Charles Murray’s Human Diversity

I pre-ordered Charles Murray’s latest book, Human Diversity, in the naive hope that I might finish it before the outrage machine masquerading as our modern press could calumniate it. It arrived on my doorstep the very day it was available in bookstores, and no sooner had I turned the last page than the New York Times, true to form, put forth a review of a book unrecognizable to the one I read. Where the book I read had been a careful exploration of the modern science of human biology, drawing predominantly on advances in genomics and neuroscience to make a series of middle-of-the-road claims about the present scientific consensus, the NYT critic found a careless bigot eager to play fast and loose with statistics. Whereas Murray, at every turn, insisted on the moral and factual impossibility of generalizing from our knowledge about people to our knowledge of any given individual, the NYT critic saw him defending “humanity’s most pernicious, wearying and stubborn stereotypes.” Readers of the book, I’m afraid, will find nothing so shocking, but that is precisely why this review, and the many like it sure to follow, are written in such alarmist terms: they do not want people reading this book. Murray well knows that if the information he presents is hum-drum by the standards of biology and neuroscience departments across the modern world, it is nonetheless anathema to the constellation of humanities and social science and anthropology departments, who have erected an orthodoxy around the subjects of race, class and gender and policed its boundaries with the enthusiasm of inquisitors. (In 2017, for example, when Murray was scheduled to give a presentation on his book Coming Apart – still, I maintain, the single best book to understand the appeal of Donald Trump to millions of Americans – at Middlebury College, both he and the professor responsible for presenting him were violently driven off the campus by protestors.)

What, then, are this book’s central claims? Murray lays them out for us at the very start, in a series of ten “propositions” that I reproduce here:

  1. Sex differences in personality are consistent worldwide and tend to widen in more gender-egalitarian cultures.
  2. On average, females worldwide have advantages in verbal ability and social cognition while males have advantages in visuospatial abilities and the extremes of mathematical ability.
  3. On average, women worldwide are more attracted to vocations centered on people and men to vocations centered on things.
  4. Many sex differences in the brain are coordinate with sex differences in personality, abilities, and social behavior.
  5. Human populations are genetically distinctive in ways that correspond to self-identified race and ethnicity.
  6. Evolutionary selection pressure since humans left Africa has been extensive and mostly local.
  7. Continental population differences in variants associated with personality, abilities, and social behavior are common.
  8. The shared environment usually plays a minor role in explaining personality, abilities, and social behavior.
  9. Class structure is importantly based on differences in abilities that have a substantial genetic component.
  10. Outside interventions are inherently constrained in the effects they can have on personality, abilities, and social behavior.

Broadly speaking, propositions 1-4 deal with biological differences in sex, 5-7 with biological differences in race or ethnicity, and 8-10 with biological differences in class. They are direct refutations of three central tenets of modern progressive thought, which is precisely why they inspire such ire: that “gender and race are social constructs,” or that “class is a function of privilege.” Note, however, that Murray does not contend that there do not exist social constructions around these ideas; he does not argue, for example, that there are not racist or sexist tropes, or even benign stereotypes of male and female behavior. What he is seeking to refute – or rather, what he claims the latest developments in neuroscience and genomics refute – is the popular conception of these categories as infinitely malleable, “constructed” merely from our cultural understanding of them. Neuroscience, as a discipline, is extremely young, but every decade brings quantum leaps in our understanding of human cognition and brain function; genomics is an even younger field, the human genome having only been fully sequenced in 2003, but its databases of genome samples grow larger by the year. I cannot hope to synthesize the arguments of this 600-plus page book in a single review, but for the sake of demonstrating Murray’s approach I will summarize some of his discussions of the cognitive and personality differences between men and women.

Let’s begin on relatively firm ground: using the two largest, most widely used models of human personality characteristics developed by psychologists, the Sixteen Personality Factor Questionnaire (16PF) and the Five Factor Model (FFM, or “Big Five”), cross-cultural patterns emerge in the personalities of the average man and the average women. The differences in personality range from the trivial to the consequential, with some of the largest personality differences corresponding to our common understanding of sex differences: men, for example, were “less open to inner feelings and emotions,” “less modest in playing down their achievements,” and less easily upset by stress and anxiety. Women scored higher than men (on average!) in the categories of warmth, attentiveness to others, sensitivity, sentimentality and aesthetic appreciation, and were more likely to be “cooperating, accommodating, and deferential.” The differences are all the more striking when they are aggregated, which is one of the reasons some social scientists, eager to downplay sex differences in personality, often seek to present the data in its disaggregated form, enabling them to claim (truthfully) that, yes, there are personality differences between men and women, but they are trivial, and pale in comparison to greater variation between individuals (ie an extroverted woman and an extroverted man are more similar than an extroverted woman and an introverted man). But this presentation of the data, Murry argues, obscures more than it enlightens, for human beings do not interact with disaggregated personalities; we engage with people at multiple levels of their personality. By way of analogy, he offers a clever example: consider the below picture of two faces.

Most people automatically sort these two faces into the categories female and male, and they do so with astonishing accuracy and rapidity. But if you restrict your gaze to any single feature – the eyes, say, or the nose, or the mouth – could you reliably predict which gender that feature belongs to? Disaggregated, the features are relatively indistinct; aggregated, clear differences emerge – the same, Murray argues, is true with male-female personality differences. Murray references a recent meta-analysis conducted by psychologists Tim Kaiser, Marco del Giudice and Tom Booth, with a sample size approaching 50,000 people, which aggregates the Mahalanobis distance (a measure of the distance between two points on a distribution, in this case the median male and female scores, by standard deviation) of the various male and female scores, arriving at a score of D = 2.71, “a huge difference that would leave only 10 percent overlap between two normal distributions.” In an article published late last year in Scientific American, analyzing the very same meta-analysis, the author summarized the findings this way: “their data suggests that the probability that a randomly picked individual will be correctly classified as male or female based on knowledge of their global personality profile is 85%.” In other words, the aggregate of the male-female personality differences looks very much like the two portraits referenced above: similar, to be sure, but noticeably different as well.

Very well, the social constructionists might say, there are observable, quantifiable personality differences between men and women, but this discovery says nothing about the provenance of those differences. We simply raise women to be warmer and more nurturing, and men to be more aggressive and risk-taking. Alas for this theory, the bad news continues: when psychologists cross-referenced their global data on male-female personality differences, they were shocked to discover that those same differences maximized in countries with greater equality, and were minimized (though still significant) in less egalitarian, less wealthy societies. Murray cites no less than five studies conducted on this topic, from as late as 2001 and as recent as 2018, all of which coalesce on the same basic conclusion: greater wealth and greater legal and social opportunities “exhibited an independent and significant association with gender differences in preferences.” The hypothesis of the social constructionists – that pervasive sexism creates the stereotypical male-female personality differences – was not only proved wrong, but backwards: the more opportunities men and women are given, the more distinct they become.

In the chapter “Sex Differences in the Brain,” Murray seeks to prove proposition #4, that “sex differences in the brain are coordinate with sex differences in personality, abilities, and social behavior.” One of the abilities that average men and women differ in is in their proficiency at math and language: the average man scores higher on mathematics tests, as well as on tests of “visuospatial” ability (mentally manipulating objects), whereas the average woman scores higher on tests relating to language. These results are consistent across cultures, ethnicities, and socioeconomic brackets. As our neuroimaging technology continues to improve, and with it our understanding of the anatomy of the human brain, the reasons for these differences comes into focus. The human brain is divided into two cerebral hemispheres separated by a visible fissure. Brain scans, conducted over time to monitor neuronal activity, have associated the left hemisphere with language production and the right hemisphere with visuospatial activity, but they also uncovered a striking difference in male-female cognition: “males primarily use the left hemisphere for verbal tasks and the right hemisphere for spatial tasks, whereas women use both hemispheres for both types of tasks,” a finding described as functional asymmetry. The working hypothesis, heavily supported by a 1996 meta-analysis of brain imaging studies, is that men’s greater lateralization, prompted by the very same neonatal testosterone exposure that transforms the default female fetus into a male, explains both their decreased verbal abilities and their increased math abilities:

Recall that the default brain is female. In the case of language processing, this means that the default is to use both hemispheres. The salient issue in analyzing the effects of testosterone on language processing is that something about the development of the male right brain is crowding out the use of the right hemisphere for language processing (which, by default, would ordinarily be used). The focus of the right hemisphere on spatial processing, driven by the impact of testosterone on the right hemisphere, is a plausible explanation.

This theory was given a significant boost when studies of brain damage were incorporated into the data: “Women’s language test scores after brain damage suffer the same effect whether the damage occurred in the left or right hemisphere, whereas men are more affected by damage to the left hemisphere.” Another piece of the puzzle was provided by the modern confirmation, via MRI, of an anatomical difference described as long ago as 1982: that the corpus callosum, “a flat, ribbonlike bundle of fibers about four inches long that lies at the bottom of the fissure between the two hemispheres” acting as a kind of physical bridge between the brain regions, is slightly larger in women than in men, when controlling for brain size.

If my summary seems exhaustive, know that it is nonetheless only a sampling of the arguments put forward in a single section of this book, which includes extensive Appendix material, as well as nearly two hundred pages of notes and references, many of them including helpful commentary from Murray about which studies have been commonly replicated and which are newer or more speculative. The end result is a book that can only have been produced after years of labour and careful study, but that is written and presented in such a manner that a total layman – someone with no working knowledge of genomics or brain anatomy or even the kind of statistical skill needed to parse some of the more complex meta-analyses – can nonetheless follow Murray’s argument with profit and pleasure. And so it irks me, to no small degree, to see a New York Times reviewer with no relevant scientific background produce a careless, mendacious and mean-spirited review, with no more thought or care than you or I might compose an email. But I am cheered up by Murray’s own optimism that we are on the verge of a great haul, produced by the astonishing new frontiers opened up to us by genomics and neuroscience. “We are like physicists at the outset of the nineteenth century,” Murray tells us, “who were poised at a moment in history that would produce Ampères and Faradays.” More hit pieces will follow, and Murray’s reputation, at least in the popular press, will not likely recover in his lifetime, but I am increasingly confident that history will vindicate his approach.