The Bell Curve: Rekindling A Dead Debate

— John Vandermeer

THE RECENT FANFARE associated with the publication of Herrnstein and Murray's The Bell Curve is surprising, given the lack of anything new in the book. It recycles two basic ideas. One is trivially obvious, the other is at best dubious.

The obvious point is that people who tend to score well on IQ tests tend to do well in our society. It is also true that people born wealthy tend to live wealthy and people born poor tend to live poor, but I doubt I will go on the talk show circuit for having made that observation.

The second point, which generates all the attention and controversy, is that cognitive ability is hereditary. One component of this argument, and the one likely to cause the most serious social damage, is the suggestion that the average IQ score differences among races are partially determined by genes, a point the authors actually concede to be pure speculation if you read their prose carefully.

The hereditarian part of their argument is cited as one of the six undeniably “true” things about cognitive ability -- “Cognitive ability is substantially heritable....” Since, according to them, this assumption is “...beyond significant technical dispute,” it is not in need of defense. And indeed one cannot find a defense in any of the rest of the book.

The Experts and Their Theories

Taking such controversial assertions as undeniably true is only positive if one defines one's experts carefully. The authors identify with the “classic” school of IQ psychologists, those who believe that IQ measures something interesting and important, and that whatever it is that it measures is substantially heritable. Undeniable truths amongst strong believers are not difficult to come by.

The general effect the authors seek to create is one of dispassionate scientists looking at all sides of the issue. They cite many of their critics, but fail to address the criticisms. They ignore challenges to their principle assertion, undoubtedly aware of the danger that the entire enterprise would fall as a house of cards if that assertion were false. If we exit their small community of propagandists -- the so-called classicists -- the debate shifts once again to where it was before the media trumped up this awful book. The evidence for a genetic difference in IQ scores between Blacks and whites is non-existent and the evidence for IQ's heritability in the first place remains questionable.

There is another aspect of this book that renders it difficult to criticize concisely. The strategy is to overwhelm the reader with statistics (there are 110 pages of statistical appendices, forty-four tables, and over ninety charts and graphs). Argument is replaced with repetition. They report on study after study that claims to show a tendency for this and that, with the ultimate aim of giving the reader the impression that a consensus is slowly building on the issue.

The method used in the natural and physical sciences of citing critical studies in this field tend to falsify their every premise. For example, a study in Germany, one of the few that really does speak to Black and white differences in IQ, is cited only in passing, and then as if it were only one of many such studies (it was quite unique, and uniquely speaks to racial IQ differences). In fact, that study is inconvenient for their thesis because it failed to find white children scoring higher than Blacks when raised together (indeed, the Black children scored higher than the whites).

In the end this book offers no new experiments or theories. It rehashes old theories and uses the same methods to trot out the same conclusions that were so obvious in the past, and sometimes reinvigorates interpretations that have already been convincingly falsified.

By identifying with a single “school,” the authors are able to frequently make the claim that “there is no longer a serious technical debate on this issue.” True enough, but remember the context. Amongst those who believe in IQ testing, who believe that there is a significant genetic component to IQ, who believe that IQ is a useful concept for social policy, there is little technical debate.

There is equally no serious technical debate about the existence of God amongst born-again Christians, the evils of capitalism amongst Marxists, nor the flatness of the earth amongst members of the flat earth society. That the experts now agree, of course, depends on the experts.

What we do know without doubt, and this is not contested in The Bell Curve, is that environment plays a very significant role in the determination of cognitive abilities, whether speaking of the ether-like “g” (general intelligence) or specific abilities. What we have been unable to show, with anywhere near the same level of confidence, is that there is a genetic component involved in cognitive abilities (within the normal range).

There may very well be a small genetic component,but try as they might, the IQ champions have yet to make the argument air tight. While The Bell Curve doesn't even try to defend the proposition,someone needs to, if I or many other skeptics are to be convinced. At the present time, the most relevant studies still are the family histories, the adoption studies, and the identical twin studies, and all of them are subject to major technical flaws that have not been (and perhaps cannot be) addressed.

Genetics, Environment and Accidents

The basic problem is this: Suppose Dick and Jane have IQ scores of 100 and 110, respectively. We know that the environmental differences they have experienced, both prenatal and during early development, have affected their scores, and perhaps have even created all of the observed difference. But there are two other forces that may be involved, genetics and random accidents of development. Honestly trying to sort out these three forces, we encounter an enormous methodological difficulty.

There is likely a strong indirect effect of both genetics and developmental accident on the environment, making their disentanglement an almost impossible statistical problem. For example, it is clear that in the United States a person born with a dark-colored skin will be treated differently by society than one born with light-colored skin -- that is because of our political history. The way society treats a person is part of the environment of that person. Thus any conceivable test will be confounded by the fact that genes, developmental accident and environment are not independent variables.

The identical twin studies offer an excellent example. (Curiously, The Bell Curve cites these studies only once, in an explanatory box on page 12.) Identical twins ought to provide the perfect experiment for disentangling environment from genetic effects. The problem is that they cannot deal with indirect effect cited above. Given that our society, to some unknown extent, categorizes people based on the way they look, people who resemble one another will be placed in similar environments by society at large. Thus, genetically identical people are bound to share certain environmental features more closely than genetically distinct people, and thus effectively be in more similar environments. This makes the comparison of identical twins less the experiment than it may seem.

Since its publication, The Bell Curve has been received by the scientific community with scorn. I know of no biologist who buys its argument, and judging from letters to the editor and op-ed pieces, social scientists have rejected it also. But that is probably not important. I remind the reader of the original article of Arthur Jensen (1969), reporting the same data and the same ideas, equally rejected by the scientific community. It was read into the Congressional Record, in its entirety, and probably had at least a small effect as part of the conservative argument against progressive social programs.

The Bell Curve is a conservative's delight for the same reasons. And with the recent changes in Congress, its publication couldn't have been better timed. The fact that the academic community rejects its arguments is of little consequence to the likes of Jesse Helms and Newt Gingrich.

ATC 54, January-February 1995

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Allowed HTML tags: <b> </b> <br> <br /> <a> </a> <em> </em> <strong> </strong> <cite> </cite> <code> </code> <ul> </ul> <ol> </ol> <li> </li> <dl> </dl> <dt> </dt> <dd> </dd> <div> </div> <img> <style> <font> </font> <blockquote> </blockquote> <hr>
  • Lines and paragraphs break automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.