998) ran to 53 pages. Cur-rently, the best short account ofthe state of knowl-edge is Steven Pinker's chapter on gender in TheBlank Slate (2002).*

Rather than present a telegraphic list of all the dif-ferences that I think have been established, I willfocus on the narrower question at the heart of theSummers controversy: i& groups, do men andwomen differ innately in chdracteristics that pro-duce achievement at the highest levels of accom-plishment? I will limit my comments to the artsand sciences.

Since we live in an age when students are likelyto hear more about Marie Curie than about AlbertEinstein, it is worth begirming with a statement ofhistorical fact: women have played a proportional-ly tiriy part in the history of the arts and sciences.^Even in the 20th century, women got only 2 per-cent of tbe Nobel Prizes in the sciences—a pro-portion constant for both halves of the century—and 10 percent of the prizes in literature. TheFields Medal, the most prestigious award in math-ematics, has been given to 44 people since it origi-nated in 1936. All have been men.

The historical reality of male dominance of thegreatest achievements in science and the arts is notopen to argument. The question is whether the so-cial and legal exclusion of women is a sufficient ex-

* A non-technical book-length treatment is Taking Sex DijferencesSeriotcsly (2004) by Steven Rhoads. Diane Halpem's Sex Differencesin Cognitive Ability (2000) is a good one-volume discussion of cog-nitive differences between the sexes. An up-to-date summary ofneuro-physiological findings about sex differences in the brain ap-peared in last May's Scientific American ("His Brain, Her Brain," byLarry Cahill). Simon Baron-Cohen's The Essential Difference (2003)is an ambitious attempt to tie together known sex differences intoan overall theory. Those who want to compare these accounts withdefenses of the no-irmate-differences position can look at VirginiaValian's IVhy So Stow? (1999) and a set of essays weighted towardsocial explanations in Gender Differences in Mathematics (2005), edit-ed by Ann Gallagher and James Kaufman.

"•" In my Human Accomplishment: The Pursuit of Excellence in the Artsand Sciences, 800 B.C. to 1950 (2003), I estimated that women ac-count for 2 percent of the significant contributors. Others havefound similar proportions.

[14]

THE INEQUALITY TABOO

planadon for this situation, or whether sex-specificcharacteristics are also at work.

Mathematics offers an entry point for thinkingabout the answer. Through high school, girls earnbetter grades in math than boys, but the boys usual-ly do better on standardized tests. The difference inmeans is modest, but the male advantage increasesas the focus shifts from means to extremes. In alarge sample of mathematically gifted youths, forexample, seven times as many males as femalesscored in the top percendle of the SAT mathemat-ics test. We do not have good test data on the male-female ratio at the top one-himdredth or top one-thousandth of a percendle, where first-rate mathe-madcians are most likely to be foimd, but collateralevidence suggests that the male advantage therecondnues to increase, perhaps exponendally.

Evoludonary biologists have some theories thatfeed into an explanadon for the disparity. In primi-dve societies, men did the hunting, which oftentook them far from home. Males with the ability torecognize landscapes from different orientationsand thereby find their way back had a survival ad-vantage. Men who could process trajectories inthree dimensions—the trajectory, say, of a spearthrown at an edible mammal—also had a survivaladvantage. Women did the gathering. Those whocould disdnguish among complex arrays of vegeta-don, remembering which were the poisonous plantsand which the nourishing ones, also had a survivaladvantage. Thus the logic for explaining why menshould have developed elevated three-dimensionalvisuospadal skills and women an elevated abihty toremember objects and their reladve locadons—dif-ferences that show up in specialized tests today.

Perhaps this is a just-so story. Why not insteadattribute the results of these tests to sociahzadon?Enter the neuroscientists. It has been known foryears that, even after adjusdng for body size, menhave larger brains than women. Yet most psycho-metricians conclude that men and women have thesame mean IQ (although debate on this issue isgrovwng). One hypothesis for explaining this para-dox is that three-dimensional processing absorbsthe extra male capacity. In the last few years, mag-nedc-resonance imaging has refined the evidencefor this hypothesis, revealing that parts ofthe brain'sparietal cortex associated with space percepdon arepropordonally bigger in men than in women.

What does space perception have to do withscores on math tests? Enter the psychometricians,who demonstrate that when visuospadal ability istaken into account, the sex difference in SAT mathscores shrinks substandally.

Why should the difference be so much greaterat the extremes than at the mean? Part of the an-swer is that men consistently exhibit higher vari-ance than women on all sorts of characterisdcs, in-cluding visuospadal abilities, meaning that thereare propordonally more men than women at bothends of the bell curve. Another part of the answeris that someone with a high verbal IQ can easilymaster the basic algebra, geometry, and calculusthat make up most of the items in an ordinary mathtest. Elevated visuospadal skills are most useful forthe most difficult items. If males have an advantagein answering those comparadvely few really harditems, the increasing disparity at the extremes be-comes explicable.

Seen from one perspecdve, this pattern demon-strates what should be obvious: there is nothing in-herent in being a woman that precludes high mathability. But there remains a distribudonal differencein male and female characterisdcs that leads to alarger number of men with high visuospadal skills.The difference has an evolutionary rationale, aphysiological basis, and a direct correlation withmath scores.

Now PUT all this alongside the historical dataon accomplishment in the arts and sciences.

In test scores, the male advantage is most pro-nounced in the most abstract items. Historically,too, it is most pronounced in the most abstract do-mains of accomplishment.

In the humariides, the most abstract field is phi-losophy—and no woman has been a significantoriginal thinker in any of the world's great philo-sophical tradidons. In die sciences, the most abstractfield is mathemadcs, where the number of greatwomen mathematicians is approximately two(Emmy Noether definitely, Sonya Kovalevskayamaybe). In the other hard sciences, the contribu-dons of great women sciendsts have usually beenempirical rather than theoredcal, vdth leading casesin point being Henrietta Leavitt, Dorothy Hodg-kin, Lise Meitner, Irene Joliot-Curie, and MarieCurie herself.

In the arts, literature is the least abstract and byfar the most rooted in human interacdon; visual artincorporates a greater admixture of the abstract;musical composidon is the most abstract of all thearts, using neither words nor images. The role ofwomen has varied accordingly. Women have beenrepresented among great writers virtually from thebeginning of literature, in East Asia and South Asiaas well as in the West. Women have produced asmaller number of important visual artists, and

[15]

COMMENTARY SEPTEMBER 2005

none that is clearly in the first rank. No femalecomposer is even close to the first rank. Social re-strictions undoubtedly damped down women's con-tributions in all of the arts, but the pattern of ac-complishment that did break through is strikihglyconsistent with what we know about the respectivestrengths of male and female cbgriitive repertoires.

Women have their own cognitive advantagesover men, many of them involving verbal fluencyand interpersonal skills. If this were a comprehen-sive survey, detailing those advantages would takeup as much space as I have devoted to a particularmale advantage. But, sticking with my restrictedtopic, I will move to another aspect of male-femaledifferences that bears on accomplishment at thehighest levels ofthe arts and sciences: motherhood.

Regarding women, men, and babies, the techni-cal literature is as unambiguous as everyday

experience would lead one to suppose. As a rule,the experience of parenthood is more profoundlylife-altering for women than for men. Nor is thereanything unique about humans in this regard.Mammalian reproduction generally involves muchhigher levels of maternal than paternal investmentin the raising of children. Among humans, exten-sive empirical study has demonstrated that womenare more attracted to children than are meri, re-spond to them more intensely on an emotionallevel, and get more and different kinds of satisfac-tions from nurturing them. Many of these behav-ioral differences have been hnked with biochemi-cal differences between men and women.

Thus, for reasons embedded in the biochemistryand neurophysiology of being female, manywomen with the cognitive skills for achievement atthe highest level also have something else theywant to do in life: have a baby. In the arts and sci-ences, forty is the mean age at which peak accom-phshment occurs, preceded by years of intense ef-fort mastering the discipline in question. These areprecisely the years during which most women mustbear children if they are to bear them at all.

Among women who have become mothers, thepossibihties for high-level accomplishment in thearts and sciences shrink because, for innate reasons,the distractions of parenthood are greater. To putit in a way that most readers with children will rec-ognize, a father can go to work and forget about hischildren for the whole day. Hardly any mother cando this, no matter how good her day-care arrange-ment or full-time nanny may be. Aly point is notthat women must choose between a career andchildren, but that accomplishment at the extremes

commonly comes from a single-minded focus thatleaves no room for anything but the task at hand.We should not be surprised or dismayed to findthat motherhood reduces the proportion of highlytalented young women who are willing to makethat trade-off.

Some numbers can be put to this observationthrough a study of nearly 2,000 men and womenwho were identified as extraordinarily talented inmath at age thirteen and were followed up 20 yearslater. The women in the sample came of age in the197O's and early 198O's, when women were activelysocialized to resist gender stereotypes. In manyways, these talented women did resist. By theirearly thirties, both the men and women had be-come exceptional achievers, receiving advanced de-grees in roughly equal proportions. Only about 15percent of the women \yere full-time housewives.Among the women, those who did and those whodid not have children were equally satisfied withtheir careers.

And yet. The women with careers were four-and-a-half times more likely than men to say they pre-ferred to work fewer than 40 hours per week. Themen placed greater importance on "being successfulin my line of work" and "inventing or creating some-thing that will have an impact," while the womenfdiind greater value in "having strong friendships,""living clbke to parents and relatives," and "having ameaningful spiritual life." As the authors concluded,"these men and women appear to have constructedsatisfying and meaningful lives that took somewhatdifferent forms." The different forms, which directlyinfluence the hkelihood that men will dominate atthe extreme levels of achievement, are consistentwith a constellation of differences between men andwomen that have biological roots.

I have omitted perhaps the most obvious reasonwhy men and women differ at the highest levels ofaccomplishment: men take more risks, are morecompetitive, and are more aggressive than women.The word "testosterone" may come to mind, andappropriately. Much technical literature documentsthe hormonal basis of personality differences thatbear on sex differences in extreme and venturesomeeffort, and hence in extremes of accomplishment—and that bear as well on the male propensity to pro-duce an overwhelming proportion of the world'scrime and approximately 100 percent of its wars.

But this is just one more of the ways in whichscience is demonstrating that men and women ftrereally and truly different, a fact so obvious that onlyintellectuals could ever have thought otherwise.

[16]

THE INEQUALITY TABOO

III

TURNING TO race, we must begin with thefraught question of whetlier it even exists, or

whether it is instead a social construct. The Harvardgeneticist Richard Lewontin originated the idea ofrace as a social construct in 1972, arguing that thegenetic differences across races were so trivial thatno scientist working exclusively with genetic datawould sort people into hlacks, whites, or Asians. Inhis words, "racial classification is now seen to be ofvirtually no genetic or taxonomic significance."

Lewontin's position, which quickly became a tenetof political correctness, carried with it a potentialmeans of being falsified. If he was correct, then a sta-tistical analysis of genetic markers would not pro-duce clusters corresponding to common racial labels.

In the last few years, that test has become feasi-ble, and now we know that Lewontin was wrong.Several analyses have confirmed the genetic realityof gtoup identities going under the label of race orethnicity. In the most recent, published this year,all but five ofthe 3,636 subjects fell into the clusterof genetic markers corresponding to their self-identified ethnic group. When a statistical proce-dure, blind to physical characteristics and workingexclusively with genedc information, classifies 99.9percent of the individuals in a large sample in thesame way they classify themselves, it is hard toargue that race is imaginary.

Homo sapiens actually falls into many more inter-esting groups than the bulky ones known as"races." As new findings appear althost weekly, itseems increasingly likely that we are just at the be-ginning of a process that will identify all sorts ofgenetic differences among groups, whether thegroups being compared are Nigerian blacks andKenyan blacks, lawyers and engineers, or Episco-palians and Baptists. At the moment, the differ-ences that are obviously genetic involve diseases(Ashkenazi Jews and Tay-Sachs disease, blackAfricans and sickle-cell anemia, Swedes and he-mochromatosis). As time gdes on, we may yetcome to understand better why, say, Italians aremore vivacious than Scots.

Out of all the interesting and intractable differ-ences that may eventually be identified, one in par-ticular remains a hot button like no other: the IQdifference between blacks and whites. What is thepresent state of our knowledge about it?

There is no technical dispute on some of thecore issues. In the aftermath of The Bell Curve, theAmerican Psychological Association established atask force on intelligence whose report was pub-

lished in early 1996. The task force reached thesame conclusions as The Bell Curve on the size andmeaningfulness ofthe black-white difference. His-torically, it has been about one standard deviationin magnitude among subjects who have reachedadolescence*; cultural bias in IQ tests does not ex-plain the difference; and the tests are about equallypredictive of educational, social, and economic out-comes for blacks and whites. However cbnttover-sial such assertions may still be in the eyes of themainstream media, they ate not controversial with-in the scientific community.

The most important change in the state of knowl-edge since the mid-1990's lies in our increased un-derstanding of what has happened to the size of theblack-white difference over time. Both the task forceand The Bell Curve concluded that some narrowinghad occurred since the early 197O's. With the advan-tage of an additional decade of data, we are now ableto be more precise: (1) The black-white difference inscores on educational achievement tests has nar-rowed significandy. (2) The black-white convergencein scores on the most highly "g-loaded" tests—thetests that are the best measures of cognitive ability—has been smaller, and may be unchanged, since thefirst tests were administered 90 years ago.

WITH REGARD to the difference in educationalachievement, the naftowirig of scores on

major tests occurred in the 1^70's and 8O's. Inthe case of the SAT, the gapS in the verbal andmath tests as of 1972 wfere 1.24 and 1.26 standarddeviations respectivdy. By 1991, when the gapswere smallest (they halve risen slightly since then),those numbers had dropped by .37 and .35 stan-dard deviations.

The National Assessment of EducationalProgress (NAEP), which is not limited tb college-bound students, is preferable to the SA t for esti-mating nationally representative trerids, but thestory it tells is similar. Among students ages nine,thirteen, and seventeen, the black-white differencesin math afe.of the first NAEP test in 1973 were L03,1.29, and 1.24 standard deviations respectively; Eornine-yeat--olds, the difference hit its all-dme low of.73 standard deviations in 2004, a drop of .30 stan-ddtd deviations. But almost all of that convergencehad been reached by 1986, when the gap was .78

* The standard deviation is a statistic that (slighUy simphfied) ex-presses the average difference of all the scores from the mean.Given a normal distrihution—a bell curve—someone who is onestandard deviation ahove the mean is at the 84th percentile. Twostandard deviations ahove the mean puts that person at the 98thpercendle. IQ tests are normed to have a mean of 100 and a stan-dard deviation of 15.

[17]

COMMENTARY SEPTEMBER 2005

standard deviations. Eor thirteen-year-olds, the gapdropped by .45 standard deviations, reaching its lowin 1986. Eor seventeen-year-olds, the gap droppedby .52 standard deviations, reaching its low in 1990.

In the reading test, the comparable gaps for agesnine, thirt:een, and seventeen as of the first NAEPtest in 1971 were 1.12, 1.17, and 1.25 standard de-viations. Those gaps had shrunk by .38, .62, and.68 standard deviations respectively at their lowestpoints in 1988. They have since remained effec-tively unchanged.

An analysis by Larry Hedges and Amy Nowelluses a third set of data, examining the trends forhigh-school seniors by comparing six large databases from different time periods from 1965 to1992. The black-white difference on a combinedmeasure of math, vocabulary, and reading fell from1.18 to .82 standard deviations in that time, a re-duction of .36 standard deviations.

So black and white academic achievement con-verged significantly in the 197O's and 198O's, typi-cally by more than a third of a standard deviation,and since then has stayed about the same. WTiatabout convergence in tests explicitly designed tomeasure IQ rather than academic achievement?The ambiguities in the data leave two defensiblepositions. The first is that the IQ difference isabout one standard deviation, effectively un-changed since the first black-white comparisons 90years ago. The second is that harbingers of a nar-rowing difference are starting to emerge. I cannotsettle the argument here, but I can convey somesense of the uncertainty.

THE CASE for an vmchanged black-white IQ dif-ference is straightforward. If you take all the

black-white differences on IQ tests from the firstones in World War I up to the present, there is nostatistically significant downward trend. Qf coursethe results vary, because tests vary in the precisionwith which they measure the general mental factor(g) and samples vary in their size and representa-tiveness. But results continue to center around ablack-white difference of about 1.0 to 1.1 standarddeviations through the most recent data.

The case for a reduction has two important re-cent results to work with. The first is from the1997 re-norming ofthe Armed Eorces Qualifica-tion Test (AEQT), which showed a black-white dif-ference of .97 standard deviations. Since the typi-cal difference on paper-and-pencil IQ tests like theAEQT has been about 1.10 standard deviations,the 1997 results represent noticeable improvement.

The second positive result comes from the 2003

standardization sample for the Wechsler Intelli-gence Scale for Children (WISC-IV), which showeda difference of .78 standard deviations, as against the1.0 difference that has been typical for individuallyadministiered IQ tests.' Qne cannot draw strong conclusions from twodata points. Those who interpret them as part of anunchanging overall pattern can cite another recentresult, from the 2001 standardization ofthe Wood-cock-Johnson intelligence test. In line with theconventional gap, it showed an overall black-whitedifference of 1.05 standard deviations and, foryouths aged six to eighteen, a difference of .99standard deviations.

There is more to be said on both sides of thisissue, but nothing conclusive. Until new data be-come available, you may take your choice. If youare a pessimist, the gap has been unchanged atabout one standard deviation. If you are an opti-mist, the IQ gap has decreased by a few points, butit is still close to one standard deviation. The clearand substantial convergence that occurred in acad-emic tests has at best been but dimly reflected inIQ scores, and at worst not reflected at all.

WHETHER WE are talking about academicachievement or about IQ, are the causes of

the black-white difference environmental or genet-ic? Everyone agrees that environment plays a part.The controversy is about whether biology is alsoinvolved.

It has been known for many years that the obviousenvironmental factors such as income, parental occu-pation, and schools explain only part of the absoluteblack-white difference and none ofthe relative differ-ence. Black and white students from affluent neigh-borhoods are separated by as large a proportional gapas are blacks and whites from poor neighborhoods.Thus the most interesting recent studies of environ-mental causes have worked with cultural explanationsinstead of socioeconomic status.*

Qne example is Black American Students in an Af-fluent Suburb: A Study of Academic Disengagement(2003) by the Berkeley anthropologist John Ogbu,who went to Shaker Heights, Ohio, to explore why

* I put aside here the explanation that has received the most pub-licity in recent years, the phenomenon labeled "stereotype threat."Its discoverers, Claude Steele and Joshua Aronson, demonstratedexperimentally that test performance hy academically talentedblacks was worse when a test was called an IQ test than when it wasinnocuously described as a research tool. Press reports erroneous-ly interpreted this as meaning that stereotype threat explained awaytile black-white difference. In reality, Steele and Aronson showedonly that it increases the usual black-white difference; if one elim-inates stereotype threat, the usual difference remains.

[18]

THE INEQUALITY TABOO

black students in an affluent suburb should lag be-hind their white peers. Another is Black Rednecks andWhite Liberals (2005) by Thomas Sowell, who makesthe case that what we think of as the dysfunctionalaspects of urban black culture are a legacy not ofslavery but of Southern and rural white "cracker"culture. Both Qgbu and Sowell describe ingrainedparental behaviors and student attitudes that mustimpede black academic performance. These cultur-al influences ofren cut across social classes.

Erom a theoretical standpoint, the cultural ex-planations offer fresh ways of looking at the black-white difference at a time when the standard so-cioeconomic explanations have reached a dead end.Erom a practical standpoint, however, the culturalexplanations point to a cause ofthe black-white dif-ference that is as impervious to manipulation by so-cial policy as causes rooted in biology. If there is tobe a rapid improvement, some form of mass move-ment with powerful behavioral consequences wouldhave to occur within the black community. Absentthat, the best we can hope for is gradual culturalchange that is likely to be measured in decades.

This brings us to the state of knowledge aboutgenetic explanations. "There is not much direct ev-idence on this point," said the American Psycho-logical Association's task force dismissively, "butwhat littie there is fails to support the genetic hy-pothesis." Actually, there is no direct evidence atall, just a wide variety of indirect evidence, almostall of which the task force chose to ignore.

As it happens, a comprehensive survey of thatevidence, and ofthe objections to it, appeared thispast June in the journal Psychology, Public Policy, andLaw. There, J. Philippe Rushton and Arthur Jensenco-authored a 60-page article entitled "ThirtyYears of Research on Race Differences in Cogni-tive Ability." It incorporates studies of East Asiansas well as blacks and whites and concludes that thesource ofthe black-white-Asian difference is 50- to80-percent genetic. The same issue of the journalincludes four commentaries, three of them writtenby prominent scholars who oppose the idea thatany part of the black-white diff'erence is genetic.Thus, in one place, you can examine the strongestarguments that each side in the debate can bring tobear.

Rushton and Jensen base their conclusion on tencategories of evidence that are consistent with amodel in which both environment and genes causethe black-white difference and inconsistent with amodel that requires no genetic contribution. I willnot try to review their argument here, or the cri-tiques of it. All of the contributions can be found

on the Internet, and can be understood by readerswith a grasp of basic statistical concepts.*

Eor those who consider it important to knowwhat percentage of the IQ difference is genetic, amethodology that would do the job is now avail-able. In the United States, few people classified asblack are actually of 100-percent African descent(the average American black is thought to be about20-percent white). To the extent that genes play arole, IQ will vary by racial admixture. In the past,studies that have attempted to test this hypothesishave had no accurate way to measure the degree ofadmixture, and the results have been accordinglymuddy. The recent advances in using geneticmarkers solve that problem. Take a large sample ofracially diverse people, give them a good IQ test,and then use genetic markers to create a variablethat no longer classifies people as "white" or"black," but along a continuum. Analyze the varia-tion in IQ scores according to that continuum.The results would be close to dispositive.

NONE OF this is important for social policy,however, where the issue is not the source of

the difference but its intractability. Much ofthe ev-idence reviewed by Rushton and Jensen bears onwhat we can expect about future changes in theblack-white IQ difference. My own thinking onthis issue is shaped by the relationship of the dif-ference to a factor I have already mentioned—"g"—and to the developing evidence for g's bio-logical basis.

When you compare black and white mean scoreson a battery of subtests, you do not find a uniformset of differences; nor do you find a random as-sortment. The size ofthe difference varies system-atically by type of subtest. Asked to predict whichsubtests show the largest difference, most peoplewill think first of ones that have the most culturalcontent and are the most sensitive to good school-ing. But this natural expectation is wrong. Some ofthe largest differences are found on subtests thathave little or no cultural content, such as onesbased on abstract designs.

As long ago as 1927, Charles Spearman, the pio-neer psychometrician who discovered g, proposed ahypothesis to explain the pattern: the size of theblack-white difference would be "most marked injust those [subtests] which are known to be saturat-ed with g." In other words. Spearman conjecturedthat the black-white difference would be greatest on

* Rushton has posted all of the articles at www.ssc.uwo.ca/psychol-ogy/faculty/rushton_pubs.htm.

[19]

COMMENTARY SEPTEMBER 2005

tests that were the purest measures of intelligence,as opposed to tests of knowledge or memory.

A concrete example illustrates how Spearman'shypothesis works. Two items in the Wechsler andStanford-Binet IQ tests are known as "forwarddigit span" and "backward digit span." In the for-ward version, the subject repeats a random se-quence of one-digit numbers given by the examin-er, starting with two digits and adding another witheach iteration. The subject's score is the number ofdigits that he can repeat without error on two con-secutive trials. Digits-backward works exactly thesame way except that the digits must be repeated inthe opposite order.

Digits-backward is much more g-loaded thandigits-forward. Try it yourself and you will see why.Digits-forward is a straightforward matter of short-term memory. Digits-backward makes your brainwork much harder.

The black-white difference in digits-backwardis about twice as large as the difference in digits-forward. It is a clean example of an effect that re-sists cultural explanadon. It cannot be explained bydifferential educational attainment, income, or anyother socioeconomic factor. Parenting style is ir-relevant. Reluctance to "act white" is irrelevant.Motivation is irrelevant. There is no way that anyof these variables could systematically encourageblack performance in digits-forward while de-pressing it in digits-backward in the same test atthe same time with the same examiner in the samesetting.

In 1980, Arthur Jensen began a research pro-gram for testing Spearman's hypothesis. In hisbook The g Factor (1998), he summarized the re-sults from seventeen independent sets of data, de-rived from 149 psychometric tests. They consis-tently supported Spearman's hypothesis. Subse-quent work has added still more evidence. Debatecontinues about what the correlation between g-loadings and the size of the black-white differencemeans, but the core of Spearman's original conjec-ture, that a sizable correlation would be found toexist, has been confirmed.

During the same years that Jensen was investi-gating Spearman's hypothesis, progress was alsobeing made in understanding g. For decades, psy-chometricians had tried to make g go away. Confi-dent that intelligence must be more complicatedthan a single factor, they strove to replace g withmeasures of uncorrelated mental skills. They there-by made valuable contributions to our understand-ing of intelligence, which really does manifest itselfin different ways and with different profiles, but

getting rid oig proved impossible. No matter howthe data were analyzed, a single factor kept domi-nating the results.

By the 198O's, the robustness and value oi g asan explanatory construct were broadly acceptedamong pyschometricians, but little was knownabout its physiological basis. As of 2005, we knowmuch more. It is now established that g is by farthe most heritable component of IQ. A variety ofstudies have found correlations between g andphysiological phenomena such as brain-evokedpotentials, brain pH levels, brain glucose metabo-lism, nerve-conduction velocity, and reaction time.Most recently, it has been determined that a high-ly significant relationship exists between g and thevolume of gray matter in specific areas of thefrontal cortex, and that the magnitude of the vol-ume is under tight genetic control. In short, wenow know that g captures something in the biolo-gy of the brain.

So SPEARMAN'S basic conjecture was correct—the size of the black-white difference and g-

loadings are correlated—and g represents a biolog-ically grounded and highly heritable cognitive re-source. When those two observations are put to-gether, a number of characteristics of the black-white difference become predictable, correspondwith phenomena we have observed in data, andgive us reason to think that not much will changein the years to come.

One imphcadon is that black-white convergenceon test scores will be greatest on tests that are leastg-loaded. Literacy is the obvious example: peoplewith a wide range of IQ's can be taught to readcompetently, and it is the reading test ofthe NAEPin which convergence has reached its closest point(.55 standard deviations in the 1988 test). Morebroadly, the confirmadon of Spearman's hypothesisexplains why the convergence that has occurred onacademic achievement tests has not been matchedon IQ tests.

A related implication is that the source of theblack-white difference lies in skills that are hardestto change. Being able to repeat many digits back-ward has no value in itself. It points to a valuable un-derlying mental ability, in the same way that per-centage of fast-twitch muscle fibers points to an un-derlying athledc ability. If you were to pracdce recit-ing digits backward for a few days, you could in-crease your score somewhat, just as training can im-prove your running speed somewhat. But in neithercase will you have improved the underlying abihty.As far as anyone knows, g itself cannot be coached.

[20]

THE INEQUALITY TABOO

The third imphcation is that the "Flynn effect"will not close the black-white difference. I am re-ferring here to the secular increase in IQ scoresover time, brought to public attention by JamesFlynn. The Flynn effect has been taken as a reasonfor thinking that the black-white difference is tem-porary: if IQ scores are so malleable that they canrise steadily for several decades, why should not theblack-white difference be malleable as well?

But as the Flynn effect has been studied over thelast decade, the evidence has grown, and nowseems persuasive, that the increases in IQ scores donot represent significant increases in g. What theincreases do represent—whether increases in spe-cific mental skills or merely increased test sophis-tication—is still being debated. But if the black-white difference is concentrated in g and if theFlynn effect does not consist of increases in g,the Flynn effect will not do much to close the gap.A 2004 study by Dutch scholars tested this ques-tion directly. Fxamining five large databases, theauthors concluded that "the nature of the Flynn ef-fect is qualitatively different from the nature ofblack-white differences in the United States," andthat "the implications of the Flynn effect for black-white differences appear small."

These observations represent my reading of abody of evidence that is incomplete, and they willsurely have to be modified as we learn more. Buttaking the story of the black-white IQ difference asa whole, I submit that we know two facts beyondmuch doubt. First, the conventional envirormientalexplanation of the black-white difference is inade-quate. Poverty, bad schools, and racism, which seemsuch obvious culprits, do not explain it. Insofar asthe environment if the cause, it is not the sort of en-vironment we know how to change, and we havetried every practical remedy that anyone has beenable to think of. Second, regardless of one's readingof the competing arguments, we are left with an IQdifference that has, at best, narrowed by only a fewpoints over the last century. I can find nothing inthe history of this difference, or in what we havelearned about its causes over the last ten years, tosuggest that any faster change is in our fiiture.

IV

ELITES THROUGHOUT the West are living a he,basing the futures of their societies on the as-

sumption that all groups of people are equal in allrespects. Lie is a strong word, but justified. It is alie because so many elite politicians who profess tobelieve it in public do not believe it in private. It is

a lie because so many elite scholars choose to ig-nore what is already known and choose not to in-quire into what they suspect. We enable ourselvesto continue to live the lie by establishing a tabooagainst discussion of group differences.

The taboo is not perfect—otherwise, I wouldnot have been able to document this essay—but itis powerful. Witness how few of Harvard's facultywho understood the state of knowledge about sexdifferences were willing to speak out during theSummers affair. In the public-policy debate, wit-ness the contorted ways in which even the oppo-nents of policies like affirmative action frametheir arguments so that no one can accuse them ofsaying that women are different from men orblacks from whites. Witness the unwillingness ofthe mainstream media to discuss group differ-ences without assuring readers that the differ-ences will disappear when the world becomes abetter place.

The taboo arises from an admirable idealismabout human equahty. If it did no harm, or if theharm it did were minor, there would be no need towrite about it. But taboos have consequences.

The nature of many of the consequences mustbe a matter of conjecture because people are sofearful of exploring them. Consider an observationfurtively voiced by many who interact with civilservants: that government is riddled with peoplewho have been promoted to their level of incom-petence because of pressure to have a staff with thecorrect sex and ethnicity in the correct proportionsand positions. Are these just anecdotes? Or shouldwe be worrying about the effects of affirmative ac-tion on the quality of government services? Itwould be helpful to know the answers, but we willnot so long as the taboo against talking aboutgroup difference prevails.

How much damage has the taboo done to theeducation of children? Christina Hoff Sommershas argued that willed blindness to the different de-velopmental patterns of boys and girls has ledmany educators to see boys as aberrational andgirls as the norm, with pervasive damage to the wayour elementary and secondary schools are run. Isshe right? Few have been willing to pursue theissue lest they be required to talk about innategroup differences. Similar questions can be askedabout the damage done to medical care, whosepractitioners have only recently begun to acknowl-edge the ways in which ethnic groups respond dif-ferently to certain drugs.

How much damage has the taboo done to ourunderstanding of America's social problems? The

[21]

COMMENTARY SEPTEMBER 2005

part played by sexism in creating the ratio of malesto females on mathematics faculties is not the ratiowe observe but what remains after adjustment formale-female differences in high-end mathematicalability. The part played by racism in creating dif-ferent outcomes in black and white poverty, crime,and illegitimacy is not the raw disparity we observebut what remains after controlling for group char-acteristics. For some outcomes, sex or race differ-ences nearly disappear after a proper analysis isdone. For others, a large residual difference re-mains. In either case, open discussion of group dif-ferences would give us a better grasp on where tolook for causes and solutions.

WHAT GOOD can come of raising this divisivetopic? The honest answer is that no one

knows for sure. What we do know is that the taboohas crippled our abiHty to explore almost any topicthat involves the different ways in which groups ofpeople respond to the world around them—whichmeans almost every political, social, or economictopic of any complexity.

Thus my modest recommendation, requiring nochange in laws or regulations, just a little moregumption. Let us start talking about group differ-ences openly—all sorts of group differences, fromthe visuospadal skills of men and women to the vi-vaciousness of Italians and Scots. Let us talk aboutthe nature ofthe manly versus the womanly virtues.About differences between Russians and Chinesethat might affect their adoption of capitalism. Aboutdifferences between Arabs and Europeans thatmight affect the assimilation of Arab immigrantsinto European democracies. About differences be-tween the poor and non-poor that could inform pol-icy for reducing poverty.

Even to begin listing the topics that could be en-riched by an inquiry into the nature of group dif-ferences is to reveal how stifled today's conversa-tion is. Besides liberating that conversation, anopen and undefensive discussion would puncturethe irrational fear of the male-female and black-white differences I have surveyed here. We wouldbe free to talk about other sexual and racial differ-ences as well, many of which favor women andblacks, and none of which is large enough to fright-en anyone who looks at them dispassionately.

Talking about group differences does not requireany of us to change our politics. For every implica-tion that the Right might seize upon (aff'irmative-acdon quotas are ill-conceived), another gives fod-der to the Left (innate group differences help ra-

donahze compensatory redistribution by the state).But if we do not need to change our pohtics, talk-ing about group differences obligates all of us torenew our commitment to the ideal of equality thatThomas Jefferson had in mind when he wrote as aself-evident truth that all men are created equal.Steven Pinker put that ideal in today's language inThe Blank Slate, writing that "Equality is not theempirical claim that all groups of humans are in-terchangeable; it is the moral principle that indi-viduals should not be judged or constrained by theaverage properties of their group."

Nothing in this essay implies that this moral prin-ciple has already been realized or that we are pow-erless to make progress. In elementary and sec-ondary education, many outcomes are tractable evenif group differences in ability remain unchanged.Dropout rates, literacy, and numeracy are alltractable. School discipline, teacher performance,and the quality ofthe curriculum are tractable. Aca-demic performance within a given IQ range istractable. The existence of group differences neednot and should not discourage attempts to improveschooling for millions of American children who arenow getting bad educations.

In university education and in the world of work,overall openness of opportunity has been trans-formed for the better over the last half-century. Butthe policies we now have in place are impeding, notfacilitating, ftirther progress. Creating double stan-dards for physically demanding jobs so that womencan qualify ensures that men in those jobs willnever see women as their equals. In universities, af-firmative action ensures that the black-white dif-ference in IQ in the population at large is broughtonto the campus and made visible to every student.The intentions of their designers notwithstanding,today's policies are perfectly fashioned to createseparation, condescension, and resentment—andso they have done.

The world need not be that way. Any universityor employer that genuinely applied a single set ofstandards for hiring, firing, admitting, and pro-moting would find that performance really is dis-tributed indistinguishably across different groups.But getting to that point nationwide will require usto jettison an apparatus of laws, regulations, andbureaucracies that has been 40 years in the making.That will not happen until the conversation hasopened up. So let us take one step at a time. Let usstop being afraid of data that tell us a story we donot want to hear, stop the name-calhng, stop thedenial, and start facing reality.

[22]