[In the following review, Krauthammer praises The Mismeasure of Man, but indicates that Gould, like those scientists whose theories Gould debunks, has his own personal biases.]

The 1964-65 New York World's Fair had a pavilion called Sermons from Science. Being one of those adolescents with an insatiable appetite for things scientific, I wandered in and was treated to fascinating films on such mysterious phenomena as resonance (that which allowed Caruso to shatter glass with his voice) and electromagnetic radiation (that which brought “Father Knows Best” into my home). Little did I know that I had blundered into a scientific sting operation. For, as soon as the lights went up, a shill stepped onto the stage, declaring that resonance and electromagnetic radiation demonstrated (by analogy, I suppose) God's invisible presence in the world. Next a pod of earnest young men appeared, brandishing pamphlets; they proceeded to work the crowd, urging a return to Christ.

That was my first unhappy encounter with the use of science for nonscientific ends. My latest encounter came from reading The Mismeasure of Man, Stephen Jay Gould's excellent case study of science in the service of ideology. Gould has written a witty, ironic, and often original history of a scientific sideshow that has been running for 150 years: the exasperating search for the psychologist's stone, the quantification of intelligence. And it is a brilliant account of how that scientific enterprise has been used to justify class and racial inequality.

The modern career of the idea of intelligence begins in the middle of the 19th century. It was then hypothesized that intelligence was determined by skull size. A second hypothesis, more accurately a belief, held that intelligence determined a race's position in the social hierarchy. These two hypotheses were twined into a kind of theoretical folie à deux: the truth of one would be assumed long enough to establish the truth of the other. For example, when Samuel George Morton, the great 19th-century Philadelphia physician and skull collector, reported that skulls of Caucasians were larger than those of Mongolians which, in turn, were larger than those of Negroes, he was satisfied that he had shown both that intelligence was a function of skull size and that the regnant racial hierarchy was justified by modern scientific evidence.

Such research might appear to some as merely a bizarre detour in the history of science but Gould shows it to be the result of the perfectly ordinary investigative activities of the greatest scientists of the time. Gould touches only lightly on the eccentrics. He concentrates on the serious work of giants like Paul Broca. More importantly, Gould is not interested in the occasional fabricated evidence, but in the glaring—and honest—distortions produced by unconscious bias. He has gone back and re-examined the raw data from which Morton, Broca, and others built their inferences. He not only demonstrates errors in calculation and computation, in measurement and inference; he shows that these errors all tend consistently in one direction—in favor of the original (racist or hereditarian) hypothesis. For example, in studying skull size Broca ignored contaminating factors like body weight, size, and nutrition. But not always. When Louis Pierre Gratiolet confronted Broca in the French Academy with evidence that German brains were 100 grams heavier than French brains, Broca responded, “It would be easy for me to show him that he can grant some value to the size of the brain without ceasing for that to be a good Frenchman.” In a remarkable display of sophisticated evidence sifting, he then proceeded to analyze systemically all the uncontrolled factors (as we call them today) that could account for the apparent disparity. He nimbly emerged with the conclusion that, all things being equal, the French brain was larger than the German. But when it came to examinations of skull size in women and Africans, Broca showed little interest in the equality of all things.

Like the other great quantifiers of intelligence, Broca was undoubtedly unaware of his selective use of his own talents. And that is Gould's major theme: the ubiquity—and innocence—of ideologized science. The path it takes is circular. It starts with a hypothesis, generated in response to an ideological need (for example, to show that human races have a hierarchical order of intelligence, whites at the top and blacks at the bottom). It proceeds to gather scientific data that fit the hypothesis. And it then returns to the hypothesis confirmed. At that point the scientist either abjures invitations to draw the necessary political implications and withdraws with appropriate scientific humility to the sanctity of the laboratory; or he jumps into the ideological fray to advocate the only enlightened social policy that “science” permits: separation of the races, strict limitation of immigration and, ultimately, the Darwinian justice of enduring social and racial inequalities.

By the 20th century craniometry had been replaced by mental testing. And, under the influence of the new science of eugenics, it set out to prove not only that classes and races are arrayed in hierarchies of intelligence but that these differences are innate, inheritable, biological, and, therefore, unchangeable. The most famous of these IQ tests was administered by Robert Yerkes to 1.75 million American inductees into the army in World War I. The results provided objective support for existing prejudices. They also lent legitimacy—and urgency—to the 1924 immigration laws which severely restricted quotas for Southern and Eastern European races. Gould has done us the service of wading through the original 800-page study which was read by almost no one and used by almost everyone. Piece by piece he takes the tests apart. He demonstrates that if they measured anything (which is doubtful, given the harried, uncontrolled, and haphazard way they were administered), they measured acculturation. Buried in the report he finds a chart showing a striking linear relationship between the score of the immigrant recruit and the number of years he had spent in the United States. The average test scores of the foreign-born recruit rose consistently with his years of residence in America. Unless one assumes that the genetic stock of immigrants coming to America declined systemically over every five-year period between 1885 and 1914, one is left with the conclusion that what these tests measured was literacy in English and acculturation to the United States. If you doubt this, test yourself on these 1914 test items: (1), Crisco is a: patent medicine, disinfectant, toothpaste, food product; (2) The number of Kaffir's legs is: two, four, six, eight; (3) Christy Matthewson is famous as a: writer, artist, baseball player, comedian. Small wonder that on a similar test 83 percent of the Jews, 87 percent of the Russians, 80 percent of the Hungarians, and 79 percent of the Italians who took it scored as morons. (It was a technical term then, denoting a condition somewhere between subnormality and imbecility.)

Gould concludes with a sophisticated, highly mathematical critique of the work of Cyril Burt and Arthur Jensen, the major modern hereditarians in the intelligence debate. He attacks their theories not so much for their peripheral racial implications, as for their core assumption: the belief that “general intelligence” exists. The idea of general intelligence was developed by Charles Spearman, an early 20th-century statistician and psychologist who derived, by the method of factor analysis, a common factor in all intelligence tests. He called it “g” and reified it as intelligence itself. Gould shows, first, that it is arbitrary, that the existence of “g” is an artifact of the tests chosen to measure it. For example, the data of intelligence testing can, with equal mathematical justification, be used to support the notion of clusters of mental abilities (mathematical, verbal, spatial, etc.) rather than a single faculty called “g.”

Furthermore, even if a universal “g” does exist, variation in scores among individuals can be explained with equal if not greater plausibility by invoking environmental rather than genetic causation. And, finally, even if one grants both that there is a “g” and that its distribution within populations is genetic, this in no way proves that the difference in intelligence between populations is genetic. An example used by the psychologist Leon Kamin is instructive. Suppose we have two bags of seed, bag “a” and bag “b.” Both have identical quantities of identical genetic mixes of seed. We plant the seeds from bag “a” on fertile ground and the seeds from bag “b” on barren ground. The distribution of heights within each population of grains is clearly related to its genes. But what of the fact that the grains from bag “a” grow higher on average than the grains from bag “b”? It is clearly the result of environment. (Similarly the fact that blacks consistently score lower than whites on IQ tests tells us nothing about genetic or innate differences between the races.)

This is not to say we should assume a priori that innate differences between groups of people cannot exist. American blacks are more susceptible to sickle cell anemia than American whites and Ashkenazi Jews are more susceptible to Tay-Sachs disease than Gentiles. There is no reason to decide in advance that genes have nothing to do with the fact that some groups are better than others at chess or basketball or the 100-yard dash. My objection to the endless debate over racial differences in intelligence is that it is a supremely uninteresting question. The very notions of race and intelligence are so vague that any conclusion about a correlation between them is bound to be vaguer still, to be of little biological interest, and to be essentially unverifiable, since we cannot control for environmental conditions affecting different races. And what socially useful results can even a positive correlation yield? Assume for a moment that we do have a positive correlation between race and a certain socially useful characteristic, say, height. Assume, further, that that characteristic is a legitimate qualification for a social role, say, to be a policeman. Now, the average Japanese is smaller than the average American. But the overlap in populations is enormous. In other words, even though a correlation between race and height exists, it is very weak. To decide whether an individual applicant is tall enough to be a policeman, one doesn't inquire about race. One measures height. Race is simply not a useful predictor of height or other human qualities society cares about.

Gould charges those obsessed with finding racial differences with ulterior ideological motives. His elucidation of the history of 19th- and 20th-century intelligence measurers bears that out quite convincingly. But Gould also lugs around his own ideological baggage. He is an unabashed egalitarian. If he were to subject himself to the very analysis he turns on his opponents, he would probably conclude that his motivation in discrediting the hereditarians stems from his belief that races and classes should be treated equally. His skillful search through the history, methods, and conclusions of IQ testing to support his case for environmental causation stands on its own. His egalitarian zeal does not affect the validity of his work, but it does, I believe, lead him to one curious omission.

One reads Gould's methodical debunking of cultural bias in intelligence tests anticipating that in the end he will reveal what he thinks intelligence is and how we should use it. He uses the word often enough and he evidently believes it has a commonsense meaning. But because Gould is acutely aware of the harm that can be done in the name of intelligence, he is wary of saying anything about intelligence that may lead to its “reification,” or anything about IQ tests that may lend them legitimacy. His views must, therefore, be inferred from his rather approving treatment of Alfred Binet, the early 20th-century inventor of the intelligence test. Binet designed his test on commission from the French government specifically to identify children with learning disabilities. The purpose was to assign them to special classes designed to help them improve, and not to stigmatize them as innately or permanently inferior. With this approach, says Gould, “mental testing becomes a theory for enhancing potential through proper education.” Gould's complaint seems to be that after Binet, hereditarians misused mental testing and perverted its original aims. In a candid revelation, Gould says,

I feel that tests of the IQ type were helpful in the diagnosis of my own learning-disabled son. His average score, the IQ itself, meant nothing, but was only an amalgam of some very high and very low scores; but the pattern of low values indicated his areas of deficit.

From this I draw certain conclusions about Gould's views on IQ tests, views so sensible that I wish he had stated them more explicitly. First, IQ tests can measure certain abilities—verbal, mathematical, spatial, etc. IQ subscores are a valuable educational tool. They allow us to identify areas of intellectual strength and weakness. On the other hand, the average score, the IQ, is a quite useless number because to arrive at a single score we must arbitrarily assign weights to different mental abilities. The idiot savant has prodigious recall and little else; Bobby Fischer has spatial and computational abilities so hypertrophied that they have crowded out his other faculties; the patient suffering from Korsakoff's disease forgets everything new he learns within about 60 seconds, but retains knowledge acquired before his illness and is otherwise intellectually intact. (Many Korsakoffians do well on IQ tests.) Who is more intelligent? Can such a question have any meaning?

It is a pity that Gould does not affirm more strongly what it is possible to say about intelligence and what limited usefulness one can ascribe to IQ tests. His voice is needed because the reputation of IQ tests has sunk so low that the one use which Binet intended—which Gould approves—is now considered illegitimate by radical egalitarians. In California, the courts have declared unconstitutional IQ tests which had been used as an adjunct in selecting children who cannot learn and who require remedial education. The system had been selecting disproportionately more black than white children. On these grounds, the judge declared the tests discriminatory. He ordered the state to find a system which would produce black and white slow learners in numbers that matched the proportion of blacks and whites in the general population. We have now returned to the classic situation of science in the service of ideology. The ideology here is that there are no innate differences between whites and blacks; therefore there should be no difference between the numbers of retarded learners among them; and therefore the tests themselves must be racist. But, the fact is, black children suffer more social and educational deprivation than white children. It is not surprising, therefore, that among slow learners in our society there should be a disproportionately larger number of blacks than whites. To deny these black children remedial help in the name of equality is as stupid and cruel as it was to deny them such help in the past on the grounds that they were genetically inferior. To the child it makes no difference whether his low score is the result of genes or environment. To the ideologue, it matters greatly. If there is a lesson to be learned from The Mismeasure of Man it is that no child should be sacrificed at the altar of ideology—even an ideology that invokes the authority of science.