Graduate School of Education Harvard University Cambridge, Massachusetts
Contrary to the implications of Christopher F. Chabris's "IQ Since The Bell Curve" [August], sciences--including
psychology--operate and progress in more than one way. Mr. Chabris focuses on the study of intelligence as it has been
carried out over the past century by psychometricians--psychologists interested in the measurement of abilities and
traits. His evidence and conclusions follow from the psychometrician's tools of the trade: paper-and-pencil tests,
delivered in settings remote from daily practice (typically classrooms or testing laboratories), often timed, and
loaded heavily with items that are linguistic, logical, or spatial in nature. Individuals trained in this
tradition--many of whom make a living from administering such batteries of tests--are reluctant to consider alternative
definitions or assessments of intellect. Often reviewing the data from a scientifically conservative perspective, they
tend to emphasize the unitary (rather than the pluralistic) aspects of intellect, and the hereditary (rather than the
environmental) explanations of individual differences in intelligence.
Fortunately, over the last two decades, the study of intelligence has become much more varied in terms of definitions,
approaches, sources of data, and tentative conclusions. Mr. Chabris mentions my theory of multiple intelligences,
developed over the last twenty years. He does not, however, mention that my theory grew out of an entirely different
source of data. Rather than creating and administering short-answer tests, I surveyed evidence from a wide range of
sciences. Abilities gained credibility as candidate intelligences to the extent that they fulfilled various criteria:
the presence of areas in the brain that serve particular functions and processes; the existence of people, like
prodigies or savants, who exhibit particular strengths or weaknesses in different areas; relevant evidence from
evolutionary psychology, cross-cultural experiences, and psychological experimentation. My initial list of intelligences
included seven: the language, logic, and spatial abilities recognized by other psychologists, and also musical, bodily,
interpersonal, and intrapersonal intelligences. After a decade of further research in this area, I believe that there is
firm evidence for an eighth, or naturalist's, intelligence, and some evidence for a ninth, or existential, intelligence.
From the perspective of most psychometricians, I have committed two cardinal sins. First, I did not define my
intelligences on the basis of test correlations, but rather out of a synthesis of information and data gleaned from the
biological, cultural, and psychological sciences. This is a perfectly valid scientific approach--it just runs counter to
what American psychometricians have done.
My second sin is hesitancy about creating a battery of tests of different intelligences. This is not a hesitation on
principle. I am interested in the nature of the different intelligences, and I am curious about the extent to which
they do or do not correlate with one another. Contrary to what Mr. Chabris maintains, I have actually been involved in
research that has sought to assess the intelligences of young children; three volumes on this research, Project
Spectrum, will shortly be issued by Teachers' College Press. Nor do I, nor do individuals associated with me, when
asked how abilities in the various intelligences are measured, respond that "we don't measure them," as Mr. Chabris
claims. We are, however, hesitant to endorse instruments that we feel are inadequate.
Assessments of multiple intelligences should fulfill two criteria. First, they need to be direct and as natural as
possible--the assessment should not occur through a paper-and-pencil instrument. Thus, a good measure of interpersonal
intelligence should examine how individuals actually interact with and evaluate one another--not how one answers
questions about such encounters. Second, they need to survey an intelligence in some detail. Whether we are dealing
with spatial, musical, or interpersonal abilities, each of these has many facets, and we should not be satisfied with a
one-shot assessment of a single facet.
Personally, I have seen my theory of multiple intelligences misused by teachers who apply quick-and-dirty assessments
and then label children as "spatial but not linguistic" or vice versa. On political grounds, I am disturbed by such
maneuvers. However, on scientific grounds, it will certainly be possible some day to assess the various intelligences
and to determine the extent of their correlations and also the heritability of each of them.
Department of Mathematics Quinnipiac College Hamden, Connecticut
Christopher F. Chabris's defense of Richard J. Herrnstein and Charles Murray's The Bell Curve reminded me of Senator George Aiken's 1966 proposal on the Vietnam war: "Declare victory and withdraw." Mr. Chabris's declaration of victory on behalf of Herrnstein and Murray includes errors small and large. On the small side--but of personal importance to my colleagues and me--Charles Murray was trained in political science, not in sociology. On the more substantial side, there is, for example, Mr. Chabris's misrepresentation of the report by the American Psychological Association's special panel on The Bell Carve, "Intelligence: Knowns and Unknowns," which appeared in the February 1996 issue of the American Psychologist. The committee did not "[back] all of the book's main claims," as Mr. Chabris asserts. Despite its strained effort to be fair to all sides, several of the panel's conclusions undercut central assertions in The Bell Curve. Here are a few examples:
Another example of Mr. Chabris's errors is his treatment of the book I and five other sociologists wrote, Inequality by
Design: Cracking The Bell Curve Myth (Princeton, 1996). Although he credits us with "hard work," he understates our
findings. We not only show, with the same survey data Herrnstein and Murray used, that social circumstances outweigh test scores in predicting life outcomes, we also show that the so-called intelligence test used in that survey was basically a measure of school performance (and thus heavily determined by teaching) and that in no way could intelligence explain historical changes in, nor cross-national patterns of, economic inequality. Furthermore, Mr. Chabris implies that there was only one other critical "scientific" response to The Bell Carve besides our own. This is incorrect. There are at least two other monograph-length studies that refute The Bell Curve--Dickens et al., Does The Bell Curve Ring True? (Brookings); and Knapp et al., The Assault on Equality (Praeger, 1996)--as well as many scholarly articles that do the same.
Mr. Chabris retreats, finally, to asking how anybody could challenge Herrnstein and Murray's "basic claim ... that smart
people [i.e., high scorers on school achievement tests] do better than dumb people." Well, if that were all The Bell
Curve was about, then the brouhaha really was a "phony controversy," as Mr. Chabris claims. But this is disingenuous.
The reason The Bell Curve made a splash lies in its assertions that differences in intelligence explain growing economic
inequality--they do not; that differences in intelligence explain racial inequality--they do not; that intelligence,
because it is genetic in origin, is largely immutable--it is not; and that, therefore, there is nothing that can or
o
ught to be done to ameliorate inequality--also, as history clearly shows, untrue.
University of Otago Dunedin, New Zealand
Christopher F. Chabris says much that needs saying. I have little sympathy with those who reject the plausibility of a
concept of general intelligence, or deny that IQ often measures something that behaves much like it. And I have no
sympathy with those who wish to discourage research on intelligence of the kind done by Arthur R. Jensen, and a positive
antipathy to those instrumental in suppressing Christopher Brand's book on intelligence. Still, I have two reservations.
First, the "Flynn effect," or massive IQ gains over time, causes more problems for IQ-based theories of intelligence
than Mr. Chabris acknowledges. No doubt there have been tremendous environmental changes over the past 150 years. But it
is not clear what tremendous change occurred between 1972 and 1982 in Holland to boost IQ scores on Raven's Matrices
Test (a good measure of general intelligence, or g) the equivalent of eight IQ points. The notion that IQ gains over the
last 80 years are intelligence gains would mean that in 1918 a majority of people suffered from mental retardation (see
my essay, "IQ Gains Over Time," in The Rising Curve: Long-Term Gains in IQ and Related Matters, edited by Ulric Neisser).
Second, Mr. Chabris's moderate account of The Bell Curve robs it of much of its dramatic impact. The meritocracy thesis
of
Herrnstein and Murray says nothing less than this: that every step toward the abolition of privilege and inequality of
opportunity takes us toward a frightful semi-caste society--one in which genes for intelligence are so correlated with
class that lower-class children have virtually no hope of bettering their lot. That should depress Right and Left alike.
I have attempted to show that genes for IQ are no more correlated with class today than they were 50 years ago, and that
evolution toward such a society is unlikely rather than certain (see my article, "Group Differences: Is the Good Society
Impossible?," Journal of Biosocial Science, 1996). The Bell Curve posits an America with both a huge and hopeless
underclass and an elite class so potent as to constitute a threat to our democracy. How the children of these two
classes are to compete on a basis of equality is not explained.
Department of Sociology University of California Berkeley, California
The argument that IQ tests are valid measures of intelligence rests ultimately on the mathematically complex subject of
statistical factor analysis, and therefore cannot be fully understood by people who lack technical training. But common
sense should convince any reasonable person that something is fishy.
A typical intelligence test asks a variety of questions, many of which are of the type one learns to answer in school.
For example, if you take a Wechsler Adult Intelligence Scale test, probably the leading IQ test currently in use, you
will be asked to define words; to answer factual and comprehension questions; and to do simple arithmetic problems.
An old version of the test, no longer in use, includes the following questions: what is the meaning of the word
"reluctant"?; name three kinds of blood vessels; why does the state require people to get a license before they get
married?; a coat that normally sells for $600 is reduced by 15 percent during a sale--what is the price of the coat?
Maybe I am missing something, but it seems perfectly plain to me that the likelihood of answering such questions
correctly is heavily affected by, among other things, the quality and quantity of the education one receives as well as
by one's intelligence.
In fact, it seems obvious to me that how well a person does on an IQ test depends on a variety of factors besides
intelligence: education, reading habits, experience with and attitudes toward taking tests, upbringing, and mental and
physical health, to name a few.
Intelligence is difficult to define precisely, but we can all agree that it refers to intellectual ability as opposed
to intellectual achievement. No one, so far as I know, thinks that physical ability can be reliably measured by having
people tested on a variety of physical skills--running, jumping, swimming, doing push-ups, etc.--and then doing a factor
analysis of the results. Why should intellectual ability be any different? Count me among those who regard the study of
intelligence as more pseudo-science than science.
Milwaukee, Wisconsin
Christopher F. Chabris succinctly restates the case that g, the general intelligence factor, exists and is measurable,
but, like so many who have addressed this topic, he does not appear to appreciate the limited significance of that
finding. For perspective, one merely has to turn a few more pages in the same issue to Joseph Epstein's article, "The
Old People's Socialist League," in which one finds George Orwell's line that "there are certain things one has to be an
intellectual to believe, since no ordinary man could be so stupid."
Mr. Chabris describes the Herrnstein and Murray theory that America is becoming "increasingly meritocratic," and without
blinking concurs that this would cause "wealth and other positive social outcomes" to be "distributed more and more
according to people's intelligence." The presumption is that intelligence can be the supreme coin of a merit market. Yet
nowhere is this so, not even in academia.
In The Bell Curve, Murray and Herrnstein, perhaps inspired by some of the recent computer-industry fortunes, proposed
that the time has come when high SAT's will be the path to and the justification for earthly power, a new order to be
sustained by mating trends. Consider how perfectly this vision conforms to a "revenge-of-the-nerds" scenario, for in it
not only do those with prodigious IQ's rate respect, but they get the girls! Will they also be the most popular kids in
high school?
Efforts to deny g simply enable IQ theorists to put off the day when they must finally answer the question: so what? It
must be understood that what is being measured when one measures g is a gradation of consciousness, and that human
consciousness is a gift of such enormity that any difference between some g and a lot of g is relatively minor: the
"dumb people" Mr. Chabris writes of actually possess a greatness in the scheme of things that the term "dumb people"
does not quite capture.
The study of g ought never to have as its goal a social, economic, or--God help us!--political exaltation of
intelligence. It should, rather, seek to place its abstractions in the real world and to increase the capacity of individuals to wrench, from whatever level of g they have, the fullest possible enjoyment of the banquet set before us.
Monterey County, California
In "IQ Since The Bell Curve" Christopher F. Chabris writes:
The most basic claim put forth by
Herrnstein and Murray was that
smart people do better than dumb
people. What is so troubling
about that? We rarely encounter
an argument over the fact that
beautiful people do better than
ugly people, or tall people do
better than short ones, though
each of these propositions is also
true.
What a simple world Mr. Chabris and his heroes inhabit. Just as people can be classified by a unilinear measure of
height, so they can be classified by unilinear measures of intelligence and beauty--with the tall, smart, beautiful
"doing better" than the poor, short, dumb, and ugly.
Though Mr. Chabris does not mention the word "eugenics," he captures its spirit. He has a very poor knowledge of
history if he does not understand the revulsion this kind of argument creates.
Monterey County, California
Department of Psychology University of Edinburgh Edinburgh, Scotland, UK
Christopher F. Chabris's well-informed article is in many ways welcome. We believe that it might help in the public
understanding of a much misunderstood and maligned area of research and hope that the article will be read widely. For
those of us who are researchers in the field, however, Mr. Chabris offers a doleful prediction: though we win all the
intellectual battles, we shall lose the war. Why? Because, he suggests, up-and-coming researchers are dissuaded from
studying intelligence differences by the bad press the field has received from a series of high-profile attacks,
including Stephen Jay Gould's The Mismeasure of Man and the responses to Richard J. Herrnstein and Charles Murray's The
Bell Curve.
But we think the propaganda war on intelligence differences need not be won; in fact, one should not even engage in the
contest, for two reasons. First, it takes place outside the realm of scientific discourse; its weapon is rhetoric and
its shield is public ignorance. Winning such a war would not equate with being correct, which is all that should matter
in science.
Second, within the relevant scientific community, the debate over the importance of traditional-style research into
human intelligence differences is nonexistent. The authoritative and disinterested report of the American Psychological
Association, "Intelligence: Knowns and Unknowns," concluded that: (1) the psychometric structure of intelligence
differences attracts broad scientific agreement; (2) intelligence differences are highly stable, moderately heritable,
and predictive of important life outcomes; and (3) intelligence differences are beginning to be understood in terms of
the brain's biology.
Such a consensus on the facts will go much further in convincing new researchers to work in the area than will locking
horns with IQ's detractors by writing impassioned articles in intellectual magazines. Perhaps Mr. Chabris's article is
itself a sign of this consensus, since he is an outsider who has been won over to the value of the area after having
apprised himself of the facts.
Edinburgh, Scotland, UK
I was grateful for Christopher F. Chabris's kind reference to what he calls my "worthy" book, The g Factor: General
Intelligence and Its Implications, and his detailing of its publishing--or should I say de-publishing?--history. The
book was first published in the United Kingdom in 1996 and then was withdrawn, as Mr. Chabris puts it,
after negative media coverage
and a frenzy reminiscent of thee
1970's. The publisher, Wiley,
also cancelled the book's
publication in the United States
before any copies went on sale.
Brand has since been fired from
his teaching position at
Edinburgh University, and has
yet to find another publisher.
In his article, Mr. Chabris asks whether the psychometricians in academia could have done more to make their case better
known. The question is a good one, especially since even new books on intelligence by Arthur R. Jensen, Richard Lynn, J.
Philippe Rushton, Michael Levin, Gunnar Adler-Karlsson, and the late Hans Eysenck are available mainly from mail-order
houses, and for the most part the general public has access to them only via libraries.
The problem has not, in fact, been a lack of publication, even in prestigious journals. Let me give a brief list here:
Tom Bouchard's study of intelligence in separated identical twins, finding 75-percent similarity, was published in
Science in 1990; the work of my Edinburgh colleagues and myself on the essential connection of IQ with "mental-intake
speed" was put to readers of Behavioral & Brain Sciences and the Lancet, and to international experts at a conference in
Berlin; my evidence that "fast-track learning" works without hindering lower-IQ children appeared in Intelligence, as
well as in my book; my criticisms of the environmentalist idea that the achievements of geniuses result chiefly from
hard work appeared in the British journal of Developmental Psychology.
So the problem is not a scarcity of published material. Rather, it is the case that psychologists are frightened to
admit that race differences in IQ are substantially heritable. In the public debates over The Bell Curve in the U.S., it
is noticeable that psychologists have contributed little.
In recent years anyone who speaks out in the media or in academia against egalitarianism puts his job on the line, as I
found out to my cost at Edinburgh University.
My own review of America's leading psychological critics of IQ today will appear shortly in Personality & Individual
Differences, but Mr. Chabris will find that it makes no difference. The social-science faculties of the English-speaking
world have become the parade grounds of neo-Stalinist egalitarians. These desperate environmentalists will not read what
Commentary recommends; and they keep discipline in universities by censorship, intimidation, and sacking.
Department of Psychology University of Wisconsin Oshkosh, Wisconsin
I want to thank Commentary for publishing the interesting discussion of the IQ debate by Christopher F. Chabris. The
article makes a variety of points that need making, particularly concerning how Stephen Jay Gould and others committed
to an extreme environmental interpretation of intelligence have managed to smear the scientific study of intelligence
with the epithet of racism, and by a remarkable variety of fallacious arguments. I have done my small part to expose the
errors in my 1994 article in Cognition and am happy to see others joining in.
I would, however, like to take issue with one point put forward by Mr. Chabris. He suggests that there is a strong
analogy between the suppression of the science of intelligence by extreme environmentalists and the earlier suppression
of behaviorism by cognitive psychologists. I do not believe this analogy has any merit. It is, rather, the behaviorists
who have spent much of their energies in this century suppressing the scientific study of cognition, personality, and
consciousness. It is only in the current decade, for example, that journals devoted to the serious investigation of
consciousness have appeared (e.g., Consciousness and Cognition and Psyche).
Behaviorists, as Mr. Chabris mentions, responded to Noam Chomsky's 1959 critique of B.F. Skinner's Verbal Behavior by
labeling it uninformed and irrelevant. Those behaviorists, however, were wrong. Chomsky's work has led, through
extension and reaction, to a rich science of linguistics, while Skinner's work on verbal behavior has led to nothing.
This is not due to political suppression but to the paucity of theoretical tools available to behaviorism with which to
account for linguistic phenomena, as Chomsky himself detailed in 1959.
Skinner's own reaction to Chomsky was extraordinary. In 1970, he wrote:
I published Verbal Behavior in
1957. In 1958 I received a 55 page
typewritten review by someone I
had never heard of named Noam
Chomsky. I read half a dozen
pages, saw that it missed the
point of my book, and went no
further.
Skinner had neither the wit nor the integrity to confront his critics squarely. It would be good to hear that today's
environmentalists, like Gould, were different, but I doubt that they are; to take just one example, Gould has never
responded, either publicly or privately, to my critique of his work.
School of Computer Science and Software Engineering Monash University Clayton, Victoria, Australia
After reading Christopher F. article, I am compelled to comment on his discussion of the so-called "Mozart effect" and
on some popular misconceptions about it. The "Mozart effect" is a term coined by the media to refer to the studies by
Gordon Shaw and myself on the enhancement of spatial-temporal performance after listening to music.
First, let me note that listening to ten minutes of a Mozart sonata is not going to make one smarter, thereby doing, as
some have claimed, "in ten minutes what years of intensive educational interventions often fail to accomplish." Our
studies with undergraduates found a short-term (10-15 minute) improvement in spatial-temporal performance, after
exposure to complex music, compared with listening to minimalist music, relaxation instructions, British trance music,
or an audiotaped story, or with performing the tests in silence. Short-term memory, as well as other areas of spatial
reasoning, was not affected. This work has been replicated in other laboratories.
Second, my testimony before a congressional committee that Mr. Chabris mentions was not in response to these Mozart
studies. I testified about the benefits of including music in the core curriculum, based on our (and others') studies
with preschoolers and kindergartners in which we found long-term improvement in spatial-temporal abilities after music
instruction, not simply listening to music.
The human brain is extremely plastic in the early years, with new neural connections being made very rapidly, partly as
a function of experience and learning. While listening to music may influence task performance by inducing some sort of
neural priming, neuroanatomical changes as a function of active involvement in music may account for the enhancements found in children. In other words, music involvement may be an experience that influences neural connections that are also relevant to spatial-temporal abilities. Similar effects of this type of influence on the adult brain are not likely to be found. We are currently performing studies with rats to help determine the possible neurophysiological changes brought about by musical enrichment.
All this suggests that the human brain has a remarkable capacity to change; it is not a static entity, and an
individual's capacities are not fixed at birth. While one's genetic code may determine one's potential intelligence, I
believe that early experience determines whether that potential will be reached. All told, the relative contributions
of nature and nurture are not yet known.
Department of Psychology Appalachian State University Boone, North Carolina
In Christopher F. Chabris's article, which I read with interest, I noted his account of the Mozart effect, in which he
mentions an experiment of mine. His description, however, misses one point that is related to the general issues raised in the article.
My experiment used a "backward-digit-span" task, which means that the subject is given a sequence of digits
(e.g., 4-3-1) and asked to report them back in reverse sequence (1-3-4). This simple task occupies a central point in
the controversies about intelligence. Arthur R. Jensen began his work with the assumption that differences in school
performance were explained by cultural differences, and he chose the digit-span task as a culturally-neutral measure of
intellectual performance. It was persistent group differences in performance with the backward-digit-span task that led
him to his current views. At present, many researchers (like Jensen) view backward-digit-span performance as a measure
of g. For an appreciation of both the nature and difficulty of the task, a reader could ask a friend to recite the
bar-code number from a magazine and then the reader could repeat the sequence backward.
Mr. Chabris's history of the Mozart effect, in which he details the problems with the original short article in Nature
by Frances H. Rauscher and Gordon Shaw, is accurate. But short reports in Nature are supposed to be followed up by more
detailed studies in specialized journals; this has not happened. Instead, the authors published a follow-up report in
Perceptual and Motor Skills, which, as Mr. Chabris notes, is a "low-prestige [journal] that many psychologists never
read."
In Nature, Rauscher and Shaw reported that they had combined the performance from three different tests because
performance among tasks was well-correlated. In their article in Perceptual and Motor Skills, however, they show
uncombined results from that experiment and indicate that the effect occurred with only one of these tasks. This is an
attempt to account for the failure of other researchers to replicate their original experiment. But in fact the two
reports of the same experiment contradict each other.
Mr. Chabris further suggests that an account of the failure to replicate the Mozart effect will never appear in a
prestigious journal of psychology. In this he may be correct. Recently a paper of mine on a Mozart effect experiment
was rejected by a prestigious journal, with an explanation similar to that of some colleagues of mine who told me that
following up on a bogus study was a waste of time.
Let me suggest here that there are two current disciplines of psychology: academic and popular. One is found in
university libraries and the other is found in the chain bookstores. Academic psychologists are very skeptical of the
Mozart effect, and for good reason. But it is a loss to the commonweal when they refrain from analyzing the claims of
popular psychology, thus allowing it to control public policy.
Intellequity Technology Services Natick, Massachusetts Dartmouth College Hanover, New Hampshire
Christopher F. Chabris is correct in stating that "many more phony controversies lie ahead" in the nature/nurture
dispute. Neither side is entirely right or entirely wrong. We all know that a tall boy has to have tall genes, a good
diet, sunshine, and plenty of practice to get to be a good, tall basketball player. Who doubts that alcohol in the
blood of a pregnant woman raises the risk of bearing a neurologically impaired child?
Nurture advocates deny the demonstrated facts of behavior genetics, but the nature side forgets that even identical
twins often differ due to subtly different in-utero environments. But some of the most important factors have been
generally ignored by academics on both sides of the controversy.
Lead and manganese are neurotoxins that influence brain development. These heavy metals damage brain structures,
interfere with normal neurotransmitters, and reduce cognitive competence and impulse control. Taking over twenty
conventional factors into consideration, industrial pollution is associated with significantly higher rates of criminal
violence (see our article, "Environmental Pollution, Neurotoxicity, and Violence," in Environmental Toxicology, edited
bye. Rose, 1998).
More recently, in a study we conducted of all communities in Massachusetts, we discovered that the use of
silicofluorides in water treatment increases the lead in children's blood and, as a consequence, increases rates of
violent crime. In general, where silicofluorides are used, we find that crime rates are doubled.
There is no end to the nature/nurture issues these considerations invoke. Here is an example. Calcium, which is
released from a pregnant woman's bones, is required for fetal neural and bone development. Exposure to lead in the
environment (paint dust, water from lead plumbing, soil) raises the potential of lead entering the gastrointestinal
tract, from which it can cross into the bloodstream. Once in the blood, especially if the diet is low in calcium, iron, and other essential minerals, lead is absorbed in the bones and in the developing brain. Little wonder that studies have consistently shown that lead or manganese absorption is associated with attention deficit disorder, hyperactivity, and lower IQ.
These effects are worse for blacks, who are six times as likely as whites to have levels of lead in their blood over 25 [micro]g/dL. Not surprisingly, according to national Health and Nutrition Evaluation Surveys, blacks' intake of calcium is only about three-quarters that of whites. But, as a study in Baltimore and Philadelphia sponsored by the Abell Foundation shows, providing school breakfasts has remarkable effects in improving students' performance and behavior. No wonder that the most effective part of Head Start programs (to the chagrin of some ideologues) was probably its nutritional component.
We cannot change genes (yet). And, as the failures of the last generation have shown, it is just as hard to change schools and teachers, not to mention the entire socioeconomic system of the United States. But we did get lead out of gasoline, and a good school breakfast is relatively cheap and effective to deliver to the poor. Maybe it will be easier to clean up the water supplied to our inner cities than to clarify the intellectual debates that professors love so much.
Department of Psychology Harvard University Cambridge, Massachusetts
Christopher F. Chabris's excellent article provides an informative analysis of post-Bell Curve scholarship on cognitive
ability. Two issues warrant further comment. First, he mentions a seminar in the Harvard psychology department during
which a member of Howard Gardner's research team discussed educational applications of the theory of multiple
intelligences. When I asked the speaker how she and her colleagues measured the different intelligences, I was startled
when she replied that they do not measure them. Because measurement is integral to psychology, it would seem desirable
to develop reliable and valid quantitative indices of the distinct intelligences postulated by this interesting theory.
Second, cognitive ability is related to certain mental disorders as well as to various social problems. Intrigued by
Herrnstein and Murray's work, my colleagues and I tested whether cognitive ability predicts the severity of
post-traumatic stress disorder (PTSD) symptoms in Vietnam combat veterans (see our papers in the Journal of Consulting
and Clinical Psychology, 1998, and the American Journal of Psychiatry, 1995). After controlling statistically for the
extent of combat exposure, we found that the lower a veteran's pre-combat IQ score, the more severe were his PTSD
symptoms today. Thus, higher cognitive ability may protect against the development of chronic PTSD symptoms following
exposure to horrific events.
Chapel Hill, North Carolina
It was gratifying to see the article by Christopher F. Chabris supporting the idea that there are indeed important
differences among people in IQ or general intelligence, with which I heartily agree. In the article, Mr. Chabris
mentions my book Human Cognitive Abilities (1993) and notes that it "offers the most extensive factor-analysis of
mental tests." But I regret that he was not able to include a full discussion of my three-level theory that includes
not only general intelligence at the apex of a hierarchy but also a set of broad abilities like general reasoning and
spatial ability at a second-order level, and a set of perhaps 50 or 60 special abilities, like native-language
vocabulary and even "absolute pitch" possessed by some musicians and others. Thus, my theory straddles the two sides
of the argument about whether there is only one general intellectual ability or a multiplicity of abilities.
Mr. Chabris rightly notes that Stephen Jay Gould's critique of general intelligence "has been widely accepted." But by
whom and on what basis? It has long irked me that so many otherwise reputable social scientists subscribe to Gould's
statements about mental abilities. In his writings, Gould has a wondrous way of sounding like a scientist who has
carefully studied factor analysis--the statistical procedures used to identify varieties of individual differences--and
has discovered certain fatal flaws in it. But he is wrong, both in his claim that the notion of general intelligence
cannot be supported by factor analysis and in his belief that factor analysis can say something about the hereditary
basis of mental ability (see my article in the January-February 1995 issue of Intelligence).
Quite apart from Gould's views, I have also been bothered by the reluctance of many social scientists and others to
accept the idea--or in fact the scientific finding--that variations in people's mental abilities may at least be partly
founded on genetics. I find it odd that while most people acknowledge the partially genetic basis for many physical
diseases and psychological disorders, they regard it as distasteful and unacceptable to think about cognitive ability
in terms of genes. We can regret that genes affect many aspects of our lives, but it is more rational to accept this
fact and think about ways of helping people (children and adults) to overcome the limitations that may be imposed on
them by adverse genes.
Transaction Publishers Rutgers University Piscataway, New Jersey
Christopher F. Chabris's article is both informative and courageous. I do not write to take issue with its substance
but simply to make a small but important correction: Hans Eysenck's book, Intelligence, is not a posthumous volume, as
Mr. Chabris says, but one that was completed in its entirety before his death. While this may appear as nitpicking,
there is a tendency to treat posthumous work as a lesser, even inauthentic, effort. Having sat with Eysenck through the
final stages of manuscript preparation, I can assure Mr. Chabris and Commentary readers that Intelligence is not only
"accessible and entertaining" but entirely authentic.
Toronto, Ontario, Canada
"IQ Since The Bell Curve" makes several significant points about the importance of general intelligence. If anything,
Christopher F. Chabris errs on the side of not making the case strongly enough. Empirical studies by numerous
researchers have shown intelligence to be the most influential factor in occupational performance in all jobs and as
having significant influence on most other areas of life. Intelligence has been shown to be positively related, among
other things, to eminence, creativity, moral reasoning, income, motivation, practical knowledge, and motor skills. By
contrast, an equally broad category that includes such things as impulsivity, truancy, accident-proneness, racial
prejudice, and authoritarianism has been found to have a negative correlation with intelligence.
The critics of intelligence should conduct empirical studies rather than just reporting anecdotes and offering examples.
Finally, although this is a highly technical issue, it should be noted that general intelligence is the factor that
contributes the most to verbal, spatial, and mathematical ability.
Department of Psychology St. Mary's University San Antonio, Texas
Neil Seeman:
I am greatly indebted to Christopher F. Chabris for injecting considerable equanimity and good sense into the
ideologically polarized debate over the measurability of human intelligence. Since the publication of The Bell Curve,
detractors of this ground-breaking work have been motivated by the most vicious manifestations of academic jealousy
and, as Mr. Chabris tells us, shocking scientific disingenuousness.
Professor Emeritus Educational Psychology University of California Berkeley, California
Christopher F. Chabris displays a keen perception of the current state of affairs related to The Bell Curve and the
so-called "IQ controversy." Among the article's remarkable features is the fact that it provides the first public
recognition that, with notable exceptions, the community of specialists in psychometrics has been shamefully negligent
in defending its science and what it knows to be true about the g factor, or general intelligence, the existence of
which is firmly established by massive evidence and is contradicted by virtually none. The only important direction
now left for further research on the nature of g is to go directly into the human brain itself. Research has already
revealed a number of anatomical and physiological brain correlates of psychometric g. This evidence, in addition to
the high heritability of g, suggests it is a biological as well as a psychological phenomenon.
Therefore, it is probably puzzling to the general reader that Mr. Chabris, who is working in cognitive neuroscience,
should tell us that his friends were surprised by his writing an article on intelligence. This is not surprising,
considering how g has been generally ignored by cognitive neuroscientists.
Entirely aside from the politically-correct prejudice that surrounds research on intelligence, there is another reason
why researchers in cognitive neuroscience have generally paid so little attention to the g factor: the failure to
recognize the important distinction between the essential design features of the brain (the main interest of
neuroscientists), which show distinct modules and specific localization of functions for various distinct cognitive
functions, on the one hand, and individual differences in the speed and efficiency with which these functions operate
(the main interest of psychometricians and differential psychologists), on the other. But I would argue that both of
these facets are proper subject matter for brain research. I suggest as a heuristic hypothesis that the design features
of the brain--its neural structures and functions--that are necessary for the many distinct processes that enter into
information-processing, or intelligence (such as attention, perception, discrimination, generalization, learning,
memory, language, thinking, problem-solving, and the like) are essentially the same for all biologically normal Homo
sapiens, i.e., those free of chromosomal and major gene anomalies or brain damage.
Correlated individual differences in the functioning of these various information processes are a result of other
quantitative biochemical and physiological conditions in the brain, most of them highly heritable, that are separate
from the brain's essential design features, or "hard-wiring," but are, as it were, super-imposed on all of them in
common, and affect the overall speed and efficiency of their functioning.
A methodology for discovering these physical conditions responsible for the existence of psychometric g is explained
in my recent
book, The g Factor, along with examples of individual variation in the several different quantitative anatomical and
physiological conditions that are found to cause various distinct cognitive functions to be correlated and hence to
result in the wide range of individual differences in general ability.
Burkittsville, Maryland
Christopher F. Chabris's pessimism about IQ's future may easily be right, but here is an alternative scenario. The key
is his point that "people often react most defensively when challenged ... on beliefs they wish were true but suspect
at some level to be false." I think the level in question is shifting closer and closer to the surface. The science of
the IQ controversy is too one-sided for serious scholars to be unshaken by it, no matter how passionate their
ideological preferences. And there is the matter of everyday experience: can anyone who spends his life in one of
America's leading universities doubt that Richard J. Herrnstein and I were correct about cognitive stratification?
Now assume--and it is a high-probability assumption--that in the next two decades behavioral genetics comes down on the
side of the classic psychometric view of intelligence. Arguments that IQ is not substantially heritable disappear.
Arguments that g is a statistical artifact disappear. A hodge-podge of politically-correct allegations about IQ are
discredited. Add this kind of evidence to the cognitive dissonance that already nags at many of the IQ critics, and
you have the stuff of an explosion in the received wisdom about intelligence.
Ironically, the genetic story will not have much effect on our state of knowledge about the relationship of IQ to
social and economic outcomes--the topic of The Bell Curve. But the psychological effect of the genetic story on members
of the academy is likely to be profound. I confess to a hankering to be around to see it.
I am gratified by the range of responses that "IQ Since The Bell Curve" stimulated, and I thank everyone who wrote. I
am especially pleased with the correspondence that avoids rehashing past debates and instead discusses methods and
trends that may characterize the future of intelligence research. I will say more about these matters later on, but let
me begin by addressing the mistakes I have been accused of making.
Despite Howard Gardner's protestation that none of his associates would deny measuring the "multiple intelligences" his
theory proposes, one of his co-authors in fact made this denial in my presence, as Richard J. McNally corroborates. Mr.
Gardner's influential theory has often been interpreted as claiming that seven or, we now learn, eight or nine domains
of basic cognitive abilities are independent of one another in all respects, and that there is thus no such thing as
general intelligence. Since this hypothesis contradicts mainstream theory so directly, it cries out to be tested. Mr.
Gardner's letter recognizes that need, and it is good to hear that his new research will be published "shortly." But
the fact that, fifteen years after the theory itself appeared, we have yet to see the empirical support for it
illustrates more plainly than any seminar discussion the reluctance of multiple-intelligence advocates to put their
ideas to the test.
In any event, I am troubled by the path suggested by Mr. Gardner's criteria for fair assessments of intelligences. Each
ability, he says, has "many facets," all of which need more than a "one-shot assessment" before an overall assessment
can be arrived at. But do these facets themselves have facets, just as intelligence itself has seven or more facets? If
too many facets of too many things must be measured too many times (per person!), the whole process becomes unworkable,
and the theory behind it ultimately loses its scientific value for lack of testability. True, sciences operate and
progress in many ways; but the most important way is by comparing theoretical predictions with empirical facts.
As Claude S. Fischer and others have pointed out, Charles Murray received his doctorate in political science, not in
sociology as I implied, and I apologize to all concerned. More pertinently, Mr. Fischer takes me to task for neglecting
two anti-Bell Curve works. Since one has not yet been published and an advance copy was not forthcoming from its
publisher, I can comment only on The Assault on Equality. In its outlook and in the thrust of its analysis, it is
similar to Mr. Fischer's own book; like the final paragraph of Mr. Fischer's letter, attempts to refute an exaggerated
form of almost every assertion of The Bell Curve, though in a more sarcastic, denigrating tone and with decidedly less
careful attention to accuracy. It should be read by anyone who needs an extended example of the "paroxysms of denial"
(the phrase is Arthur R. Jensen's) with which some critics have responded to IQ research.
In his letter, Mr. Fischer claims that I misrepresented the American Psychological Association's post-Bell Curve report
on intelligence. I am happy that he has chosen the APA report as his battleground since, as Ian J. Deary and his
colleagues relate, it does indeed endorse the scientific concept of general intelligence and the major methods used to
study it. Mr. Fischer, by contrast, recites several quotations from the report that he believes "undercut central
assertions" of Herrnstein and Murray.
Let me reply by saying that, first, it is true that "intelligence appears as only one of a great many factors that
influence social outcomes," and certainly this is a "main claim." But, just as certainly, it does not contradict The
Bell Curve, which did not, as Mr. Fischer seems to imply, say that intelligence was the only predictor of social
outcomes. A look at its many graphs comparing the effects of intelligence and socioeconomic status shows that each has
an effect, the issue being which effect is larger.
Second, it is also true that little of the total variation in social outcomes is accounted for by variation in
intelligence. Again, however, far from contradicting The Bell Curve, this part of the APA report actually cites the
book's own results in presenting its conclusions. Mr. Fischer implies that if a predictor--in this case,
intelligence--accounts for only a "small" proportion of the variation in an outcome (say, less than one quarter), it
must not be very important. But what is a reasonable baseline against which to compare the effect of intelligence? Or,
to put it another way, how large an effect should surprise us?
Given the myriad factors and circumstances that influence people's lives over the course of decades, it is remarkable
that scores on a single test should explain even 5 percent of their differences in job performance or criminal behavior,
not to mention the symptoms of post-traumatic stress disorder studied by Richard J. McNally or the many qualities
listed by Malcolm James Reel An effect much smaller than 5 percent was considered large enough to interrupt a major
biomedical study on the grounds that it would be unethical to continue denying to anyone the "tiny" benefit from the
medication involved; the benefit was a reduction in heart-attack risk and fatality, and the medication was aspirin.
Mr. Fischer's other points are similarly problematic. He hears The Bell Curve saying what he wants to hear, not what it
actually says, and so he mistakenly hears the APA report contradicting it. His own response, one might say, like that of
many others, has been to "declare victory and withdraw."
James R. Flynn, who noticed the Flynn effect but did not coin the term--in fact, Herrnstein and Murray coined it in The
Bell Curve--asks whether, given the large gains in IQ test scores over the past 80 years, we would consider most adults
of the World War I era to be mentally retarded by today's standards. It is a better question than he seems to think. An
average 1918 adult might indeed score very low on one of today's IQ tests; confronted via time-travel with today's
culture and technology, that same adult might be so bewildered and confused that full adjustment could never occur. But
in any case, such rhetorical arguments and thought-exercises cannot, by themselves, yield an explanation of what has
and has not changed about human cognition and the environments in which it occurs, and we need such an explanation
before we can understand the meaning of Mr. Flynn's "effect."
Jack Kaplan argues that intelligence tests do not measure intelligence because individuals with more or better
education will usually do better at answering the questions. (The same argument is made by Mr. Fischer and his
colleagues with reference to the particular test used in Herrnstein and Murray's analyses.) But this assumes,
ironically enough, that intelligence is a trait that cannot be affected by education. If we follow Mr. Kaplan's logic
to its conclusion, we shall soon be holding that intelligence must be determined completely by genetic and prenatal
factors--something it is safe to presume he does not believe.
Psychometric conceptions of intelligence commonly give roles to both "fluid" intelligence, loosely corresponding to
the ability to process information in novel ways, and "crystallized" intelligence, or the stored knowledge and methods
that assist in problem-solving. Both are associated with general intelligence, but they influence one another as well,
and develop differently across an individual's life span. Perhaps Mr. Kaplan considers only fluid intelligence to be
"true" intelligence, but it is not possible to separate "intellectual ability" completely from "intellectual
achievement." Even if we considered only measures of fluid intelligence as predictors of job performance and other
outcomes, the results would probably not differ greatly, since fluid intelligence is the strongest component of general
intelligence.
Kelly P. Ambrose stresses the importance in life and in society of factors other than intelligence. The point is
undeniable, and it is not denied by psychometric theorists, who are fully aware that whatever intelligence does not
predict must be explained by all sorts of other traits and contexts and their interactions, as well as by an almost
infinite succession of virtually random idiosyncratic circumstances that are beyond analysis. This does not mitigate
the fact that intelligence is often the single best predictor of who will succeed or fail in a given occupation or
situation.
Because I noted that people with certain characteristics do better than people without them, Jeff Hittenberger accuses
me of eugenicism. It is hard to imagine a more fallacious inference, but people who simply state empirical facts are
accused all the time of harboring noxious and veiled policy goals. Indeed, this subtle form of guilt by association can
even transcend generations. Thus, a recent letter to the New York Times from one Stephanie Olson claimed that "the
National Socialist party in Germany drew upon the work of American scientists like ... Dr. Arthur Jensen, who argued
for the genetic inferiority of minorities." As it happens, Jensen was ten years old when Hitler came to power.
Mr. Hittenberger might profit from rereading the paragraph in my original article that followed the one he quotes. As I
wrote there, "whenever intelligence is said, `race' is heard; whenever race is said, `genetics' is heard; and whenever
genetics is said, `inferiority' is heard--even though these issues are not necessarily connected in any way." To this I
can now add, regrettably, that whenever the subject of human differences is addressed, "eugenics" will be cried in
reply.
Ian J. Deary and Christopher Brand, though they inhabit the same city, might as well live on different planets.
Mr. Brand's planet, unfortunately, has a much less hospitable climate. Since his book was "de-published" and he was
fired from the faculty of the University of Edinburgh, he has been merely waiting--literally on tables. Mr. Deary is
rather more fortunate: in his world of science, where Mr. Brand used to live and where he did important work relating
mental-processing speed and other correlates to g, the enterprise of intelligence research continues apace.
Arthur R. Jensen and Charles Murray ask what will happen next in this field, but offer different answers. Lamenting
the
neglect of individual differences by cognitive neuroscientists, Mr. Jensen suggests that they are confused about the
distinction between those mechanisms of the brain that implement particular functions common to all human beings and
those properties that cut across such "modules" and give rise to overall speed and efficiency, or general ability. It
is true that cognitive neuroscientists are more interested in comparing people with monkeys and rats than with other
people, but I believe this tendency owes more to a simple lack of interest than to any misunderstanding on their part.
Cognitive neuroscientists are in fact uniquely equipped to write a crucial chapter in the behavioral-genetic story of
intelligence. Genes are linked to behavior by the brain, where the genes are expressed and the behavior is originated.
I alluded in my original article to the work of an international research team led by Robert Plomin of the Institute of
Psychiatry in London, which recently discovered that one variant of a specific gene is found more frequently in high-IQ
children than in average-IQ children. Thanks to the Human Genome Project, we also know that this same gene codes for a
protein known as "insulin-like growth factor-2 receptor." What we do not yet know is how this protein is involved in
the sequence of events that leads to mental functioning--in particular, which areas or properties of the brain are
influenced by the protein, how and when it influences them, and how they in turn influence cognition.
Discovering such multilevel chains of causation is beyond psychometrics, but it is the defining goal of cognitive
neuroscience. Until we do discover them, the relationship between genetics and behavior will remain as mysterious as
the eerie similarities between identical twins that are traditionally used to document it.
Ironically, it may turn out that to develop a cognitive understanding of the genetics of intelligence, neuroscientists
will depend on the very individual differences they have heretofore ignored. As the Plomin group's research shows,
linking specific genes to specific brain functions requires finding people who differ in each and measuring the
associations among those differences. A reconciliation of the psychometric and cognitive approaches to intelligence
would be a wonderful legacy of the welcome entry of genetics into psychology.
In "IQ Since The Bell Curve," I predicted more phony controversies even as our understanding of human cognitive
abilities expands in the coming years. Charles Murray disagrees, forecasting a long-awaited and sudden acceptance by
academics and the public of the overwhelming evidence for the reality and heritability of intelligence. Such an outcome
would, in a sense, vindicate Richard J. Herrnstein and himself, and we can only hope things turn out that way.
COPYRIGHT 1998 American Jewish Committee
COPYRIGHT 2000 Gale Group
Commentary Magazine