Herrnstein Revisited

The publication two years ago of Professor Richard Herrnstein's Atlantic Monthly article about I.Q. testing and genetic-racial differences in human intelligence carried the discussion of these issues into the political arena. The article by Joseph Alsop (the Globe, August 27) and some comments in the September 3 issue of Time magazine have focused new attention on the subject.

Herrnstein's article was embarrassingly misinformed on several counts; I would like to remark here on contravening evidence from the substantial scientific research about these matters extending over many decades.

Early I.Q. testing and its modifications by Stanford, Wexler, Terman et al are all based on the studies of Binet and Simon at the end of the last century. The original I.Q. testing technique was a careful attempt to replace the rather arbitrary methods of judging human potential by measurement on a sounder scientific scale. However Binet and Simon knew that their "measuring scale of intelligence," properly speaking, does not measure true intelligence, because intellectual qualities are not superposable, thus cannot be measured as linear surfaces but are rather a classification, "a hierarchy among diverse intelligences." They wrote: "intelligence, better mentality, encompasses many different faculties, including sensation, emotions, and others." They considered the most important faculty within intelligence to be judgment, "otherwise called good sense, practical sense, initiative, the faculty of adapting oneself to circumstances. To judge well, to comprehend well, to reason well, these are the essential activities of intelligence."

As they knew and said at the time, I.Q. is not a scale of such human intelligence, but something very different: it expresses the relationship between a so-called "mental age" and the chronological age of a child in development (as a ratio, hence the name "intelligence quotient"). The technique was conceived and designed to test children, not adults. The notion that such a ratio makes for comparison among adults was and is ridiculous. This is clear especially from looking at the capacities of exceptionally intelligent adults: comparing the I.Q.'s of, say, Einstein, Judge Brandeis and the theologian Paul Tillich says nothing at all about the comparative capacities for judgment, good sense, initiative, comprehension, and reasoning of these men. It is well known that some great writers can hardly add two and two, while first-rate mathematicians stumble over spelling and common English.

I.Q., then, measures not intelligence in adults, but something else, namely chronological age versus mental age, or better, "developmental age," in children: the psychological development at a given chronological age. Since the rate and scope of such development varies enormously in the population, I.Q. can change over time: persons do not have a fixed I.Q. from birth to death, nor do all children even from childhood through adolescence. The ten-year-old with an I.Q. of 100 may score 120 by the end of adolescence; the child with an I.Q. of 130 may drop to 120 by age eighteen. Beyond late adolescence the I.Q. itself is meaningless, and the modern widespread testing of I.Q. among late adolescents and young adults as a criterion for admission to universities is particularly harmful, since history shows that many persons who later made enormous contributions to society did poorly on I.Q. tests and were excluded from universities. These remarks do not mean that if an industrial firm requires specific capacities in a worker, or a university department needs certain capabilities in its students, an I.Q. screening is not helpful and justified; but it should be clear that what is being tested then is not human intelligence, but faculties needed for specific tasks. Even in these contexts, but more critically in the context of school testing, it is important to evaluate the results of an I.Q. test in the light of the previous experience of the person entering the test situation: his mood, attitudes, motivations, anxieties affect his performance and depend on emotional and cultural factors which are almost never properly evaluated. The test itself is no more than a diagnostic, a sophisticated kind of x-ray; and even the most "objective" x-ray plate is interpreted differently by different experts.

Hopefully, this clears the confusion between what an I.Q. test does and what it does not do. But Herrnstein is even more confused about the relation between human intelligence and heredity. My own extensive studies on the factors which produce retardation have demonstrated beyond doubt that extremely low I.Q.s among children (below 55) are due to neuro-psychiatric disorders, the large collection of developmental disorders and illnesses which interfere with mentation. These are to a limited extent of a genetic nature (metabolic errors) and to a larger extent due to external influences which have operated upon the embryo or the growing child. (This population with extremely low I.Q.s includes children of highly intelligent parents; indeed, a so-called idiot or imbecile comes usually from intelligent parents.) The higher I.Q.s (55-70) among the mentally retarded are more often due to genetic factors. However, no one has ever found a racial pattern in the group of intellectual inadequates. My studies (funded by the National Institute of Health) have provided evidence that the composition of this group shows no relationship to racial factors, but a definite relationship to those parents in the population who are less intellectually endowed than others. Generally the I.Q. of a child lies between 15 points (about one standard deviation for the roughly normal distribution of I.Q. in the population at large) above the higher I.Q. of the two parents, and 15 below the lower I.Q. Thus a man with a supposed I.Q. of 130 married to a woman with a supposed I.Q. of 120 produces mostly children with I.Q.s between 145 and 105. Again, two parents both of I.Q. about 110 will ordinarily produce children with I.Q.s between 125 and 95, i.e. from "quite bright" to low-normal.

Thus if men and women of superior intelligence intermarried only among themselves, the dire consequence would be indeed as Herrnstein predicts--a hereditary, intelligent elite far removed from a much larger pool of undistinguished and anonymous persons. However Herrnstein ignores entirely the known history and scientific studies showing that highly intelligent persons usually mate with persons with other qualities and that consequently intelligence tends constantly to equalize in the population. As is well known, successive generations of once-outstanding families return, sometimes gradually but sometimes quite abruptly, into the anonymity of the general pool; conversely, outstanding individuals regularly emerge from quite ordinary families.

The sum of these objections to Herrnstein's claims is this. Human intelligence is composed of many diverse personality faculties, the interaction of which is not ascertainable, and demonstrably not tested by I.Q. tests, which measure only development rates among children and do this only in terms of certain faculties which may be desirable and needed under certain conditions but are quite useless in other situations. The genetics of human intelligence is not scientifically well enough established to be used as a political weapon; the propagation of intelligence through generations obeys sociological, psychological and biological laws which override completely the possibility of political predictions based on the partial knowledge given by I.Q. measurements.

Dr. Clemens E. Benda is an internationally recognized authority on mental retardation and child development.

Some great writers can hardly add two and two...