Most of this stuff is well known to researchers, as acknowledged, and I don't think that finding IQ is only 40% or 50% heritable will make people's political hopes and aspirations that much more realistic. The remaining 50% or so is everything including the prenatal environment, any diseases or developmental irregularities, family environment, schools, friends, books people have read, etc. There's just much less room for social intervention to create equality than assumed in public policy debates, where nobody will say that rich kids are smarter than poor kids in part because of their genes.
A key point is that the replication crisis has shown that all research and research publications that were not pre-registered are suspect, especially in the field of academic psychology. Confirmation bias is a major factor in the conclusions researchers draw behind the scenes, and to be fair, this observation applies to "environmentalist" research as well. It's not like Cyril Burt was a bad apple in an otherwise healthy barrel of psychology research. The replication crisis shows us that a sizable portion of the barrel contained rotten apples that made it past peer review. That, by the way, is the only valid defense of Burt's "study" I can come up with.
How do you jump from emotions to ideology? That is a crazy interpretation of my perspective!
As is explained in this post (and many of my other posts), what is fundamentally at stake for me is that my mathematical career (and more specifically my ascension to a level that I had long thought I was biologically incapable of) made me rethink my entire perspective on language, logic, intuition, and cognition.
This is a very important topic, adjacent to the IQ debate, but of prime importance to many practitioners of mathematics—including the many Fields medals and Abel prizes who have read and praised my book.
The adjacency is this: it is very unlikely that math talent could be as malleable as I think it is (alongside Descartes, Einstein, Grothendieck, Feynman, and many other) if IQ were to be 80% heritable.
The idea that we mathematicians could all be delusional about our own life experience is emotionally-loaded, but this has nothing to do with ideology.
And, for completeness: I used to believe that math talent was innate, but my confrontation with high-level research made me change my mind. (So I really had no prior political objection to hereditarianism.)
A good chunk of IQ hereditarian claims are based on the 1990 Minnesota Study of Twins Reared Apart (Science Magazine) IQ study. DZ-apart twins (DZAs) constituted the study's control group. Critics have argued that the DZA IQ correlations were suppressed/hidden, and that if they hadn't been suppressed/hidden, the Minnesota researchers would have found no evidence in favor of above-zero IQ heritability. No one has ever shown that the critics are wrong, not even Bouchard, in his 2023 response article. Care to comment?
Eh, hereditarianism seemed to be the stronger argument to me despite me not wanting it to be true emotionally. It was only when I saw that it too has its problems (PGS portability, missing heritability, etc) that I changed my mind towards being more partial towards environmentalism.
Interestingly, better schools and pre-natal care and nutrition etc. all improve measures of intelligence AND intelligence is largely determined by forces outside of soclal engineer's control...
There are many things that evade our conscious control, and yet there are many things that we can control, and many more that it is worth trying to control, even if in the end we only manage to capture a fraction.
"Rich kids are smarter than poor kinds in part because of their genes"
I see where we're going here. This is part of the determinism that fuels people like Stephen Miller and RFK Jr. If someone gets sick, they must have deserved it somehow; they were in fact predetermined to get sick by some divine plan, and therefore a) everyone who hasn't yet gotten sick (or say, in a car accident, or some other misfortune) has plot armour, and can't be touched. Meanwhile anyone anything bad happens to deserved it, since they weren't The Main Character.
Survivor bias.
This is the weirdness underlying the current eugenicists. They don't want to just kill certain swaths of the population, like the OG Nazis, which was horrific enough; the new, American Nazis have adopted a sort of American Health Insurer Nazism, in which they want to kill anyone who might have ended up exhibiting the weakness of getting sick.
It's a sort of calvinist thing. You worth was predetermined by some sort of God Sorting Machine, which decided where you would be when certain germs were floating through the air. Bad luck was your fault to begin with! It's very circular.
They use the language of genetics to imagine all of this, even though the vast majority of people are susceptible to things like, say, the measles. Genetics has nothing to do with who is rich and who isn't; this is again, Nazism by way of American Class Hierarchies. Basically medieval bloodline of kings descended from God, etc, but with a gloss of scientific-looking decoration on the top.
This is really how Nazis think, btw. If you think rich kids are inherently smarter you are probably a Nazi. But then, there are many types of smartness, and the idea that IQ is all that meaningful is pretty stupid. Most people who think so have very little social smarts, since their creepy ideas don't play well at parties.
If you open Wikipedia's page on Richard Hanania you will very quickly find that he is (was, if you believe him when he says he changed) white supremacist and wrote for Breitbart, as well as being a contributor to the planning stages of Project 2025, the current fascist project to take over america.
So to answer your question, "Yes, and we have documentary evidence that proves it".
"The remaining 50% or so is everything including the prenatal environment, any diseases or developmental irregularities, family environment, schools, friends, books people have read, etc." Well, friends and books are certainly malleable! And these 50% altogether are also harder to measure, which opens space for hopes and aspirations (though perhaps of a less goal-oriented kind).
That said, I do find the 50% picture fairly cheerless (even knowing that I'm fairly high up on it), so I tend to agree with your wider point even while agreeing with David's arguments against taking the data at face value. The gulf between the twin studies and the numbers that a blank-slate view would require is just too large, and there is so much dynasticity in academia that hereditability should be the null hypothesis.
An optimistic way to look at it, even if the "accessible to agency" fraction is measured at a low value: this is a population average, it might well be that the people who try hard (for themselves or for their kids) do achieve substantial success, but there's just not many of them.
Also: 50% isn't proven to be the right figure, and IQ itself isn't the outcome that matters.
I commend you for the effort. If only strong counter arguments like this could be beamed into the heads of the other 3,499,999.
Reading this makes me feel like on some level whatever science I come in contact with I can’t trust unless I’ve done the work myself. It’s not true, but to see a post I fully believed as true turn out to have this level of debate is startling.
I feel ya, but this piece confirms that a disciplined, astute, and self-aware approach makes for very persuasive science indeed. The real culprit is the gee-wiz effect of modern media, the bread and butter of its commercial existence. Social media is culpable to a great extent, but the twin studies phenom’s media profile obviously pre-dates Century 21’s great contribution to humanity’s capacity for mass delusion. (Now if we could just get some of those purebred hereditarians to read the whole thing it might just make the world a nicer place!) As a 35 year high school teacher I never doubted the influence of genetics, but, having taught in rural, suburban, and inner city settings I never doubted the influence of social and environmental factors either. Now that I’m retired I’m thinking about working on my math - with renewed hope!
Another point, perhaps g is 70-80% heritable, but IQ is less so, and math talent even less so. Just like height is more heritable than basketball ability.
I think you guys are looking for a conclusion rather than just looking at the evidence. “Yeah IQ might not be heritable but what it attempts to measure is.” What!? Why is it so important for intelligence to be sealed at birth for you? Maybe its because that kind of simplicity, ironic, gives you peace about the world. It explains so much. If the world is more complex, then you are on less stable ground and you must deal with the challenges that ground presents.
On the contrary, why does it seem so important to YOU that intelligence is not molded at birth? Indeed, it's much more convincing to suppose that it is YOUR skeptic crowd which, at root, is concerned less with finding the truth and more with a pleasant fantasy. Are you actually saying that "everyone has the same potential-ish and can work hard to achieve their dreams :)" is the complex, brutally honest and rigorous choice? Utterly laughable how you people try to talk down to hereditarian proponents.
This presumably goes both ways: maths ability could also be highly heritable when measured independently of the other items that attempt to measure G. Has this been done?
But I take your point that the non-g components of math ability probably are decently heritable too, even if less than the g one. It probably helps to be able to obsess over a particular problem for a few months. That sounds like something that is probably genetic, though less so than IQ. Still, I would guess that math ability, while fairly heritable, is less heritable than IQ.
Some of this is slightly inaccurate, maybe read Kevin Mitchell’s book ‘Innate’.
Just as anecdotal evidence, my mother was adopted. She met her biological family as an adult, and was stunned by their shared personality traits compared to her adoptive family. I’ve always found that compelling because I saw it myself. Admittedly, n=1 but there are numerous other cases like it.
I don't see that anyone has argued against the idea that many personality traits can be inherited or are somehow otherwise "innate".
What they have pointed out, however, is that attempts to show that intelligence is hereditary are generally not scientifically sound for a number of reasons.
Most of this is far beyond my intellectual pay grade, and in retrospect it’s obvious all the different studies used the same few data sets, but as an eighteen year old psychology major I vaguely thought that every time you wanted to run a heritability study, you had to go out and find a fresh group of identical twins who had been separated at birth and run experiments on them. How many of these twins could there possibly be?! I remember wondering. Surely at some point the scientists are going to run out! It’s funny and kind of delightful to discover, thirty years later, such a complicated story hiding behind that little flicker of confusion.
This is a textbook semantic fidelity failure. A phrase (“twins reared apart”) becomes a talisman that people treat as an airtight causal design. Once the label hardens, a chart can borrow its authority even when the underlying data is tiny, confounded, and historically messy. Really valuable to see the trope dismantled at the level of mechanism, not vibes.
I caught Cremieux grossly misrepresenting statistical data on the link between vaccines and autism. He presented a twin study of MMR as statistical ‘proof’ that vaccines do not cause autism. I pointed out that the cited study compares just one vaccine against any other vaccines (no control for vaccination status except MMR), so it only shows, at best, that MMR is not more likely to cause autism than other vaccines. He didn’t deny my argument, but stopped short of conceding that his claim was misleading. An expert in statistics is not free to claim innocence on such a basic logical error, or they would be abrogating the status and reliability as an expert.
I’m glad to see you are a centrist! I think most hereditarians are in fact centrists (i.e., believe enviroment has a sizeable impact). But most environmentalists are almost “forced” to be blank slaters.
There is a reason for this: mathematically as you remove sources of environmental variance you are left with only sources of genetic variance. (And obviously as of yet it’s impossible to remove genetic variance.).
So, environmentalists are left with the issue that if they ever succeed in achieving their goal of perfect enviromental equality, they will only have succeeded in creating a world that is perfectly genetically deterministic.
So *that* is why they are blank slaters. They are forced into it by necessity.
Also, I know you focused on twin studies, but they are really only one piece of a very persuasive picture along with adoption studies and kinship studies, which don’t suffer many of the flaws you identified here and are overall quite consistent with twin studies. Twin studies should just be looked at as one piece of a large bayesian picture.
(E.g., google the Texas Adoption Project; this estimated heritabilty at 0.78 without using twins reared apart.)
I think that you are misframing this as a "nature vs nuture" opposition, while a lot of the variance is attributable to non-genetic factors that aren't amenable to socio-economic dimensions.
I don’t understand the range restriction counter. Heritability is always environmentally constrained (i.e., if somebody suffers from malnutrition heritability will obviously be impacted.)
Wouldn’t we want a heritability for “typical” environments? Why doesn’t couch heritability with the caveat that the envioronment is in fact typical?
I.e., what is the heritability for a typical student at Overland Park Kansas high school?
Your argument boils down to “heritability is low in really bad environments”.
Well, yeah. That is indisputable. Nobody is saying that heritability is 0.8 in sub-saharan Africa for instance.
If you want heritability for "typical environments", then you need to build a database of IQ variance among students coming from "typical environments."
IQ variance is 15 for the general population, but it is certainly much lower for these "typical" students.
So when you observe that the typical IQ gap between MZA twins is 6 points, you can't conclude "this is much smaller than the expected value of 15, therefore IQ is mostly heritable". You should conclude "this is somewhat smaller of the expected value of X (with X << 15), so IQ is somewhat heritable".
Besides, my argument doesn't boil down to “heritability is low in really bad environments".
It boils down to: "this is demonstrably bad science and the smoking gun is that Bouchard has spent the past 35 years refusing to release his control group data".
It is incorrect *without* the adjustment, of course.
I haven't looked at the methodology, but 0.4-0.8 is a pretty wide range ;)
[One issue with all these subtleties, which is behind my conclusions 1 & 2, is that model credibility naturally decreases with complexity, for good reasons (curse of dimensionality / p-hacking) and bad reasons (people, including me, tend to get bored...
I fundamentally understand the desire to estimate heritability, and I fundamentally understand the allure of MZA studies, twins reared together studies, and adoption studies, and I looked at them with an open eyes... my personal conclusion is that the core matter is too complex and messy for solutions to be clear-cut... It's a sad conclusion, not only that I'm happy to reach.]
> Heritability is, by construction, a population-level aggregate. Before it can inform policy-making (or even personal decision-making), it must be interpreted at the level of individuals.
Well, no, not really. An effect can be minor at the level of individuals but still be hugely important for group outcomes and society overall. For examples, a small difference in means or variance between two groups leads to much larger disparities at the tails (share of each group in the 1% extremes of the trait).
Yes and no. It may matter for the polygenic tails, but no-one knows for sure where one is situated (in the absence of an accurate test) and, for the "random" person, the effect is negligeable.
As for Scott Alexander's post, I had scanned through it but not fully read. My general feeling is that he's slowly moving away from his hereditarian priors (which I completely understand since I started out with the same priors — my earlier post attempts to respond to his "parable of talents" post: https://davidbessis.substack.com/p/beyond-nature-and-nurture). But this process takes time and I'm disappointed to see that he often continues to dismiss the centrist viewpoint (eg, I definitely don't view myself as "nurturist", as he labels us here https://www.astralcodexten.com/p/the-good-news-is-that-one-side-has) — but this is an evolving picture and he seems to be moving in the right direction.
BTW, I do hope he reads this post and finds it some merit (he's super smart, but he's not omniscient and he's not a statistician). I liked that he seems to have carefully read what Sasha wrote (despite Cremieux trying to dissuade him by disparaging Sasha and painting him as an ideologue) — this is a sign of being open on a topic where he started from a very rigid position.
That is a popular argument among people who want intelligence to be hereditary and among people who want various abilities to be gender linked, etc., but it assumes all sorts of unknown things about the distribution and phenomenon. Average life span among Danes may be significantly higher than average life span among Pakistani villagers, but that doesn't tell you anything about which population has more people of age 100+.
Despite my inability to really process a fair amount of this, I think I grasped the basic story of this author’s attempt to explain a complicated topic to others, but first to himself. Fascinating, and certainly one kind of essay that I would almost certainly never read if it weren’t on substack. Another thing drawing me to it is that @Rafe Meager restacked it. A Rafe Meager restacking often leads me to slog through material that’s puzzling and far above my pay grade, but also riveting when I start to understand pieces of it. To sum up— it’s good to inherit intelligence, but also good to subscribe to people whose minds can dance around yours. No matter how “smart” you are, it’s not a bad idea to have smart friends.
Your most important point: "For all their performative battles, hereditarians and blank-slatists operate within the same deterministic frame (genetic determinism vs social determinism), denying any meaningful role for human agency and the messy, noisy process we call life."
Yet most of the comments continue to assume that determinism of one type or another controls for 100% of outcomes. And by "outcomes" I don't mean test scores: I mean success in life. My view is that individual human agency, for those who choose to exercise it, is the dominant variable. Free will, which cannot be explained by deterministic models of the universe, nevertheless is axiomatic to all human thought, which otherwise is cosmic noise. This is what separates humans from jellyfish and gnus.
Do you know if anyone has done research with “hyperdonors” of sperm like Jonathan Jacob Meijer? Obviously this would have its own limitations, as zygotes don’t contribute genes in a predictable pattern and people undergoing in vitro fertilization are not a broad cohort, but 500 kids with a single father and separate households would be interesting.
The other part that is such nonsense in the whole heredity argument is the notion that genes are switches for broad phenotypical attributes. I have achieved a high level of success academically, but I think much of my skillset came from absurd oversensitivity (which sucks, btw). I think my brain had to cope with an excess of stimuli and became conditioned to pattern-creation and matching.
I don’t think I have some gene for good pattern matching.
Hasn’t it become fairly clear that an individual genome has limited predictive power with regard to the development of the individual? Rather, it seems that the role of the genetic material is more that of a library rather than a formational hierarchy. In other words, the entire chemical payload present in zygotes is at least as important as the genes themselves and the capacity for individuals to be influenced by environment is far greater than a system based simply on genetic heritability. Anyway, just my opinion, but it all seems to line up…
if that was true identical twins wouldn't end up being..identical, no? There must be ~zero noise and ~zero sensitivity to initial conditions in development bc if there was, it would compound over the thousands, millions, billions (?) of branch points in development and we would expect to see some difference in face, height, etc.
I read all of this and commend your effort and good faith intellectual pursuit. Kudos. Now that you’ve tackled the heritability variable, please dig into IQ. I was a little miffed that this notion (IQ), so clearly developed and contingent on environmental acquisition of information, was discussed as if it was fixed and static and measuring what it claimed to measure. I felt the piece would have benefitted from a more expansive discussion of what IQ is purported to measure/be and whether these tests get at that in a valid way. To interrogate IQ as valid, fixed, extant, etc. is all the more relevant to this discussion as the point of this article was to examine and challenge assumptions of the “purity” or “cleanliness” of twins reared apart studies. I think a rigorous analysis of IQ in the same spirit of this piece on heritability would make a valuable companion.
Ah, another long story. The issue is that IQ is mildy meaningful, and it's hard to build a debunk of the exaggeration of something that is mildly meaningful — the IQ mafia will instantly label me a flat-earther (they already do).
Ashamed to say I took those Jim Twin anecdotes at face value. I suppose the conformity is still striking, but that’s the 1950s for you. I don’t doubt genetics played a part, there’s just not the empirical record I thought. Man- psychology has a spotty record.
First-time reader here! I'd like you to know that this reads like a transcribed HBomberguy video and I mean that as the highest possible compliment. I'm going to show this to my mother!
It's far from certain that Cyril Burt committed fraud, rather than simply being sloppy with data retention in the way many from his era were.
https://arthurjensen.net/wp-content/uploads/2014/06/1991-jensen-1.pdf
Most of this stuff is well known to researchers, as acknowledged, and I don't think that finding IQ is only 40% or 50% heritable will make people's political hopes and aspirations that much more realistic. The remaining 50% or so is everything including the prenatal environment, any diseases or developmental irregularities, family environment, schools, friends, books people have read, etc. There's just much less room for social intervention to create equality than assumed in public policy debates, where nobody will say that rich kids are smarter than poor kids in part because of their genes.
Why do you frame this in terms of "political hopes and aspirations" when this is absolutely not what is at stake in my critique?
(And re Burt: Jensen has his opinion, but it is far from representing the consensus.)
A key point is that the replication crisis has shown that all research and research publications that were not pre-registered are suspect, especially in the field of academic psychology. Confirmation bias is a major factor in the conclusions researchers draw behind the scenes, and to be fair, this observation applies to "environmentalist" research as well. It's not like Cyril Burt was a bad apple in an otherwise healthy barrel of psychology research. The replication crisis shows us that a sizable portion of the barrel contained rotten apples that made it past peer review. That, by the way, is the only valid defense of Burt's "study" I can come up with.
You talk about your emotional reaction to hereditarianism, which seems to suggest you’d rather it not be true for ideological reasons.
How do you jump from emotions to ideology? That is a crazy interpretation of my perspective!
As is explained in this post (and many of my other posts), what is fundamentally at stake for me is that my mathematical career (and more specifically my ascension to a level that I had long thought I was biologically incapable of) made me rethink my entire perspective on language, logic, intuition, and cognition.
This is a very important topic, adjacent to the IQ debate, but of prime importance to many practitioners of mathematics—including the many Fields medals and Abel prizes who have read and praised my book.
The adjacency is this: it is very unlikely that math talent could be as malleable as I think it is (alongside Descartes, Einstein, Grothendieck, Feynman, and many other) if IQ were to be 80% heritable.
The idea that we mathematicians could all be delusional about our own life experience is emotionally-loaded, but this has nothing to do with ideology.
And, for completeness: I used to believe that math talent was innate, but my confrontation with high-level research made me change my mind. (So I really had no prior political objection to hereditarianism.)
https://davidbessis.substack.com/p/beyond-nature-and-nurture
David, Hanania est une fraude totale. Il n'a pas les compétences techniques requises.
C'est probable — et ça explique certainement sa manière de répondre à côté du sujet !
A good chunk of IQ hereditarian claims are based on the 1990 Minnesota Study of Twins Reared Apart (Science Magazine) IQ study. DZ-apart twins (DZAs) constituted the study's control group. Critics have argued that the DZA IQ correlations were suppressed/hidden, and that if they hadn't been suppressed/hidden, the Minnesota researchers would have found no evidence in favor of above-zero IQ heritability. No one has ever shown that the critics are wrong, not even Bouchard, in his 2023 response article. Care to comment?
Sealioning
Eh, hereditarianism seemed to be the stronger argument to me despite me not wanting it to be true emotionally. It was only when I saw that it too has its problems (PGS portability, missing heritability, etc) that I changed my mind towards being more partial towards environmentalism.
So the nuanced take isn't "intelligence is largely genetic", but "intelligence is largely determined by forces outside of soclal engineer's control".
Which is, of course, false since better schools and pre-natal care and nutrition etc. all improve measures of intelligence.
Interestingly, better schools and pre-natal care and nutrition etc. all improve measures of intelligence AND intelligence is largely determined by forces outside of soclal engineer's control...
There are many things that evade our conscious control, and yet there are many things that we can control, and many more that it is worth trying to control, even if in the end we only manage to capture a fraction.
Life is messy, we all need to get over it.
You sound like a Nazi. Are you a Nazi?
"Rich kids are smarter than poor kinds in part because of their genes"
I see where we're going here. This is part of the determinism that fuels people like Stephen Miller and RFK Jr. If someone gets sick, they must have deserved it somehow; they were in fact predetermined to get sick by some divine plan, and therefore a) everyone who hasn't yet gotten sick (or say, in a car accident, or some other misfortune) has plot armour, and can't be touched. Meanwhile anyone anything bad happens to deserved it, since they weren't The Main Character.
Survivor bias.
This is the weirdness underlying the current eugenicists. They don't want to just kill certain swaths of the population, like the OG Nazis, which was horrific enough; the new, American Nazis have adopted a sort of American Health Insurer Nazism, in which they want to kill anyone who might have ended up exhibiting the weakness of getting sick.
It's a sort of calvinist thing. You worth was predetermined by some sort of God Sorting Machine, which decided where you would be when certain germs were floating through the air. Bad luck was your fault to begin with! It's very circular.
They use the language of genetics to imagine all of this, even though the vast majority of people are susceptible to things like, say, the measles. Genetics has nothing to do with who is rich and who isn't; this is again, Nazism by way of American Class Hierarchies. Basically medieval bloodline of kings descended from God, etc, but with a gloss of scientific-looking decoration on the top.
This is really how Nazis think, btw. If you think rich kids are inherently smarter you are probably a Nazi. But then, there are many types of smartness, and the idea that IQ is all that meaningful is pretty stupid. Most people who think so have very little social smarts, since their creepy ideas don't play well at parties.
If you open Wikipedia's page on Richard Hanania you will very quickly find that he is (was, if you believe him when he says he changed) white supremacist and wrote for Breitbart, as well as being a contributor to the planning stages of Project 2025, the current fascist project to take over america.
So to answer your question, "Yes, and we have documentary evidence that proves it".
"The remaining 50% or so is everything including the prenatal environment, any diseases or developmental irregularities, family environment, schools, friends, books people have read, etc." Well, friends and books are certainly malleable! And these 50% altogether are also harder to measure, which opens space for hopes and aspirations (though perhaps of a less goal-oriented kind).
That said, I do find the 50% picture fairly cheerless (even knowing that I'm fairly high up on it), so I tend to agree with your wider point even while agreeing with David's arguments against taking the data at face value. The gulf between the twin studies and the numbers that a blank-slate view would require is just too large, and there is so much dynasticity in academia that hereditability should be the null hypothesis.
An optimistic way to look at it, even if the "accessible to agency" fraction is measured at a low value: this is a population average, it might well be that the people who try hard (for themselves or for their kids) do achieve substantial success, but there's just not many of them.
Also: 50% isn't proven to be the right figure, and IQ itself isn't the outcome that matters.
yeah
I commend you for the effort. If only strong counter arguments like this could be beamed into the heads of the other 3,499,999.
Reading this makes me feel like on some level whatever science I come in contact with I can’t trust unless I’ve done the work myself. It’s not true, but to see a post I fully believed as true turn out to have this level of debate is startling.
Thanks for doing the work to clarify.
Thank you! Now we’re down to 3,499,998 :)
I feel ya, but this piece confirms that a disciplined, astute, and self-aware approach makes for very persuasive science indeed. The real culprit is the gee-wiz effect of modern media, the bread and butter of its commercial existence. Social media is culpable to a great extent, but the twin studies phenom’s media profile obviously pre-dates Century 21’s great contribution to humanity’s capacity for mass delusion. (Now if we could just get some of those purebred hereditarians to read the whole thing it might just make the world a nicer place!) As a 35 year high school teacher I never doubted the influence of genetics, but, having taught in rural, suburban, and inner city settings I never doubted the influence of social and environmental factors either. Now that I’m retired I’m thinking about working on my math - with renewed hope!
Another point, perhaps g is 70-80% heritable, but IQ is less so, and math talent even less so. Just like height is more heritable than basketball ability.
I think you guys are looking for a conclusion rather than just looking at the evidence. “Yeah IQ might not be heritable but what it attempts to measure is.” What!? Why is it so important for intelligence to be sealed at birth for you? Maybe its because that kind of simplicity, ironic, gives you peace about the world. It explains so much. If the world is more complex, then you are on less stable ground and you must deal with the challenges that ground presents.
On the contrary, why does it seem so important to YOU that intelligence is not molded at birth? Indeed, it's much more convincing to suppose that it is YOUR skeptic crowd which, at root, is concerned less with finding the truth and more with a pleasant fantasy. Are you actually saying that "everyone has the same potential-ish and can work hard to achieve their dreams :)" is the complex, brutally honest and rigorous choice? Utterly laughable how you people try to talk down to hereditarian proponents.
This presumably goes both ways: maths ability could also be highly heritable when measured independently of the other items that attempt to measure G. Has this been done?
Probably not
But I take your point that the non-g components of math ability probably are decently heritable too, even if less than the g one. It probably helps to be able to obsess over a particular problem for a few months. That sounds like something that is probably genetic, though less so than IQ. Still, I would guess that math ability, while fairly heritable, is less heritable than IQ.
Some of this is slightly inaccurate, maybe read Kevin Mitchell’s book ‘Innate’.
Just as anecdotal evidence, my mother was adopted. She met her biological family as an adult, and was stunned by their shared personality traits compared to her adoptive family. I’ve always found that compelling because I saw it myself. Admittedly, n=1 but there are numerous other cases like it.
I don't see that anyone has argued against the idea that many personality traits can be inherited or are somehow otherwise "innate".
What they have pointed out, however, is that attempts to show that intelligence is hereditary are generally not scientifically sound for a number of reasons.
Most of this is far beyond my intellectual pay grade, and in retrospect it’s obvious all the different studies used the same few data sets, but as an eighteen year old psychology major I vaguely thought that every time you wanted to run a heritability study, you had to go out and find a fresh group of identical twins who had been separated at birth and run experiments on them. How many of these twins could there possibly be?! I remember wondering. Surely at some point the scientists are going to run out! It’s funny and kind of delightful to discover, thirty years later, such a complicated story hiding behind that little flicker of confusion.
This is a textbook semantic fidelity failure. A phrase (“twins reared apart”) becomes a talisman that people treat as an airtight causal design. Once the label hardens, a chart can borrow its authority even when the underlying data is tiny, confounded, and historically messy. Really valuable to see the trope dismantled at the level of mechanism, not vibes.
I caught Cremieux grossly misrepresenting statistical data on the link between vaccines and autism. He presented a twin study of MMR as statistical ‘proof’ that vaccines do not cause autism. I pointed out that the cited study compares just one vaccine against any other vaccines (no control for vaccination status except MMR), so it only shows, at best, that MMR is not more likely to cause autism than other vaccines. He didn’t deny my argument, but stopped short of conceding that his claim was misleading. An expert in statistics is not free to claim innocence on such a basic logical error, or they would be abrogating the status and reliability as an expert.
I’m glad to see you are a centrist! I think most hereditarians are in fact centrists (i.e., believe enviroment has a sizeable impact). But most environmentalists are almost “forced” to be blank slaters.
There is a reason for this: mathematically as you remove sources of environmental variance you are left with only sources of genetic variance. (And obviously as of yet it’s impossible to remove genetic variance.).
So, environmentalists are left with the issue that if they ever succeed in achieving their goal of perfect enviromental equality, they will only have succeeded in creating a world that is perfectly genetically deterministic.
So *that* is why they are blank slaters. They are forced into it by necessity.
Also, I know you focused on twin studies, but they are really only one piece of a very persuasive picture along with adoption studies and kinship studies, which don’t suffer many of the flaws you identified here and are overall quite consistent with twin studies. Twin studies should just be looked at as one piece of a large bayesian picture.
(E.g., google the Texas Adoption Project; this estimated heritabilty at 0.78 without using twins reared apart.)
I think that you are misframing this as a "nature vs nuture" opposition, while a lot of the variance is attributable to non-genetic factors that aren't amenable to socio-economic dimensions.
This is explained here: https://davidbessis.substack.com/p/beyond-nature-and-nurture
[As for adoption studies, they suffer from range-restriction, as discussed in my post & backed by the Pinker quote.]
I don’t understand the range restriction counter. Heritability is always environmentally constrained (i.e., if somebody suffers from malnutrition heritability will obviously be impacted.)
Wouldn’t we want a heritability for “typical” environments? Why doesn’t couch heritability with the caveat that the envioronment is in fact typical?
I.e., what is the heritability for a typical student at Overland Park Kansas high school?
Your argument boils down to “heritability is low in really bad environments”.
Well, yeah. That is indisputable. Nobody is saying that heritability is 0.8 in sub-saharan Africa for instance.
No, your reasoning is mathematically incorrect.
If you want heritability for "typical environments", then you need to build a database of IQ variance among students coming from "typical environments."
IQ variance is 15 for the general population, but it is certainly much lower for these "typical" students.
So when you observe that the typical IQ gap between MZA twins is 6 points, you can't conclude "this is much smaller than the expected value of 15, therefore IQ is mostly heritable". You should conclude "this is somewhat smaller of the expected value of X (with X << 15), so IQ is somewhat heritable".
Besides, my argument doesn't boil down to “heritability is low in really bad environments".
It boils down to: "this is demonstrably bad science and the smoking gun is that Bouchard has spent the past 35 years refusing to release his control group data".
It’s not “incorrect” per se, it just requires an adjustment for the lower variance.
Which people have done and still get 0.4-0.8 estimate for adoption studies.
It is incorrect *without* the adjustment, of course.
I haven't looked at the methodology, but 0.4-0.8 is a pretty wide range ;)
[One issue with all these subtleties, which is behind my conclusions 1 & 2, is that model credibility naturally decreases with complexity, for good reasons (curse of dimensionality / p-hacking) and bad reasons (people, including me, tend to get bored...
I fundamentally understand the desire to estimate heritability, and I fundamentally understand the allure of MZA studies, twins reared together studies, and adoption studies, and I looked at them with an open eyes... my personal conclusion is that the core matter is too complex and messy for solutions to be clear-cut... It's a sad conclusion, not only that I'm happy to reach.]
> Heritability is, by construction, a population-level aggregate. Before it can inform policy-making (or even personal decision-making), it must be interpreted at the level of individuals.
Well, no, not really. An effect can be minor at the level of individuals but still be hugely important for group outcomes and society overall. For examples, a small difference in means or variance between two groups leads to much larger disparities at the tails (share of each group in the 1% extremes of the trait).
(BTW, have you read https://www.astralcodexten.com/p/missing-heritability-much-more-than ?)
Yes and no. It may matter for the polygenic tails, but no-one knows for sure where one is situated (in the absence of an accurate test) and, for the "random" person, the effect is negligeable.
As for Scott Alexander's post, I had scanned through it but not fully read. My general feeling is that he's slowly moving away from his hereditarian priors (which I completely understand since I started out with the same priors — my earlier post attempts to respond to his "parable of talents" post: https://davidbessis.substack.com/p/beyond-nature-and-nurture). But this process takes time and I'm disappointed to see that he often continues to dismiss the centrist viewpoint (eg, I definitely don't view myself as "nurturist", as he labels us here https://www.astralcodexten.com/p/the-good-news-is-that-one-side-has) — but this is an evolving picture and he seems to be moving in the right direction.
BTW, I do hope he reads this post and finds it some merit (he's super smart, but he's not omniscient and he's not a statistician). I liked that he seems to have carefully read what Sasha wrote (despite Cremieux trying to dissuade him by disparaging Sasha and painting him as an ideologue) — this is a sign of being open on a topic where he started from a very rigid position.
That is a popular argument among people who want intelligence to be hereditary and among people who want various abilities to be gender linked, etc., but it assumes all sorts of unknown things about the distribution and phenomenon. Average life span among Danes may be significantly higher than average life span among Pakistani villagers, but that doesn't tell you anything about which population has more people of age 100+.
Despite my inability to really process a fair amount of this, I think I grasped the basic story of this author’s attempt to explain a complicated topic to others, but first to himself. Fascinating, and certainly one kind of essay that I would almost certainly never read if it weren’t on substack. Another thing drawing me to it is that @Rafe Meager restacked it. A Rafe Meager restacking often leads me to slog through material that’s puzzling and far above my pay grade, but also riveting when I start to understand pieces of it. To sum up— it’s good to inherit intelligence, but also good to subscribe to people whose minds can dance around yours. No matter how “smart” you are, it’s not a bad idea to have smart friends.
Your most important point: "For all their performative battles, hereditarians and blank-slatists operate within the same deterministic frame (genetic determinism vs social determinism), denying any meaningful role for human agency and the messy, noisy process we call life."
Yet most of the comments continue to assume that determinism of one type or another controls for 100% of outcomes. And by "outcomes" I don't mean test scores: I mean success in life. My view is that individual human agency, for those who choose to exercise it, is the dominant variable. Free will, which cannot be explained by deterministic models of the universe, nevertheless is axiomatic to all human thought, which otherwise is cosmic noise. This is what separates humans from jellyfish and gnus.
Amazing article!
Do you know if anyone has done research with “hyperdonors” of sperm like Jonathan Jacob Meijer? Obviously this would have its own limitations, as zygotes don’t contribute genes in a predictable pattern and people undergoing in vitro fertilization are not a broad cohort, but 500 kids with a single father and separate households would be interesting.
The other part that is such nonsense in the whole heredity argument is the notion that genes are switches for broad phenotypical attributes. I have achieved a high level of success academically, but I think much of my skillset came from absurd oversensitivity (which sucks, btw). I think my brain had to cope with an excess of stimuli and became conditioned to pattern-creation and matching.
I don’t think I have some gene for good pattern matching.
Hasn’t it become fairly clear that an individual genome has limited predictive power with regard to the development of the individual? Rather, it seems that the role of the genetic material is more that of a library rather than a formational hierarchy. In other words, the entire chemical payload present in zygotes is at least as important as the genes themselves and the capacity for individuals to be influenced by environment is far greater than a system based simply on genetic heritability. Anyway, just my opinion, but it all seems to line up…
if that was true identical twins wouldn't end up being..identical, no? There must be ~zero noise and ~zero sensitivity to initial conditions in development bc if there was, it would compound over the thousands, millions, billions (?) of branch points in development and we would expect to see some difference in face, height, etc.
I read all of this and commend your effort and good faith intellectual pursuit. Kudos. Now that you’ve tackled the heritability variable, please dig into IQ. I was a little miffed that this notion (IQ), so clearly developed and contingent on environmental acquisition of information, was discussed as if it was fixed and static and measuring what it claimed to measure. I felt the piece would have benefitted from a more expansive discussion of what IQ is purported to measure/be and whether these tests get at that in a valid way. To interrogate IQ as valid, fixed, extant, etc. is all the more relevant to this discussion as the point of this article was to examine and challenge assumptions of the “purity” or “cleanliness” of twins reared apart studies. I think a rigorous analysis of IQ in the same spirit of this piece on heritability would make a valuable companion.
Ah, another long story. The issue is that IQ is mildy meaningful, and it's hard to build a debunk of the exaggeration of something that is mildly meaningful — the IQ mafia will instantly label me a flat-earther (they already do).
Ashamed to say I took those Jim Twin anecdotes at face value. I suppose the conformity is still striking, but that’s the 1950s for you. I don’t doubt genetics played a part, there’s just not the empirical record I thought. Man- psychology has a spotty record.
-Sage
First-time reader here! I'd like you to know that this reads like a transcribed HBomberguy video and I mean that as the highest possible compliment. I'm going to show this to my mother!