Some Personal Reflections on the Genetics of Intelligence
(This essay was originally published on January 16, 2018, as part of the Genetics and Human Agency project, organized by Eric Turkheimer at the University of Virginia and funded by the John Templeton Foundation. The GHA website is being retired and I am copying posts here. I have corrected a few errors in this essay but otherwise, it is unchanged from its previous version.)
Last March, I developed a terrible sore throat — eating felt like swallowing crushed glass, and my lymph nodes were tender to the touch. For various reasons (laziness, moving for sabbatical and then moving back, relying only on my OB-GYN during the many years I was either pregnant or nursing), I don’t have a primary care physician, and instead, queue up in urgent care with hungover undergrads and yoga instructors without health insurance whenever I am sick enough to require medical attention.
My daughter’s preschool posts the sick updates on the door to her classroom with a laminated sign. “Today there are ____ kids out because of ____________.” It’s a type of medical ESP: What illness will I come down with tomorrow? Back in March, the sign read “4” and “strep throat.”
After the urgent care physician (muscles, gelled hair, medical degree of questionable quality) examined my throat, he was skeptical about strep throat. “You don’t have white streaks on your tonsils. You don’t even have tonsils.” I insisted on the strep test anyway.
The test was positive, ergo I had strep throat, and I was sent home with a round of highly effective antibiotics.
*****************************************************************
Back in 2008, when I was a graduate student doing my clinical internship, I was depressed. Boston is cold. I worked at a psychiatric hospital where my office was in the basement. Every morning I had to shovel my car out of the snow; there was no time to do research; my boyfriend lived in Virginia. When I burst out crying one afternoon, my supervisor (psychoanalyst, pristine white bob, black Eileen Fisher) wasted no time in giving me the number of a Harvard Square psychiatrist.
I didn’t have to insist on any tests. The psychiatrist asked me a few perfunctory questions about sleeping and appetite and crying, did his due diligence about suicidal ideation, and I was sent home with a prescription for Lexapro.
I took the pills for a few months then threw them away when I moved to Texas. There was so much sunshine. Who needs Lexapro when there is so much sunshine? I’m still not sure if the pills did anything. Maybe I just needed a kindly gentleman of a certain age to nod reassuringly while I talked about how lonely and cold I was that year.
*****************************************************************
Talking about whether something is a biological construct is slippery and difficult. Strep throat is a biological construct. When faced with conflicting information between my symptoms and the presence of streptococcal bacteria, the biological indicator won the day, because it’s more than an indicator. The bacterial infection is having strep throat.
Depression is not (yet?) a biological construct. Lots and lots of time and money have been spent trying to find biological correlates of being depressed — plasma ACTH and inflammatory cytokines and hippocampal volume and the serotonin transporter gene and even a few SNPs from the latest GWAS. But none of that information was necessary to diagnose me with depression in 2008, and none of it is necessary now. In 2008 and in 2018, to be depressed is to burst out into tears in inappropriate professional settings, and when the behavior conflicts with the biological indicators, we go with the former and ignore the latter.
*****************************************************************
Is intelligence a biological construct?
More difficult, and more interesting — should intelligence be defined at the level of biology? Is it just a matter of time before we discover enough genetic variants that we shift, perhaps slowly or perhaps all at once, to a new understanding of intelligence at the level of the genome?
In a blog post last summer, the physicist-turned-geneticist Steven Hsu summarized the scientific consensus on intelligence as follows:
“0. Intelligence is (at least crudely) measurable
1. Intelligence is highly heritable (much of the variance is determined by DNA)
2. Intelligence is highly polygenic (controlled by many genetic variants, each of small effect)
3. Intelligence is going to be deciphered at the molecular level, in the near future, by genomic studies with very large sample size.”
Hsu claimed that “every single listed participant in the above debate” — a list that includes me — “would agree with those claims.” But in my view, 0–2 are empirical claims that have been settled rather definitively, whereas 3 is an altogether different kettle of fish.
Intelligence could potentially be a biological construct that we currently see through a glass, darkly, the way 19th-century physicians understood tuberculosis. This seems to be what people mean when they say that intelligence is going to be “deciphered at the molecular level” or “understood at the level of the genome” — that intelligence will be like strep throat, something we define at the level of the underlying biology.
Or, maybe not.
What we do know: Intelligence is massively polygenic, with very, very many associated genetic polymorphisms that each probabilistically contributes only a very, very little to phenotypic variation; all the genetic variants involved have yet to be identified; it is unclear if the variants that have been identified are actually the causal loci; there is not yet a clear understanding of the mechanisms that link genetic variation with phenotypic variation, and these mechanisms might be quite indirect and dependent on environmental mediation or moderation.
It is at this point in the conversation that dudes-on-the-internet often feel compelled to point out to me that genome-wide association studies (GWAS) have identified lots and lots of specific genetic variants as associated with intelligence, or with educational attainment (which many researchers view as an easily measured proxy phenotype for intelligence). And, even more trait-associated SNPs will be identified as GWAS sample sizes continue to climb into the stratosphere. I’ve now been in the audience at least twice when researchers presented results from the most recent iteration of the educational attainment GWAS (so-called “EA3”), and yes — the number of associated SNPs keeps increasing, as does the percent phenotypic variance (within European ancestry samples) in test scores and other educational outcomes explained by measured SNPs.
To be clear: I am not a GWAS hater, and I think GWAS of educational attainment and intelligence have been carefully done and have produced results that are immensely valuable to the scientific community.
But, at the risk of being redundant, associated SNPs are not necessarily causal variants; there is not (yet?) a clear understanding of the biological or environmental mechanisms for how the identified genetic variants affect intelligence; and those mechanisms might be indirect, with multiple, intervening, socially-dependent processes.
To use an old thought experiment proposed by Sandy Jencks: if we lived in a dystopian society that didn’t allow red-headed children to go to school, we would certainly find SNPs that were associated with intelligence and with educational attainment. (They would even differ between races!) We would not, however, properly understand intelligence as a biological construct.
I don’t think that all of the genetic variants associated with intelligence will turn out to be related to intelligence via red-heads-can’t-go-to-school-type mechanisms, i.e., culturally-specific, social responses to arbitrary phenotypic characteristics. But, Jencks’ fundamental point still stands. As Turkheimer pointed out in “Heritability and Biological Explanation,” and Kendler pointed out in his writings on psychiatric nosology, and Pritchard and colleagues pointed out most recently in their paper on the “omnigenic model” — lots of tiny correlations with genetic differences do not, in and of themselves, produce a coherent biological explanation of intelligence “at the level of the genome.”
I honestly am not sure whether our conception of intelligence will and should change from a psychological phenotype with biological correlates to a biological phenotype that is “deciphered at the molecular level.” Maybe Eric Turkheimer’s gloomy prospect is right, and knowing the SNP variants associated with intelligence will help us appreciate the nature of intelligence as well as knowing where the pits are on a CD helps us appreciate whether a song is good or not. Or maybe James Lee is right, and if we just know enough about which genes and what they do and how the gene products are related to each other, if we keep on keepin’ on with ever bigger data, intelligence will cohere at the level of the genome.
Is intelligence the new consumption? How do you know? How would you know for sure? If you are super confident in your answers, I personally suspect you haven’t thought about the question very deeply.
*****************************************************************
One last medical story: Last January, I had to have uterine surgery. Coming on the heels of two C-sections in two years, I was terrified at the prospect of general anesthesia — the black wave of unconsciousness crashing down, and then slowly swimming toward wakefulness at the other end.
My OB-GYN (Texas accent, good blowout, tasteful diamonds, and tasteful Botox) offered to tie my tubes while she’s “already down there.” I am 35 years old. I have two children ages 5 and under. My children’s father and I are both research-active faculty at an R1 university, and we spend about 50% of our take-home pay on various forms of childcare. By any reasonable standard, I should be done with having children.
And yet. When it came to the moment of actually foreclosing on my future fertility — on the possibility of another new life — I couldn’t do it.
*****************************************************************
In 1927, Carrie Buck had tubal sterilization while imprisoned in the Virginia State Colony for Epileptics and the Feebleminded in Lynchburg, Virginia.
In 1910, Madlener introduced his technique for crushing and ligation of the fallopian tube. By 1919, the technique had been used in 89 procedures, resulting in 3 deaths. It is now described as a technique that is “only of historic significance as it should be abandoned due to high failure rates.” The first laparoscopic tubal sterilization was not performed until 1936, in Switzerland.
The favored anesthetic agent for abdominal surgery in the 1920s seems to have been ether. It was “flammable and explosive, and its after-effects were extremely unpleasant. They were described by Henderson and Coburn in the following terms: ‘patients… were flaccid, cyanotic, pallid, or grey, with empty veins, weak peripheral pulses, and depressed respiration; it was one to three hours before consciousness returned, and this was followed by nausea, vomiting, and retching for some time afterwards.’ ”
What must have Carrie Buck been feeling when she was swimming up toward consciousness? Imprisoned, separated from her child, robbed of her fertility, cut open, retching, grey.
Whenever I hear Oliver Wendell Holmes’ decision in the Buck v. Bell Supreme Court case (“three generations of imbeciles are enough”), I think about how nauseated Carrie must have felt from the ether. I think about the pink pearly abdominal scars left by obstetric surgery.
*****************************************************************
Carrie Buck’s sterilization was the result of a policy that was motivated, at its root, by revulsion for poor people and for Black people, but it was justified by research on genes and intelligence.
There’s now a marker for Buck’s grave in Charlottesville, Virginia, where I was a graduate student. This year, in Charlottesville, white nationalist Richard Spencer led a crowd of torch-bearing protesters: “What brings us together is that we are white, we are a people, we will not be replaced.”
Everything old is new again.
Compared to when I started graduate school in 2003, there’s a new insouciance among academics and writers about the genetics of intelligence. Some part of this is a bit relieving. After more than a decade in a semi-defensive crouch, I get the appeal of being able to say, plainly and even a little bit obnoxiously — “You can measure intelligence and it’s heritable and it matters for life outcomes!”
But on the other hand, I think the attitude of “We’re just asking questions? What’s wrong with asking questions?” is morally blinkered. The nonchalance with which scientists and commentators are suggesting that we might screen embryos for polygenic scores or that we might use gene-editing technologies to genetically engineer millions of people or that White people might have evolved to be smarter than Black people because winters are colder in North America is — to be frank — appalling. There is a middle ground between “let’s never talk about genes and pretend cognitive ability doesn’t exist” and “let’s just ask some questions that pander to a virulent online community populated by racists with swastikas in their Twitter bios.”
Last summer, my colleagues and I wrote an article about genes and intelligence in Vox, which was tweeted out by the white supremacist publication, American Renaissance:
“Only a few straw men remain b/t liberal academics and our view on race and IQ. What happens when they fall away?”
You tell me: what happens?
No, really, what happens?
