Human Diversity is Charles Murray’s attempt to save America (if not the entire Western liberal order).
Everything we’re allowed to say about race, sex, & class is wrong, & everybody knows it. The systems built around the post-1960s progressive consensus have failed, & are now collapsing under the weight of their own contradictions - but the mechanisms that enforce that consensus are stronger than ever. So with Human Diversity, Murray is wielding his tenure as a shield, & trying to reopen the conversation for the rest of us, before it imposes itself catastrophically.
(If there’s any hope for liberalism at all, I think this is it. It’s not enough to point out that the progs are wrong & their dumb world is burning down around them - the joy has gone out of shitposting because it’s stating the obvious at this point. Nobody has a compelling alternative vision - no one has succeeded in integrating these dangerous edgy facts that we know into a normal worldview. Claiming to believe that Adolf Hitler was the last blood of Atlantis & he’s in suspended animation under the Antarctic ice is a way of daring the reader to ask you a serious question about the ramifications of your beliefs so you can call him a g*y ret*rd. And that can be hugely enjoyable on Twitter, but eventually I’ve got to figure out what I’m going to tell my kids.)
So, Murray is almost painfully patient & cautious & restrained & precise. Human Diversity is written to be an archive of what is indisputable, & he’s only willing to cop to his conclusions when the data makes them dead obvious (& sometimes not even then.)
Murray’s central theses are 1) that social scientists know a lot of things about race, sex, & class that they have been pretending not to know, & 2) that what they know isn’t all that scary, & we should be less nervous talking about it. As you might imagine, the first argument is pretty straightforward; the second is a lot more complicated.
He has the easiest job of reassuring the reader on the question of sex. His selection of studies suggests almost cosmic symmetry & complementarity - men & women really do appear to be “not better or worse, just different”. There are some differences in aptitude, but these matter most in the extremes, & are drowned out by much more dramatic differences in preference. Women like flexibility, job security, work-life balance, & working with people; men like risk, money, & working with systems.
His implied policy recommendation is to let everybody do what they want, & then respect their revealed preferences. Which makes it pretty hard to get mad at him, so he has liberty to explore a couple fun “women be shopping” takes (more on this below). Again, the point is to assure you that you can believe your lying eyes - the stuff you know from intuitive observation of men & women is basically right, & there’s a way to approach those truths without being an asshole.
The section on race is way, way less specific: Murray makes no attempt to compare individual racial groups on any cognitive or behavioral axis whatsoever. Instead, he makes a very limited argument that race is an objective phenomenon, & that cognitive & behavioral differences between races are virtually certain, given some pretty deep-in-the-weeds (but apparently uncontroversial) findings in population genetics.
He’s totally silent on what those cognitive & behavioral differences might be - because, he says, race isn’t as thoroughly studied as gender (he argues the state of the science is 15-20 years behind) - but I suspect it’s at least partly because the results don’t show the same elegant gameplay balance that he finds for sex. So his uncertainty is not so much empirical as moral: how do you hold these facts in your mind without being an asshole. (This is not a trivial problem!)
He goes as far as to say [paraphrasing], “there are definitely racial differences in behavior & cognition, & DIE activists are already getting upset about mental health treatments designed with White Brains in mind. So somebody, somewhere is going to have to talk about this soon - but not me, not in this book”.
The section on class feels like a review of The Bell Curve. IQ is strongly predictive of cognitive ability & socioeconomic success, & most of the egalitarian success of public education in the last 200 years has come from identifying & elevating high-IQ poor people, rather than improving anyone’s native ability. Now that smart people are relatively easy to find (& have found each other), education reinforces class divisions rather than making them more porous.
Again, the subtext as he’s rattling off all these regression models is a moral wrestle: Murray’s pretty sure you can’t change IQ, period - & there seems to be a lot of other cognitive/behavioral traits you can’t change, either. So how do we reward smart & high-functioning people for talents that society needs, while recognizing that those talents are an unearned gift? And how do you think about people on the back half of the curve, who can never be taught to contribute in a ways that society finds really useful? How can you permanently foreclose that hope without being an asshole? (Like I said, I think he has more to say about sex differences because it’s easier to make the optimistic case there.)
Murray’s certitude about the role of IQ and class makes his shrugging about race differences look a little disingenuous - but he has to survive office Christmas parties like everybody else, & I don’t really blame him. (UPDATE: it looks like he was saving up material for the book he released two weeks ago, entitled Facing Reality. Maybe he figured out a moral approach. Will be reading that next.)
Anyway it’s a great book - strongly recommend you read it for yourself. I put together the notes & reactions below as I let the book wash over me on a road trip, but this isn’t nearly all the good stuff - the weeds of it are very interesting & accessible, especially in the gender section.
Sex
Women answer some spatial/math questions with comparable accuracy, but they take longer to get there. I wonder how much this is determined by neuroticism, rather than actual processing speed. I know women who take a long time to differentiate left from right - not because this is cognitively demanding, but because they go through 2 or 3 layers of second-guessing each time. (“ok my left thumb & forefinger make an ‘L’, for left - wait, is that what an ‘L’ looks like? Shit”)
Murray notes that egalitarian countries are more sex-segregated professionally, not less - which he takes to mean that sex differences are innate, & women in sexist societies aren’t free to realize their feminine nature professionally. But he also notes that the widened gap is mostly the result of men breaking away from baseline, both physically & mentally, so that doesn’t make sense. My guess is the causality works the other way: Gender equality doesn’t make men prefer male jobs - rather, you need endemic male-brained profit-maximizing autismo to walk a society blindly into gender equality. That would be funnier, anyway.
Men & women with comparable IQs are deceptively different, because they use different toolboxes to get the same answers. For example, women react emotionally & record memory in the same hemisphere of the amygdala, while in men the two are segregated - which may explain why men have an easier time with traumatic memory. This also potentially explains why women in a fight will recall things you did ten years ago: the memory & the negative affect are much more tightly linked.
He argues that empathy (which women excel at) is distinct from systematizing intelligence (which men excel at), which is odd to me. Isn’t empathy just the ability to simulate a complex system (another person’s environment & emotional state), & predict the outputs of that system? “How would I feel if I were in their shoes?” Maybe this is something that I do in a very male-brained way, & it’s routed totally differently for women.
Men experience novel positive stimuli much more intensely than women, but quickly become habituated, while the reverse is true of women - negative stimuli are stronger, & tend to linger. There’s a pretty easy evolutionary just-so story here - primitive males needed to chase novelty & take risks in order to succeed, while females mostly needed to not fail - don’t alienate the tribe, don’t feed your baby anything poisonous, don’t get eaten. Also explains contemporary stereotypes of male/female curiosity, risk-taking, novelty-seeking, addiction.
Men are much higher-variance than women, across the board - both from one man to another, & among the competencies in an individual man. (A very smart woman is more likely to be reasonably competent at everything, while a very smart man is more likely to have some real deficits.) Murray suggests this is because the doubled X chromosome imparts some redundancy & consistency, whereas if something goes wrong on a man’s X, there’s no backup. Saying that the very smartest men are probably smarter (albeit weirder) than the very smartest women is about as spicy as Murray gets in the whole book.
Murray implies that the West has pretty much ended sex discrimination: at this point, if men dominate some fields & women others, this is largely down to innate ability & preference. But he claims that women should excel in people-focused roles such as CEOs, lawyers, politicians, directors, but those are still very heavily male fields. You’ve got to believe either that gender discrimination is still very strong in these fields, or that the difference between men & women is a lot more complicated than just “people versus things”.
Murray says that high-IQ women are “more resistant to anti-feminist propaganda”, which ignores the absolute universal triumph of pro-feminist propaganda in the West. But “brainwashing” in either direction is probably the wrong way to look at this. If your society affords higher social status to professional work than family work, you’ll probably find professional work more attractive, even if you think your society’s assessment is wrong. This is especially true if you’re agreeable & concerned with social cohesion, as women tend to be. The low social status of homemaking is an actual social reality that has consequences for women who choose it - it’s not a fiction installed in individual women’s minds.
Anons like to point out that there has been no feminist rush into sanitation or construction work, which according to Murray is true: in low-class professions, there basically hasn’t been any sexual revolution at all. When women face a choice between low-class female work or low-class male work, they overwhelmingly choose low-class female work. This suggests that for the vast majority of women in high-class male-dominated fields, it’s not a question of having been “liberated” to do what they innately want to do - it’s just that the status calculus for high-class women has changed, & they have followed the incentives.
Murray is very enthusiastic about the closing of the gender gap thus far, & the unlocking of female potential in the professions - but he seems to miss the loss of what was once accomplished by women outside the professional sphere. He acknowledges that money & career isn’t everything - but to him, this seems to mean that it’s ok for a woman to work 40 hours a week instead of 70, or take a 6-month sabbatical with her baby before starting daycare.
Race
Machine learning algorithms, naively examining unlabeled genetic data, very easily find the same discrete racial groups that humans do. The method, called “K-means clustering”, involves showing the computer some raw data, & telling it to identify clusters that naturally appear in the data. As you ask it to identify more & more clusters, you can identify a tree of relationships within the data. (See below.) For human genetic data, the boundaries are very crisp & easy to find. Murray makes a very detailed argument for exactly why this is the case (rather than, say, a smooth gradient from deep-brown to milk-white as you go from the Congo to Finland), but it was pretty technical & it’s enough to say that it is the case: race is not a social construct. Anyway the breakout of the cluster analysis was pretty cool. It goes like this:
Common to hear progressives argue that meaningful genetic drift (at least above the neck) is impossible on human historical timescales. Murray argues that natural selection for a given trait would move slowly if it were dependent on a single mutation that had to crawl through the whole population over generations, likely with significant inbreeding. but if a trait is polygenic (i.e. affected by many genes), & all those genes are already floating around to varying degrees throughout the whole population, then selection pressure can cause them to emerge, as a suite, very quickly. This is why breeding Siberian silver foxes for tameness generated a puppy-like, curly-tailed, domestic-dog phenotype within a few generations. The suite of adaptations was present to varying degrees in all the foxes, so they didn’t need extreme inbreeding & culling to spread the trait. This suggests it may be precisely the haziest & most multivariate adaptations (intelligence, attractiveness, conscientiousness) that can be selected for most rapidly.
Murray argues that ‘FBI crime statistics’ shouldn’t scare people: “we’ve learned to live with a 90% differential in homicide rates based on the Y chromosome, so we should be able to tolerate the smaller variations that exist between races”. That makes a certain amount of sense, but we treat men & women quite differently, both culturally & as a matter of policy, to accommodate the male propensity for violence. I don’t think there’s a clean or gentle way to generalize that to race. Stay tuned for his new book, I guess.
Class
IQ discourse is weird. We’re told that IQ doesn’t just make you a better software developer - all else equal it makes you better at any job, all jobs. A higher-IQ farmer is meaningfully better at farming. As well as more honest & prosocial & mentally stable & tall & handsome. If it’s The universal adaptation, with no drawbacks, why aren’t we smarter? No pre-modern culture viewed intellect as the Yardstick of Humanity the way we do. They certainly don’t seem to have consciously chosen mates for intelligence. Why was fitness so complex & multivariate then, & reducible to a single number now? Was it just mechanization? If Farmer Eoforhild was so smart, how come my longax is in his spine?
Twin studies are also weird. Adoption is such a rarified circumstance - the kids would come disproportionately from struggling households, their adoptive families would be disproportionately wealthy/stable, & the whole situation of being raised by someone arbitrarily different from you genetically has got to introduce all sorts of confounders. I wonder if this has been studied.
Murray’s argument that nurture is irrelevant to child outcomes is pretty weird given the enormous volume of nurturing that humans instinctively do, & the extent to which they’re rewarded - socially & neurochemically - for doing it. Virtually all parents, in all societies, do a lot more than just “keep your kids alive & don’t horrifically abuse them” - which Murray seems to think is all that can be done about child outcomes.
A lot of this section is just a struggle with free will. The literature divides all determinants of human behavior into Genetics & Environment - but then they examine genetically identical people in extremely similar environments, & they still find substantial variation, so they chalk that up to Non-Shared Environment (NSE) - but everything they’ve tested as potential “NSE” factors is bogus - virtually no explanatory power. It’s possible that there’s some hidden reserve of NSE variables that no one has considered yet that are super determinative, but the more parsimonious explanation is that people aren’t meat robots & they make choices, at least some of the time.
Murray spends a lot of time showing the ineffectiveness of various Dem Programz aimed at cognitive/behavioral outcomes among lower-class, low-IQ populations, so as to establish that people pretty much are what they are. But the fact that our monstrous public education establishment can’t do something is not evidence that it can’t be done. Seems pretty clear that interventions in individual children’s lives can make a difference - there are elements of interpersonal alchemy & Providence that make these things happen, but they clearly happen. We just don’t know how to summon them at will & mass-produce them, which may not be all bad.
Murray says all mental illnesses are spectrum illnesses, with normally distributed polygenic scores - in other words, you do not either “have” or “not have” schizophrenia, anxiety, depression. Murray says the binary model of mental illness encourages researchers to think only of the pathological end of the distribution - but we could learn a lot from trying to understand the experiences of people who have great strengths, on the other side of the distribution. How do particularly happy people think & act differently? How is their neurochemistry different? Are there tradeoffs?
> A higher-IQ farmer is meaningfully better at farming. As well as more honest & prosocial & mentally stable & tall & handsome. If it’s The universal adaptation, with no drawbacks, why aren’t we smarter?
I've seen it suggested that IQ (and height and general health) are downstream (to at least some degree) of a kind of general factor of genetic health ("mutational load" is one term that comes up here). The idea is that external impositions almost always cause net damage, so some people are just closer to "the ideal human" and these people have better outcomes in just about every sphere, just because of having fewer mutations or whatever.
The question of why this genotype isn't more universally dominant is interesting. One factor going into it may be exposure to external mutagens, but it's not clear how much this varies over times and places. Another is that the "higher-quality" genomes may or may not confer an actual selective advantage -- much discussed in this conversation is the "IQ shredder" phenomenon, where economically successful areas have below-replacement birthrates and end up sucking the most successful subsets of surrounding (poorer, more fertile) areas in and causing them to have fewer children. It seems that in modern times male IQ has been a net fertility benefit, but female IQ has been a net detriment, presumably due to smarter women having greater success in education/career/etc and thus prioritizing it over family. But I don't know if there's any way to quantify the magnitude of this effect over generations.
> Murray’s argument that nurture is irrelevant to child outcomes is pretty weird given the enormous volume of nurturing that humans instinctively do, & the extent to which they’re rewarded - socially & neurochemically - for doing it. Virtually all parents, in all societies, do a lot more than just “keep your kids alive & don’t horrifically abuse them” - which Murray seems to think is all that can be done about child outcomes.
I think this is basically the result of over-interpreting twin studies. The modal twin study is not actually a study of identical twins raised apart -- this circumstance is too rare to get a lot of data on. Rather, it's a comparison of pairs of identical twins (raised together) and fraternal twins (raised together). The idea is that, to the degree that identical twins are more similar than fraternal twins, this must be the influence of genes over environment. But this mechanism is obviously not capable of isolating the real effects of nurture/environment, because almost all pairs of children raised in separate households are also born of different parents. And adoption studies are 1. much rarer and 2. may well come with inherent confounders, as you note.