|Galton Institute Home Page||June 1996 Newsletter Contents||Newsletter Index|
Review: The g Factor: The Science of Mental Ability. Jensen, Arthur Robert. Praeger, London, 1998. Pp 648. Price $39.95.
“The ground fact is that every science, in order to deserve the name, must necessarily be built up on some pregnant system of principles.”
Charles Spearman, The Nature of ‘Intelligence’ and the Principles of Cognition (1923).
1999 is the thirtieth anniversary of the publication, in the Harvard Educational Review (henceforth HER), of “How Much Can We Boost IQ and Scholastic Achievement?”. The refutations that rained down on its author, Arthur Jensen, are now gone, “sunk, … with the tumult [that] they made” (Carlyle). Although estimates of heritability* vary, behavioural geneticists subsequently upheld Jensen’s claim that IQ is highly heritable. Thomas Bouchard’s work at Minnesota on monozygotic twins raised apart was a milestone in this context. Bouchard was one of over fifty professors who signed a declaration on intelligence published in the Wall Street Journal on December 13th, 1994, which claimed that “… genetics plays a bigger role than does environment in creating IQ differences among individuals”.
Jensen’s HER article is a still pertinent critique of the false assumptions which have driven compensatory education. Supporters of such programmes, it seems, lost sight of the distinction between two contrasting sources of individual differences, namely, heredity and society. For as Rousseau observes, there is an artificial inequality established by social convention, “inégalité morale, ou politique”, consisting of the gradations in power and prestige between different members of society. But there is also a natural inequality contingent on the inborn mental differences between people, the variation in what Rousseau calls the “qualités de l’esprit, ou de l’âme” (Discours Sur L’Origine Et Les Fondements De L’Inégalité Parmi Les Hommes, 1754).
In fields of endeavour where reality testing prevails, practitioners are prepared to reconsider the basic theories guiding their efforts, whenever this is necessary. Timely opportunities for such reconsideration arise when “bridges do not stand, when aircraft do not fly, when machines do not work, when treatments do not cure …” (Jensen, HER). The working assumption in compensatory education has been that most children are born with the same basic capabilities. Yet despite the manifest failure of compensatory education, its supporters have steadfastly ignored the possible contribution of genetic factors to individual differences. Contra Rousseau, differences in educational performance have been unwaveringly attributed to environmental factors. As a consequence, huge sums of public money have been spent in the United States on ultimately Sisyphean special programmes to try and raise mental ability. The cost of the Milwaukee Project (1966-1973) was $14 million or a staggering $23,000 per IQ point raised per child. The annual expenditure of Head Start, an ongoing federal pre-school intervention set up in 1964, is $2 billion.
According to Charles Spearman, a general mental ability, g, enters into every complex cognitive task necessitating an eduction of relations or of correlates. Clearly, g is remarkably resistant to attempts to force feed it. As the aforementioned Wall Street Journal declaration conceded, “… we do not know how to manipulate it [the environment] to raise low IQs permanently”. Any improvements in test scores arising from special educational interventions have generally proved too ephemeral and circumscribed (ie not spread across a range of tests). In the case of the Milwaukee Project, improvements in scores, rather than representing any underlying increase in g, were the result of what is graphically called “teaching to the test”. The determined efforts to promote equality of opportunity in the American school system (by bussing etc) have likewise failed to produce equality of educational outcome. As Jensen has remarked, g is “the one individual pupil characteristic over which schools can have little or no control …” (“Spearman’s g and the Problem of Educational Equality”, 1991). Yet is this relative non-plasticity of g so surprising? If intelligence is a “psychobiological variable”, as Jensen maintains, it must have a biological substrate.
Charles Spearman was the first psychologist to demonstrate the existence of g empirically. Its existence can be inferred, he argued, from the positive correlations between measures of various cognitive abilities, as revealed by factor analysis, which was itself invented by Spearman. In The Abilities Of Man (1927), Spearman observed that if a particular test is correlated with other operations involving g, this test must also elicit g.
In his latest book, The g Factor, Professor Jensen reminds us that both Francis Galton and Herbert Spencer also suspected that intelligence is a general or unitary ability. Yet prior to the publication of Spearman’s paper “General intelligence, objectively determined and measured” (1904), the attempts to ascertain the possible correlations between different cognitive abilities had been vitiated by the problem of attenuation. In this historic paper, Spearman proposed, inter alia, a formula for correcting the attenuating effect of random errors upon any correlation. Jensen evidently shares Spearman’s opinion that the discovery of g constituted “a Copernican revolution” in psychology.
The sub-title of The g Factor, to wit, The Science Of Mental Ability, is well chosen. For it was the fervent wish of Spearman, to whom the book is dedicated, that individual differences in mental ability would eventually be explained in terms of evolutionary biology and differences in properties of the brain. Jensen shares this ambition to give psychology “a genuinely scientific foundation” (Spearman). To this end, he (Jensen) has zealously encouraged links between psychometrics and other disciplines, notably evolutionary biology, evolutionary psychology, behavioural and population genetics and neuroscience. The g Factor presents a wealth of fascinating data, in this context, on the biological correlates of g. It is apparently correlated with individual differences in the following; a) brain size, (taking the head as a proxy for the brain); b) cerebral glucose metabolism (inversely correlated with IQ); c) the frequency and amplitude of spontaneous brain waves; d) the average evoked potential in response to an electrical stimulus; e) nerve conduction velocity. Since the publication of The g Factor, Professor Jensen has obtained a database containing evidence of an inverse correlation between g and both high and low levels of testosterone.
In a recent interview, Jensen describes g as “a property of the information-processing capacities of the brain …”. He speculates in The g Factor that speed of neural activity may be the key to individual differences in mental ability. This would explain the correlation of IQ with reaction times (RTs) to elementary cognitive tasks. Further research on individual differences in the processing speed of working memory might be fruitful, he suggests. Such differences may be linked, in his view, to the degree of myelination of neurones in the brain. It should be recalled that Spearman regarded g as a measure of a person’s mental energy (using the term energy metaphorically).
Richard Lynn says The g Factor is “a major contribution to our understanding of human intelligence” (notice, however, that Professor Lynn does not accept Jensen’s view that the concept of intelligence is, in Spearman’s words, “cankered with equivocality”, and ought therefore to be abandoned by scientists). The distillate of a lifetime’s research, it is indisputably Jensen’s magnum opus. Thus far, however, it has hardly received the critical attention which it merits. Certain publishers even declined to take on this book. And some influential journals, notably the Times Higher Educational Supplement, refuse to review it. These days, unwelcome books are no longer burnt. They are simply ignored. Professor Jensen is to be spared no ignominy by his intellectual enemies, it would seem.
It is, of course, Jensen’s position on race differences in intelligence which makes him beyond the pale for certain editors and publishers. As J Phillipe Rushton (who clearly is not given to understatement) has declared, “the evolutionary psychology of race differences has become the most politically incorrect topic in the world today” (Race, Evolution, and Behaviour). The hypothesis that the differences in mental ability between racial groups (as demonstrated by IQ tests), are in part determined by genetic factors, was first broached by Jensen in his HER article. This violation of the race taboo immediately turned him into a hate figure for the left, one of differential psychology’s swelling band of martyrs, whose latest member is the hapless Saint Christopher Brand.
Yet since Jensen has been vindicated by his fellow scientists on the heritability of individual differences in mental ability, he is understandably disinclined to tailor his position on that of group differences. He remains a recusant on this question, having perhaps concluded, like Macbeth, that he is “… in blood stepped in so far, that, should [he] wade no more, Returning were as tedious as go o’er”.
Chapters eleven and twelve of The g Factor, entitled “Population Differences in g”, and “Causal Hypotheses”, respectively, are the book’s core. The exegesis of the properties and correlates of g in the earlier chapters is a prelude to a painstaking demonstration, in these two lengthy chapters, of the patent inadequacies of the non-genetic explanations of the white-black (W-B) IQ difference. These explanations are generally founded on tenuous sociological notions such as alienation, caste status, culture of poverty, low self-esteem, peer culture and white racism. The premise of all these theories is that IQ differences between groups are cultural not genetic differences. Jensen calls this assertion of the dissimilarity of within group and between group differences the “dual hypothesis”. In opposition to it, he proposes his “default hypothesis”. The latter states that the W-B IQ difference is determined by the same environmental and genetic factors which account for within group variation, and in approximately the same ratios.
One exponent of the dual hypothesis singled out for criticism by Jensen is the sociologist and economist Thomas Sowell. Sowell claims that the “Flynn effect” (ie secular increase in IQ scores throughout the world) proves that the W-B IQ difference cannot be genetically determined. For if inter-group mental test differences are primarily determined by genetic factors, how, Sowell demands, could national IQ could have risen over time in the USA, given the greater fertility of social and ethnic groups with lower average IQ?
In Race And Culture (1994), Sowell proposes that the facility for handling abstract thought is itself a cultural variable and that in groups which lack this facility, the motivation to perform abstract tasks may also be depleted. He maintains that a paucity of abstract, intellectual interests is a distinguishing feature of the black American sub-culture. This helps explain why white-black test score differences are greatest on abstract (culture reduced) questions, in his view. And why American Negro soldiers given IQ tests during World War I apparently tended to “lapse into inattention and almost into sleep”.
The mean IQ of black Americans is approximately 85. Sowell predicts that this black IQ deficit will eventually prove to be a transitional phenomenon. He takes comfort from the fact that American Jews, Poles and Italians also once performed badly on IQ tests. The subsequent improvement in these groups’ mental performance suggests that environmental causes generally determine such group deficits, in his estimation. For Sowell, further cultural assimilation of black Americans indicates the direction from which an improvement in black IQ scores and educational achievement can come. He holds up American Jews and Asian Americans, in particular, as minorities worthy of emulation because of their positive attitudes towards education, saving, work and the family.
However, none of the sociological theories deployed to explain away the W-B IQ difference, including Sowell’s, have disposed of the evidence that genetic factors play a significant role in determining this difference. The following telling points, made by Jensen in The g Factor, represent a summary of this evidence;
1) The W-B IQ difference is greater when subjects are asked to repeat numbers in reverse order (backward digit span), compared to normal order (forward digit span). Jensen regards this is an illustration of “Spearman’s hypothesis” that the size of the W-B IQ difference is proportionate to a test’s difficulty or g loading.
2) Black IQ scores are validated by their correlation with independent measures of educational achievement and job performance. In other words, there is no psychometric evidence that IQ tests are culturally biased against Afro-Americans.
3) Differences in reaction times to elementary cognitive tasks cannot be determined by cultural variables or motivational factors. Blacks, on average, have slower reaction times than whites.
4) The W-B IQ difference has remained unchanged for eighty years, a period of significant environmental improvements for blacks.
5) Recent Asian immigrants, on average, have higher IQ scores than US blacks.
6) The mean IQ difference between black and white children increases at each successively higher level of parental socio-economic status. This phenomenon is only explicable in terms of the genetic law of filial regression to mediocrity, enunciated by Galton. Similarly, when blacks and whites were matched by Jensen for an IQ of 120, their siblings average IQ was found to be 113 for whites and 99 for blacks.
7) There is a pronounced dysgenic trend within the Afro-American community, due to a differential birth rate. This problem is possibly compounded by inbreeding depression, contingent on race differences in sexual behaviour (Jensen, personal communication with the present writer). This dysgenic trend can only enlarge the W-B IQ difference.
8) The Minnesota Transracial Adoption Study demonstrated that adoption of black children by white, middle class families does not lead to significant long-term IQ gains.
9) The W-B IQ difference is not simply a difference in socio-economic status. For as Sowell himself concedes, “Asian American students from families with incomes of $6,000 or less scored higher on the mathematics portion of the SAT [Scholastic Aptitude Test] than black … American students from families with income of $50,000 or more”.
10) Blacks in Africa who have never been subjected to slavery have lower average IQ than Afro-Americans who have. This finding contradicts anthropologist John Ogbu’s ingenious theory that the lower average IQ of Afro-Americans compared to whites is a concomitant of their caste status.
A Darwinian long before the leftist intelligentsia substituted a politically correct type of Darwinism for defunct Marxism, Jensen believes that race differences in behaviour, including differences in cognitive ability, must have an evolutionary origin. For human races, considered as relatively isolated breeding populations, have inhabited different geographical locations. They have therefore been subjected to varying degrees and forms of selection pressure. It seems noteworthy, in this context, that paleoanthropologists such as Chris Stringer have emphasised the increases in brain size and intelligence during the evolution of the hominids which preceded anatomically modern humans. Stringer also accepts that climate has been the main factor determining the size of the human cranium. For a smaller, dolichocephalic cranium is, he says, more adaptive in a hot climate, and a larger, brachycephalic one is more adaptive in a cold climate. Jensen’s working definition of the term race, to wit, populations having “differences in the frequencies of different alleles at many gene loci ” (page 424), recalls that of the population geneticist Theodosius Dobzhansky, in Heredity And the Nature Of Man.
The g Factor confirms the impression of The Bell Curve that social problems appear alarmingly intractable, viewed from a hereditarian perspective. Jensen notes, for example, that a small but not insignificant proportion of the members of our increasingly information intensive society lack the mental ability to fill in a form or operate a video recorder. The social exclusion contingent on such low cognitive ability represents part of the downside of a meritocracy.
A similarly fatalistic impression is made by Jensen’s observations on ethnic group differences in crime rates. These highly conspicuous differences, he suggests, are connected with the different proportions of various groups falling into the high risk, lower IQ stratum. Put differently, bell curves differ across racial groups. Thus, 60% of US blacks have an IQ below 90, and the peak crime rate occurs in the IQ range 75 to 90. The significance of g in relation to other key social problems, notably those relating to education, employment, health and welfare is likewise underscored. Concerning education, in the Unites States, an IQ below 70 is categorised as mentally retarded. 15.95% of blacks compared to only 2.3% of whites fall into this category.
Professor Jensen is an incisive critic of affirmative action, which he re-defines as “institutionalised discrimination”. He considers that this policy of racial preferences has been an unqualified disaster, inimical to economic efficiency, educational progress, social justice and harmonious race relations. As a method of promoting people beyond their level of ability, however, he concedes that affirmative action has been an outstanding success. For as Linda Gottfredson has demonstrated, blacks in the US are now numerically over-represented in the higher, “g demanding” professions in relation to the B% W% ratio in the IQ recruitment range of these occupations. In “Spearman’s g and the Problem of Educational Equality”, Jensen has eloquently described the uniformly destructive effects of the lowering of selection standards for certain groups in order to achieve racial parity in college admissions. He witnessed these baneful effects first hand at Berkeley, where the college rules were changed so that students from particular ethnic minorities could not fail courses.
Psychometrics is now the dismal science. Jensen shares Richard Lynn’s concern that there is an underlying dysgenic trend, throughout the advanced countries, leading to a decline in genotypic intelligence. They contend that this decline has been masked by an increase in phenotypic intelligence, due to both non-genetic biological factors (in particular, better nutrition and health care) and some genetic factors, such as a reduction in inbreeding. This is the explanation, both maintain, for the “Flynn effect”, which represents a real increase in g. They predict that once the quality of nutrition reaches the optimum, phenotypic intelligence will then begin to fall, if dysgenic fertility persists. Professor Lynn, it will be recalled, addressed this all important issue in Dysgenics. (Newsletter issue 25 June 1997.)
The g Factor does not exaggerate the social significance of cognitive ability, however. A sufficient level of g is deemed necessary but not sufficient for individual success. Across the range of occupations, which demand different levels of mental ability, g is characterised as a threshold variable. The importance of other personality factors is acknowledged. As Spearman himself pointed out, students can do very well at mental tests but not at university (but not vice versa!) because of the influence of effort or zeal on success.
A R Jensen has been called the foremost living representative of the Galtonian tradition in psychology, that “unique amalgam of evolutionary biology, behavioural genetics, psychometrics and neuroscience” (Rushton). Burt, Cattell, Eysenck, Galton and Spearman were previous distinguished representatives of this “London school”. He (Jensen) is disapproved of in certain influential circles, however, where the bizarre complaint is made that he pursues lines of inquiry which “worry people”. In these benighted circles, it is not acceptable to treat race differences scientifically. For by so doing, we might undermine the egalitarian assumption that homo sapiens is exempt from the universal laws of variation, selection and speciation.
Mankind’s perverse tendency to persecute its educators (or eductors!) was bewailed by Thomas Huxley, who concluded that there was therefore no sadder spectacle than the history of humanity. It is earnestly to be hoped that Professor Jensen, who is this year’s Galton Lecturer, will ultimately prevail over the Philistines who are trying to silence him again.
*Heritability is a technical term in genetics. It means the proportion of phenotypic variance due to variance in genotypes. It is estimated by measuring various kinship correlations. The heritability of IQ increases with age. In his HER article, Jensen suggested that the heritability of intelligence might be as high as 80%.