What Is Intelligence?

To the Editors:

It is always interesting to note the reactions of a critic who is caught with his pants down, and Stephen Jay Gould’s reply to my letter [NYR, October 23] is no exception. Gould, in the original review of Jensen’s book which I criticized had denied the usefulness of the concept of general intelligence; I pointed out that recent “work with non-cognitive tests, such as measures of reaction time, speed of sensory discrimination, and in particular EEG evoked potentials,” none of them tainted by cultural factors, and all reactions to simple sensory stimuli, correlated high with IQ as measured by traditional tests. I went on to say that “In our own recent work EEG-evoked potentials show correlations with Wechsler IQ as high as does the Wechsler IQ with, say, the Binet IQ.” I concluded that “No assessment of the hypothesis of the existence of intelligence as a powerful factor can be considered meaningful which leaves out some of the most important sources of evidence supporting it.”

How does Gould react? He immediately falsifies the argument, by saying that “even if the correlations were as high as Eysenck claimed, what would it mean? It wouldn’t validate a notion of inborn general intelligence.” He then goes on to argue that the data would not support the inborn nature of the factor, but note that I was not concerned with the genetic aspects of intelligence but merely with the usefulness in scientific analysis of the concept of intelligence as such! This side-stepping of an inconvenient argument, by pretending that something quite different is intended is typical of Gould’s biased way of reasoning and arguing.

He also quotes some rather low correlations obtained by earlier workers between EEG-evoked potentials and intelligence; I had referred to our own work in my letter, which gave correlations of .84. Gould again criticizes what I didn’t say, and fails to deal with what I did say. Similarly, he does not deal with speed of sensory discrimination tests at all, although these too have resulted in correlations in the 70s and 80s. One could go on for a long time discussing the squirmings of a prejudiced reviewer caught out in the act, but what has been said will be sufficient to establish the point. Gould complains that I make ad hominem attacks on him; all I said was that he is not an expert in the field, and that technical books should be reviewed by experts. His attempted answer to my letter is clear evidence that I was right; Gould is not trying to assess Jensen’s book on academic and scientific grounds, but simply scrambling around for any facts or arguments that may support his prejudices, regardless of the true position. The smokescreen he produces to hide the nakedness of his argument may fool the non-expert; it is unlikely to fool those familiar with the literature.

H.J. Eysenck

The Bethlem Royal Hospital

and The Maudsley Hospital

London, England

Stephen Jay Gould replies:

I don’t wish to engage Mr. Eysenck in a protracted debate about the relative exposure of our respective arses; nonetheless, I can’t resist noting that his initial remark surprised me because I thought I had caught him in the same unenviable posture he ascribes to me. After all, he wrote a letter filled with ad hominem attacks and a single substantive point—and I was able to show that his only point is refuted directly in Jensen’s book, the very volume he was trying to defend against my supposedly inept onslaught.

Now Eysenck tries to wriggle out by describing Jensen’s lengthy tabulations as “some rather low correlations obtained by earlier workers.” Well, Jensen’s book appeared in 1979 with the latest results from many independent laboratories. The best we can say for Eysenck’s 0.84 (a figure not cited in his first letter) is that it doesn’t square with a great deal of contemporary work done with equal care. But, let my bumbling amateurishness yield to acknowledged expertise. On the subject of correlation between IQ and supposed non-cultural measures of neurological speed, I quote a paragraph from Donald D. Dorfman’s review of Eysenck’s latest book (The Structure and Measurement of Intelligence, Springer, 1979) in the British scholarly journal Nature (April 17, 1980). Dorfman, a psychometrician, is professor of psychlogy at the University of Iowa.

His review of the evidence for a correlation between IQ and amplitude and latency of evoked cortical potentials is misleading. The figures that he displays in support of a strong correlation are taken from Ertl’s early work which even Eysenck admits “suffered from technical and methodological deficiencies” (page 50). He then asserts that Shucard and Horn have also obtained “quite sizable correlations between AEP’s [averaged evoked potentials] and IQ” (page 50). In fact, those investigators reported a correlation of only +0.24 for fluid intelligence and an absence of correlation for crystallized intelligence in the article cited by Eysenck. He also presents data from “our own laboratories” collected in about 1973. No details are given and no reference is made to any relevant publications in scholarly journals.

Eysenck adds one additional charge in this letter: he accuses me of unfairly recasting the argument in terms of inborn intelligence when he was only trying to defend the utility of a notion of general intelligence, whatever its source. Did Eysenck read my review? It was primarily a critique of the basic historical argument for general intelligence, not a commentary on the nature-nurture problem. My first letter addressed the same theme. I argued that the concept of g, or general intelligence, was falsely based upon a mathematical abstraction—the first principal component of the positive correlation matrix of mental tests—that represents only one way among many of treating the same data. Therefore, g has neither logical nor physical necessity. Correlations between IQ and neurological reaction do not affirm the reality of g; correlations themselves do not imply cause.

Do come in from the cold, Mr. Eysenck.