Observations on the heritability of intelligence

The reason I think polygenicity is important in this case is that it means there is a huge mutational target that natural selection has to keep an eye on. The constant production of new mutations in sperm and egg cells, the fact that so many of them could affect intelligence, and the fact that they will tend to do so negatively, should, in my opinion, make it harder to push intelligence consistently upwards, when new mutations will constantly be pulling it back down.

Again, I would argue this is a different situation to many other traits. For any trait, new mutations are likely to degrade, rather than improve, the developmental program and biological pathways underlying it. But for some traits, the “goal” of that program is to hit a species-optimal set point. Mutations affecting that program could mean you miss high or miss low – there’s no reason to expect to go one way or the other, really (as far as I can see).

For intelligence, following my argument above, the goal is to hit the maximal level possible. New mutations will thus not just replenish genetic variation affecting the trait (in either direction, as in standard models of stabilising selection); they will tend to push it downwards.

Now, maybe someone will tell me why that actually doesn’t matter, but it seems to me that this will tend to oppose any efforts of directional selection to push intelligence upwards in any given population. Whether that is true or not (or the size of the effect it could have) may depend on how much the trait is dominated by the effects of rare mutations. Various lines of evidence suggest that the collective influence of such mutations on intelligence is very substantial.

That is from Kevin Mitchell.  I do not feel qualified to judge his claims, but nonetheless found the discussion of interest.


Bending over backwards to avoid what is just obvious makes you not intellctuallly honest.

Intelligence, or the perception of intelligence is many things. A less intelligent child adopted and raised by a Tiger Mom would probably test as a genius. Many child prodigies "coincidently" excel in the same field that one of their parents does. It's almost as though the parent works with the child from the time it is a toddler until 1st grade and the child excels. It is like magic (sarcasm).
There is a paradox in intelligence. Many really intelligent people do not excel in life or meet their potential. Many average people do both of these things. It appears that it is often difficult to parlay perceived intelligence into success.

"A less intelligent child adopted and raised by a Tiger Mom would probably test as a genius." And yet endless studies of adopted children say "not so".

As the author admits, IQ is the #1 predictor of success in just about every field. I suggest you read up on the topic you are discussing.

How could you possibly think that this hadn't been tested before, anon?

Yeah, the author doesn't seem to engage directly with the evidence. I buy the case that you might not prima facie expect big intelligence gaps between populations, but the evidence for these gaps is so strong at this point that you really need to directly engage that vs. speaking in generalities about what you'd expect mutations to do.

I think it might be true that the intelligence gaps simply are not very large in some "absolute" sense - i.e. the raw computing power of different groups is, at some level, very similar - but the overall cognitive infrastructure of humans magnifies those differences immensely. SSC writes a little about this idea here: http://slatestarcodex.com/2017/08/02/where-the-falling-einstein-meets-the-rising-mouse/

And, as always, in a perfect world this simply wouldn't matter., individual variation comfortably outstripping variation between groups and whatnot, and we could spend time figuring out how to enable human flourishing at the variety of cognitive levels that exist in any group however you want to define it. This only matters because of the persistence of zero-sum (really, negative-sum given the memetic spread of its assumptions) tribal thinking.

Well if it's genetic then you would not expect expensive environmental interventions to succeed, whereas if the gaps are environmental then you might expect them to succeed. So the question of genes versus environment seems to matter regardless.

The real issue is the interaction between genotypes and environments. A particular Gene may only be advantageous in certain environments.

Look at it like crops, my field of work. The best, highest yielding corn variety, if put into poor soil will yield lower than an old open pollinated landrace variety. It is made to thrive in ideal conditions, but can't handle adverse weather.

The real reason why it matters is trying to find interventions that will help. The black-white IQ gap exists; if it is due to environmental reasons, those are potentially correctable (and indeed, we know that some of the gap historically was - ending segregation decreased the size of the gap). If the problem is genetics, then nothing short of genetic modification will close the gap.

If black children do worse than white children by X amount due to genes, then we can look at whether or not a majority black school is in fact living up to its potential or not; if it is worse by X + Y rather than just X, then we know that there's a problem. If it is meeting or exceeding X, then it is doing a good job.

"I think it might be true that the intelligence gaps simply are not very large in some "absolute" sense - i.e. the raw computing power of different groups is, at some level, very similar"

Lebron James doesn't have 20 times the ability at basketball than the average NBA player, so why is he compensated that way? Google wasn't 250 times better than Yahoo, so why was it valued that way? Minor differences in ability result in outsized rewards in competition.

Conclusion: differences in life outcome by demographic are not necessarily evidence of demographic discrimination.

This strikes at the root of the leftist ideology, of the new soviet man, of the basis of all difference to be power struggle. That is the only reason this is considered controversial. It is controversial in the same way that heliocentrism was controversial to the church.

Yes - I'd self-identify as a leftist, but I agree. My point not that any of that is wrong, just that our observations on group intelligence aren't necessarily inconsistent with what Mitchell says. Intelligence may well be kept within a narrow absolute band by the factors he mentions, but, as you allude to, small *absolute* differences can cause comparatively large relative effects.

We don't really have any meaningful sense of absolute intelligence; IQ is a relative scale rather than an absolute one. There are humans which are four times stronger than other humans; it is possible that there are humans which are four times more intelligent than other humans as well. Or it is possible that the smartest human is like, a hundred times smarter. We have no idea. That SEEMS unlikely, but it is worth remembering that some people can barely read, while others can understand and follow very complex and esoteric arguments written in aggressively dense language.

Reminds me of the intelligent design argument that the eye is irreducibly complex. Except now it’s the converse: the eye is so complex, it must be easily and constantly reducible.

Actually, considering just how many ways eyes function in various parts of the animal world, using a human eye alone as proof of either evolution's and/or intelligent design's weaknesses or strengths is pretty reductionist, and clearly irrelevant.

However, as for how bananas were designed to be opened, well, that is monkey play, as shown by this primate, without using language at all - https://www.youtube.com/watch?v=-GyXUgzCz2Y

"I do not feel qualified to judge his claims"

But the least qualified commenters will have the most to say.

And it will be so entertaining to see.

How does his model account for the known example of humans evolving way higher intelligence (at least in some specific areas) than our nearest relatives, in a few million years? We know that happened, we can see that we're much smarter than chimpanzees, and we can see increasingly complex tools and art and such in ancient sites from humans and other hominids over time, which presumably reflects increased intelligence.

Language complexity is the core of intelligence. Human evolution is about the revolution in our ability to communicate. It drove intelligence and intelligence drove language complexity.

new mutations will constantly be pulling it back down.

Will they? How does he know? What is interesting is his assumption that more intelligence is a good thing - "For intelligence, following my argument above, the goal is to hit the maximal level possible." This is not a surprise. He is an academic and so academics tend to think the production of people like them is the peak of all evolution. But perhaps there is an optimal level of IQ, say, 102. Any higher than that and you start to get problems that reduce your fitness. There is some reason to think so. After all, it was not the chess club getting laid in High School.

But leaving that aside, why would most movement be down? We have evolution to make sure that does not happen - the really stupid die. If the average IQ is N and someone with an IQ of N+d has D more children, while someone with an IQ or N-c has C fewer children, (for all values of N, c, C, d and D) then even if a hundred times more children are born with IQs below N than above, the population's average IQ will grow.

However the main problem is his argument from disbelief. He may well think that IQ could not evolve in different ways in different continents, but that does not prove it did not. After all, East Asians evolved to have body shapes suited for Siberian weather - short, and stocky in order to retain heat, while people from the Sudan (in the general sense) evolved long thin limbs to radiate heat. Not to mention people across Africa tend to have black skin. While Tibetans and Peruvians do better with higher altitudes. So if the races evolved in stable enough environments to change their bodies, there is no reason to think their minds did not evolve differently as well.

Although, as he said, we ought to look at environment first. However Black America's nutrition has improved a lot. The gap in academic performance remains.

"But perhaps there is an optimal level of IQ, say, 102. Any higher than that and you start to get problems that reduce your fitness. There is some reason to think so. After all, it was not the chess club getting laid in High School."

These are some conclusions from the Terman survey of high IQ children who were later tracked in life:

"He found that gifted children did not fit the existing stereotypes often associated with them: they were not weak and sickly social misfits, but in fact were generally taller, in better health, better developed physically, and better adapted socially than other children. The children included in his studies were colloquially referred to as "Termites". The gifted children thrived both socially and academically. In relationships, they were less likely to divorce." (wikipedia)

A survey of Mensa members found that they suffered higher rates of mental illness than the general population:


Maybe Mensa members are not representative of smart people generally, but still...food for thought.

Right, Mensa members self-select entering a group that acknowledges high IQ.

I would suspect that well adjusted people with high IQs don't join MENSA, because they have better things to do with their time (work, family, friends, hobbies, community, etc.)

I was going to add that. I have a friend who is a very high IQ mathematician, and I think he would laugh at joinin Mensa: "And take away time working on math and playing chess?!"

"Not to mention people across Africa tend to have black skin." (Citation Needed)

Haha, this made my morning!

TC has (as usual) placed his chip firmly in the "intelligence is a single (self-consistent) phenotype" camp. I think it is obvious that intelligence is emergent and multifaceted and almost as obvious that there are severe down-sides of having a 3-sigma IQ. I am also not aware that we are as a single world-wide population in an environment which favors higher intelligence. After all: economists.

Probably because you're not familiar with the field

Can you point to some data actually showing that there's some negative consequences to greater intelligence?

Now, there are pretty clearly negative consequences to some ways of *getting* higher intelligence--for example, the brute-force approach of just making the brain bigger probably works for making your kids smarter, but go very far in that direction and they won't make it out the birth canal. But if you assume that there's an optimal level of intelligence, that should say that even cost-free ways to raise intelligence won't be selected past some optimal point.

Ever higher intelligence doesn't have to be bad, it only has to be unnecessary.

When someone is perfectly able to hunt, gather, farm, enage in commerce, flirt, dance, that might be enough.

Bonus points if you have a singing voice.

Chicks definitely dig the guitarist and playing the guitar takes a big brain. Females select for "creative" output. Seems the nerds borrow the infrastructure.

Women’s estimated fertility predicted their short-term (but not long-term) preference for creativity over wealth, in both their desirability ratings of individual men (r=.40) and their forced-choice decisions between men (r=.46). These preliminary results are consistent with the view that creativity evolved at least partly as a good genes indicator through mate choice.


Who sez economists are smart? They aren’t, and they needn’t be. What they are is woke to the truth that there’s no such thing as a free lunch. TANSTAAFL.

So, here is a link that explores in depth this recent fascination of MR concerning science in general - https://www.youtube.com/watch?v=0Rnq1NpHdmw

(With bonus referencing of a Time cover titeled 'Those Asian-American Whiz Kids')

I remember that cover. It was great!

Suggesting that maximal intelligence is "the goal" seems an odd proposition. Intelligence has diminishing marginal utility in an evolutionary sense, and, above some level may have nominal or potentially even negative effects on fecundity. There are also energetic (metabolic) and physical (birth canal) constraints that interplay with intelligence.

Perhaps reading a bit on neutral and nearly neutral theory and mutational landscape would bring insight.

Your puny brain has no idea how much it will be enhanced by the time I get through with it.

You can thank me in 2045.

"Intelligence has diminishing marginal utility in an evolutionary sense, and, above some level may have nominal or potentially even negative effects on fecundity."

Citation? Everything I've read is that there are no diminishing returns to more g. Very smart people as a group, are much more successful in terms of life outcomes, than the merely smart. And the further our you go on the distribution the more success you see.

In evolutionary terms what's required isn't that they be "successful" but that they have lots of children who in turn have lots of children.

The selection quoted by Tyler doesn't mention that point either.

Exactly this. The human brain accounts for ~25% of our basal metabolic rate, compared to ~8% in other apes. Who says there’s no downside to maximizing intelligence?

That's like saying an investment with a great return has a "downside" because of it needs money to be invested.

Yeah, the brain accounts for 25% of metabolic rate but it also accounts for the generation of ~100% of energy needed to fuel ALL metabolic functions.

It is not like an investment, because the returns to most investments are linearly related to the sum invested, whereas the energy obtained by an organism has a nonlinear relationship to its intelligence

I see your point but I disagree. If anything, the higher the intelligence, the more exponential the return.

It is true that an organism needs only so much energy per hour/day/week etc., but it is also true that "security" is a normal good - i.e the more, the better almost without limit.

Having a lot of intelligence close to guarantees not only your metabolic needs right now but very far into the future, minimizing risk and the effect of uncertain and changing environment.

It's like having a big bank account. You surely don't need it today but it's a great buffer.

Same with muscles -guerrillas don't need that much muscle to pick grass blades but occasionally they need enormous power to defend their troop and fend off competition. Is it then very wasteful to have overwhelming power?

Depends on if the local food source can supply enough nutrients to support heavy musculature.

You need to make a distinction between two situations we might be in:

a. There's an optimal level of intelligence somewhere in the middle, and getting smarter than that is a bad evolutionary strategy, even if you can do so at no cost.

b. You get smarter by making tradeoffs with other things (bigger head, more energy used by the brain, higher probability of being crazy, whatever), so there's an optimal tradeoff point between high IQ and other things.

I don't see any reason why (a) should be true in general. You can imagine some societies or situations where it would be true, but remember we're talking about evolution over hundreds or thousands of generations--most societies don't stay the same for that long.

I think it's clear we must be in situation (b), and in fact that humans have gone way down the path toward trading off other things to get better at being smart. Thus, we spend lots of energy on our brains, we're born helpless and stay helpless for several years, our women die in childbirth because babies heads will barely fit out the birth canal, etc.


By the way, add "it might be useful to believe bullshit" to (b).

Smart people are better at producing convinving, fascinating, engaging bullshit.

Arguments like:

What I mean is that intelligence went from being subject to strong directional selection (which may have exhausted its potential over millions of years) to being subject to strong purifying selection

seem more of the "macro-economist kind" than the "scientist kind".

Similar to that old joke, an economist doesn't pick up a $100 bill lying on the ground arguing that if it were genuine someone would already have taken it.

And in his guardian article linked to, he writes: "We need to get away from thinking about intelligence as if it were a trait like milk yield in a herd of cattle, controlled by a small, persistent and dedicated bunch of genetic variants that can be selectively bred into animals from one generation to the next. It is quite the opposite – thousands of variants affect intelligence, they are constantly changing, and they affect other traits. It is not impossible for natural selection to produce populations with differences in intelligence, but these factors make it highly unlikely.'

Does this make it clear that there cannot be a mutation targeting a narrower feature like memory which is gradually selected for?

Given he does not know what causes intelligence - although it is nice he is willing to concede any genetic role at all - it is interesting to note he is sure you can't breed for it.

There is a body of work suggesting Ashkenazi Jews have been breeding for intelligence. So that would suggest otherwise.

First, let's define intelligence. ... What? No takers? What a surprise! Pseudo-academic discussions on intelligence remind me of the public discussions on "energy". What the academic (skilled in the art) uses the term for is NOT the same as what you and I are using it for. If everyone was born with an IQ of 180 (say) then what would happen to the military? Rape, murder and lying are as much survival strategies as compromise or subservience. It would be interesting to be around when A.I. is able to model intelligence to such a degree such that we can model two societies which differ only in that specific. My bet is stupid is often a good thing for a society. After all: economists, cops, and the foot soldier.

How are they using the term, then?

"What the academic (skilled in the art) uses the term for is NOT the same as what you and I are using it for."

Citation please

Citation for the need to cite, please.

He is the one making the assertion, which is counter to what most people believe.

there is a huge mutational target that natural selection has to keep an eye on.

Natural selection doesn't have an eye to keep on anything.

Well, there is the old observation that the brain is not a truth engine, it is a survival engine. I think that adequately explains why high intelligence is not common, is an outlier.

Dithering about the nature of the universe is not always the best use of an organism's time.

It also should explain why small differences between groups are pretty boring, because none of those really deal with the rareness of high intelligence.

"It also should explain why small differences between groups are pretty boring, because none of those really deal with the rareness of high intelligence."

Wrong - small differences in Mean result in huge differences in the right tail (rareness of high intelligence) and the left tail (retardation).

There is more than one anonymous on this page, so don't be confused there.

But in terms of what I wrore at 5:24, and your response, I think you have a problem in the data. We know IQs north of 150 are possible. They exist. If evolution really favors that, why aren't they all 150? That is, why aren't we all 150?

Why aren't we all better looking?

Haha, to paraphrase Yogi Berra, "If we were all 150, we wouldn't be 150."

The test is always renormed so that the mean is 100.

Try it with "the absolute intelligence we now label 150."

Basically is the ability to draft Feynman diagrams broadly necessary, and if so, why can't everyone?

As an aside to the audience, an IQ of 150 puts you in the top 0.1%


As our environment has gotten more complex, intelligence becomes more important. A subsistence farmer can improve his and his families well being by noticing patterns and experimenting with different seeds, etc.

This issue has become more important as of late for a number of reasons. Didn't Ferguson have an institutionalized system where they preyed upon the less 'intelligent' or competent? You go after less effective and organized people and bleed them dry. And all this in an effort to keep a community with no resources to deal with the increased complexity of the economy, where intelligence gains a disproportional advantage.

If you structure a society where people below an iq threshold have no place, and year by year increase that threshold by increasing the complexity of almost every necessary task, you will have more and more people fall off the edge. Those types of societies don't last very long.

The first signs of catastrophe will always show up among communities the least able for various reasons to respond. And the immediate reaction from both sides of the political ideological spectrum is to call these people stupid. the far right call the blacks stupid, the moderate right to extreme left call the white Trump voters stupid.

You were an engineer right? You built into your designs a margin, the extent of the margin depending on the expected effects of failure.

I think we have hit the design margins of an effective and working society. Communities with no design margin for various reasons are failing, in some cases catastrophically.

I don't know the answer except that every bit of information is critical to figuring something out.

That is nothing more than your subjective demand for a certain kind of intelligence.

What if, instead, I just want people to be happy? Or good Christians?

What you want doesn't matter

I think Cowen would suggest that Mormonism is the answer (Orthodox Judaism probably is also a good a substitute). I believe that Steve Sailer called the idea "simple rules for simple folks". However, for that to take hold widespread in society, the upper middle class and elites would have to start "preaching what they practice," to borrow a phrase from Charles Murray.

Here are some other ideas. There probably are large gains available if the government were to try and decrease crime and incarceration while simultaneously trying to prop up the incomes and employment of non-college educated men. High taxes on alcohol (and banning the sale of marijuana while still making it legal to possess, use, and grow a small amount for personal use of friends and family) might help as well. Free long acting contraception also would likely help with family stability and parental investment by delaying the age at which non-college educated women have children (they are likely to have higher earnings the older they are when they first start having children). I suspect that free long acting contraception would also decrease the incidence of men and women who have children with more than one partner, which itself would make those men and women much more marriageable. Lead abatement is an obvious investment to make, as are classes teaching parenting and other skills to low income mothers and fathers.

Excellent comment. There's another pithy saying, "Yale or jail." The nurturists believe the fundamental problem is not enough educational opportunity so disadvantaged groups can become Einsteins, so the option of ending formal education at age 16 (by which a lot of people are frustrated and bored to tears with schooling) and apprenticing at a trade is off the table. Yale or jail.

The goal is to give everybody a dignified life commensurate with their abilities, not a society where the non-degreed are pariahs.

If you start from the premise that everybody is equally educable, then in an equal-opportunity society being a prole must stem from some conscious flaw of character. This warps policy debate in all sorts of ways. Thus, prole whites given their "privileged" status, must be morally lacking. Immigrants are fungible; it doesn't matter whether we import German bourgeois or Somali Muslims. And on and on.

I think it should be said that if we had policy aimed towards getting most everyone a dignified life commensurate with their abilities, what Somali refugees we did accept would have a place to find that kind of a dignified life too.

Mitchell's argument is that all environments select and have always selected relentlessly for higher intelligence. This is why only the top 5% or so in IQ has children in any given society, LOL.

In a certain way, I feel bad for having inherited my intelligence while so many nice people have to go through life with second-hand brains or worse. I mean, it is not my fault, it is the system. Don't hate the player, hate the game and all that. But sometimes I can't sleep.

I feel your pain.

The problem, as I see it, isn't "nature" or "nurture." It's the educational system fills, or clogs, students' varying, say, IQ capacities with utter garbage.

I mean, I did nothing to inherit my intelligence. I just chose good parents. Maybe one shouldn't be more intelligent than other people just because one came from the right womb.

Thiago - there's an easy way to correct this injustice. Eat some lead.

Yes, but are our moral instincts about envy or about efectivelly helping other people? Do we want to hurt the well-off or we want to help the downtrodden? Or do we just want to signal virtue?

I mean, would becoming stupid help stupid people? Would society be better off without intelligent people? I don't think so. Maybe intelligence unequality is an inescapable part of human condition and soaking the intelligent or killing the intelligent to eat their brains won't help. No matter what the government does, the gap in intellectual performance between intelligent people and stupid people remains. Maybe it is how things are supposed to be.

More accurately, your parents selected good breeding partners. "Marry well" is a good credo for a society. Culture is the feedback loop for genetics.

"Second hand"? Who was the surgeon?

I meant battered, useless brains, but, yes, "second-rate brains" would have been a better choice.

It sounds like a Monty Python sketch. John Cleese walks into a fish and chip shop. Behind the counter is Eric Idle.

JC: "Battered brains and chips, please, my good fellow."

EI, in oily voice: "Certainly, sir. Cow brains, lamb brains, pig brains, sheep brains?" Pause. JC is about to answer when IE continues: "Salmon brains, trout brains, herring brains, haddock brains." Pause again. JC is about to answer when IE continues: "Penguin brains, albatross brains, eagle brains, heron brains." And again: "Kangaroo brains, wallaby brains, koala brains, platypus brains".

JC cuts in: "Whitebait brains, please."

IE: "Certainly, sir. Beer batter, garlic batter, tomato batter, or weasel batter."

Graham Chapman walks in and says " Stop there, just stop there. This is getting really silly." And now for something completely different.

This is a dumb model. But let's say that individuals reproduce asexually (cloning). And let's say that IQ is the sum of N variables that can take a value of either 1 or 0 but are flipped from 0 to 1 or 1 to 0 with a probability P. Then it follows that there there can be individuals who can expect there descendants to become dumber. If the IQ of the parent / original is denoted by IQ_t and the IQ of the descendant / copy is denoted by IQ_{t+1}, then there exists an IQ_t for which it holds that E[IQ_{t+1}] 0, it follows that E[IQ_{t+1}] < IQ_t.

That's how I'd interpret the claim.

Apologies for the typos.

For some reason a proof was omitted.

E.g. Let's say that IQ_t = N, then E[IQ_{t+1}] < IQ_t.

Your model doesn't have to assume asexual breeding. It just has to stipulate that people who reproduce do so with people of similar intelligence.

Indeed, children of parents with average intelligence are sometimes less intelligent than their parents. As often as you'd expect ...

You better campaign for burkas.

I know that I don't. It's a simplifying assumption that didn't appear crucial to me.

This is bad genetics, plain and simple.

First, evolution is NOT goal oriented. It cares not a whit whether offspring get smarter, fatter, taller, or uglier. Only that they produce offspring.

Second, most mutations are what we call neutral mutations. They have zero effect, thus they are not selection targets. Of the non-neutral mutations, most will be lethal. The remaining non neutral, non lethal mutations are available as selection targets. Given the quantitative nature of complex traits, a single mutation is likely only able to account for a small fraction of the variability of the trait. Probably 2-5%. So it is basically irrelevant. Only under strong selective pressure for a trait will you see any effect over time. Otherwise, the mutations proportions generally vary randomly in the population until it either becomes fixed it disappears by random chance. This is ignoring things like linkage disequilibrium that can alter the probabilities.

Third, this is NOT different from other traits. Plant breeding is basically all quantitative genetics. Yield, plant height, stalk strength, disease resistance, etc are work rare exceptions all controlled by many genes. Yet breeders have made a career exploiting mutations to create the modern plant breeds. In on man's lifetime yields have more than doubled.

Most non-neutral mutations are lethal? That doesn't seem right, either. Don't they just make you marginally dumber/slower/shorter/whatever? Isn't that what's implied in the term genetic load?

When dealing with human evolution we also need to remember the humans are social animals which means that individuals have resources (often enormous ones) outside themselves they can call on for survival. It isn't just "me against nature". Some mutations are so lethal that the social resources just won't matter and the individual is doomed. But even in the Old Stone Age people could survive with less fatal mutations because their tribe helped them.

Genetic load is simply unfavorable mutations that reduce fitness. Reduced fitness is rarely lethal, even in non social organisms.

Look at it this way, most organisms have been around long enough that a mutation at a site isalmost guaranteed to have occurred in the past. If you can't find one, it is generally safe to assume it is highly selected against and possibly immediately lethal. Given the high homology between gene classes between species, I think it's safe to assume most are lethal at least in the long run.

Also keep in mind that genetics is always confounded by environment. The GxE interaction is very important. A very deleterious Gene mutant in one environment may be necessary for survival in another. Generic load is the accumulation of these in organisms, but is often very useful for species survival under extreme duress. Removing diversity may increase many individual traits fitness and increase narrow sense fitness while increasing the long run odds of species Extinction.

Mitchell: "I recognise that the conceptual framework presented here is unorthodox and people will no doubt take issue with various points. The overall point is that all this stuff is complex and much of it is unsettled. At the very least, it is therefore important to consider the parameters of polygenicity, mutational target, pleiotropy, mutational load, and general relationship to fitness – as well as how they all interact – in the discussion of whether it is in fact likely – a priori – that systematic genetic differences in intelligence would arise between ancient population groups. (As opposed to simply asserting that any trait that is heritable is subject to directional selection, as if these other factors are irrelevant)."

Of course, complexity is not what most people can handle, partly for the reason that they lack the intelligence to grasp the complex factors must less understand them. I thought Mitchell's discussion of natural selection interesting, especially the part where natural selection inhibited greater intelligence rather than expanding it. Thus, for example, larger brains might contribute to greater intelligence but requires a larger head to house the larger brain, the larger head making childbirth more difficult and inhibiting some physical attributes. Like Mitchell says, it's complex, much more complex than the simple assertion that there are genetic differences in intelligence between races.

It all boils down to "why don't bunnies build space ships?"

From there, "why don't average people, from any 'group,' build spaceships?" isn't a heavy lift.

Maybe I should have said raccoons. Raccoons are dexterous enough to build a spaceship. And yet Across the Known Universe there is only one raccoon, in legend or fable, who has managed it. Long odds.

"This stuff is complex and much of it is unsettled" was certainly not the message Mitchell put forward in his original Guardian piece where he asserted that there are unlikely to be heritable race differences in IQ. He developed an appreciation for the complexity of the question when various population geneticists and others slammed his Guardian article as simplistic and uninformed.

I stopped reading after "Guilty as charged"

Yes, this piece seems to basically retract the original Guardian piece (which I saw previously). He must have gotten a lot of pushback from colleagues. Makes me wonder what I wondered when I read Reich's book: how long will major universities continue to permit a discipline which yields such politically incorrect results?

y81 says:"Makes me wonder what I wondered when I read Reich's book: how long will major universities continue to permit a discipline which yields such politically incorrect results?"

In 2018, as long as major authors contributing to said discipline are writing New York Times Bestsellers that contribute to the understanding the discipline.

One very big problem here is the semantics, beginning with "intelligence". Of course one prefers more rather than less "intelligence", right? In today's complex world, the ability to multi-task is associated with greater "intelligence"; indeed, I can picture all those boy wonders in Silicon Valley spending their days multi-tasking. Admission: I cannot multi-task, so I must not be as "intelligent" as those boy wonders, right? But multi-tasking has its downside, namely the ability to focus on a single task. I may not be able to multi-task, but I can focus like a laser. Would natural selection, therefore, try to find the sweet spot between the ability to multi-task and the ability to focus, giving up a little "intelligence" in the process?

There is no trade-off between intelligence and ability to focus. Quit the opposite

But both collapse when a girl in a bikini walks by.

Monkeys that we are.

Which is useful, because otherwise we'd go extinct.

I doubt the boy wonders in Silicon Valley are characterized by flitting from one task to another. Some have described how long it takes to get into the headspace where they can work; they have to rebuild the models in their head before they can commence working on it, and that takes time.

I suspect that is exactly what you do when dealing with something complex.

'In today's complex world, the ability to multi-task is associated with greater "intelligence"'

This is the first time I have seen or heard this asserted. I am pretty sure you are confusing "intelligence" with "managerial skills" or even simply "highly compensated", an understandable mistake given that managers are paid, often highly, for managing rather than taking standardized tests.

It seems there is an ideological axe to grind here. I am profoundly uninterested in that angle.

Some animals breed for intelligence every cycle, and humans have bred dogs for intelligence. Working breeds have characteristics that are inherited that can be described as intelligence. There are specialized traits such as the herding or hunting, and more general intelligence with guardian dogs and useful farm breeds that could herd, hunt, guard, etc. As the more showy characteristics were targeted in breeding, the intelligence disappeared.

I have watched Osprey raise their young in the late summer. There is no single trait that determines survival, but intelligence would be one of them. They deal with multiple inputs in a constantly changing environment, and some young do it quicker than others. It is quite remarkable; they hatch mid July, by the end of August they are physically mature and begin flying. By the third week of September the adults head south and the young survivors feed up on the last Kokanee spawns then are gone by the first week of October. Over that time they are constantly practising the skills of survival. Many don't.

It seems there are measurable differences in IQ by community. What that means is difficult to determine; we can by nurture make people stupid, maybe a little the other way, but the raw material needs to be there. IQ has increased over the last few generations in the west; is that from breeding or from better nutrition? The kids who were head and shoulders taller than their parents in Quebec when I grew up likely had higher IQ as well; they never missed a meal and had vegetables all winter.

If there is a variability, then policy makers would best focus on making sure that through their policies they don't complicate people's existence any more than absolutely necessary. I have yet to see any politician do that, so if there is a difference in IQ between races, and it disadvantages certain races that are also economically disadvantaged, then those pasty white bureaucrats writing incomprehensible regulations are the most racist and disgusting people imaginable.

“those pasty white bureaucrats writing incomprehensible regulations are the most racist and disgusting people imaginable.”

Yes, you don’t even have to make this statement conditional

"Bred for Intelligence" generally means "bred for a trait that we associate with intelligence" like the ability to understand and follow simply verbal commands.

"like the ability to understand and follow simply verbal commands."

No, that would be for run-of-the-mill type of dog. In practice, breeding for intelligence would result in an dog being able to associate a thousand objects with their individual names or ideally, the individual names of a hundred sheep or cattle, or goats and retrieve it on demand and with enthusiasm.


Different dogs have different levels of ability to do those kinds of things. Some dogs only understand a small number of simple commands, other have a much larger repertoire of things they understand and will perform. That's what we mean by "canine intelligence"-- doing better and doing more in regards to dog-normal things. It's the same with humans: human high intelligence is simply being better at human-normal mental tasks. Get back to me when dogs develop the ability to produce actual human language and debate abstraction-- or humans develop the ability to foresee and manipulate the future.

Excellent comment. Thread winner.

about heritability and mutations: https://www.bipartisanalliance.com/2018/05/age-of-fathers-mutation-and-reproduction.html

Much writing in intelligence in a rag like the Guardian is just dishonest or ignorant claptrap. To give this chap his due he is far above that sort of stuff. He does acknowledge that: (i) intelligence is a thing, (ii) IQ ... tests do measure something real that relates to intelligence, (iii) IQ scores really are heritable.

Moreover in reviewing some prominent writers he says "The positions of ... thus state that we should, by default, expect there to be genetic differences affecting IQ (indeed, all traits) between populations that were genetically isolated from each other for long periods." That is a calm and well considered resume of their views.

This all seems to me to be a fine platform from which to advance his arguments which, he frankly admits, are based on his own opinions.

I don't think he'll prove to be right, mind.

> For intelligence, following my argument above, the goal is to hit the maximal level possible.

This doesn't seem right. Intelligence is only desirable (from an evolutionary perspective) insofar as it lets the person with high IQ reproduce at higher rates and have offspring who are also successful in reproducing. But in my experience the correlation between ease of finding partners and intelligence is only positive up to a point.

Your comment ignores the reality that humans are social animals. Even if an intelligent person has few or no children, if he enables his larger kin group to prosper and reproduce more there will be more genes of his sort in the future and the probability of people like him being born by the shuffling of the genetic deck which includes those "intelligence genes" whether they are manifest in every individual or not.

That's a good point, and one I didn't think of, but I still don't think it implies intelligence should be maximized (by natural selection). I think you hit a limit beyond which intelligence isn't much help for survival. Are the people in society who make the most important decisions the smartest? No. The smartest people (in terms of pure IQ) are math and physics PhDs and they're largely irrelevant in regards to any real world decision making.

Over long enough time periods what I outline would mean that populations with more intelligence-producing genes would survive and increase, even if those genes were not realized at their most active settings in most individuals in those populations.

If it is true that intelligence is a massively polygenic trait, would it also be likely that genes that impact intelligence also impact other traits as well? I am thinking that if there are genes impacting both intelligence and other traits, that we would perhaps see differences in mean IQ (or test scores) in populations of people whose ancestors were genetically isolated from each other, because selection pressures on traits that we wouldn't immediately recognize as related to intelligence would in fact actually have somewhat of an impact.

yes it is true. but it is also true IQ is not intelligence - there is no doubt correlation, but IQ changes over time and tests scores need to be re-centered (average = 100). Now if IQ were purely genetic, this would not happen at the rate test scores go up. There are some confounding variables in there.

So is it likely that genes that impact intelligence (or g, or IQ, or whatever) also impact or influence other traits?

Maybe. Foxes bred for domesticity start to get floppy ears and curled tails.

Just replace "maximal" with "optimal" - problem fixed!

I didn't see any discussion of sexual selection in the thread yet. I think it is important when it comes to intelligence. A la peacock's feathers phenomenon ...

Suppose that the commonly-invoked argument that "selection is just not as strong nowadays in order to survive and reproduce" is correct in general. There still is a lot of room left for assortative mating based on preferences for high intelligence (even frivolous or downright counterproductive preferences - e.g. risk of autism) especially among some groups. This alone will ensure stable, if not increasing, intelligence over time (at least for said groups).

Children of older fathers have more mutations than the children of younger fathers, about 2-3 times as many. By studying how full siblings varry in IQ based on their father's age at conception, we can establish the IQ reduction associated with a single generation of new mutations. What is it?

0. More precisely, it's somewhere north of 0, but too small to be detectable by any of the studies.

How many IQ points can a generation gain with selection for IQ? Plug the numbers into the breeder's equation for yourself.

It's demonstrably false to claim you can't select for IQ within the nautural human range. Maybe you couldn't go to 300, but it would be easy to generate group differences as large as we see today.

He speculates that all humans have already been selected for "maximum" intelligence. But clearly we aren't all Isaac Newton so .... I guess he means 100 is the "maximum group mean"? But that is oxymoronic. The average won't be the maximum unless there's no variation. As long as there's individual variation, you can increase the group mean by killing a disproportionate number of dumb people. Or is there some group selection feedback mechanism that always magically drives the average back to 100?

He is leaving out the second part of evolution - natural selection. Traits that increase fitness will be disproportionately passed on. Mutation is random, natural selection is not. The author seems to know what he is talking about, maybe he has focused too much on the brain and has forgotten some basic biology.

He's trying to argue that intelligence is uniquely exempt from selection. This is absurd on its face, so he tries to salvage it by arguing that it was *so* useful that all humans everywhere were selected for "maximum" intelligence with a single global ceiling. Implicitly, he's saying "human intelligence" was a one-time binary switch that affected all of humanity equally.

These desperate arguments are the kind of thing people resort to when they aren't willing to abandon a bad position.

The main difference between intelligence and physical traits is that it's easier to bullshit about intelligence. And that is what this guy is doing.

Isn't it rather that IQ / intelligence is now so much more important in describing poverty differences that, for political reasons, the truth about IQ heritability must be denied. The truth is not PC.

the idea that such pressures would align with continental divisions remains inherently implausible in itself, to my mind.

It seems the Mitchell is trying to create a theory that, a priori discounts measured IQ differences in different US ethnic groups. The facts show measured IQ differences are real, which seems to already falsify his theory. Much like bees flying falsified the "physics theory which proves bees can't fly" -- which many normal folk learned. Frisbee like extra edge turbulence lift was missing from the early physics calculations.

His theory might be mostly right but is certainly incomplete because the actual data falsify it. Whether it's IQ tests or some other measure of g.

The higher IQs of Jews, 95% or so of the men having Abraham's Y chromosome (the only one that is inherited in a line), also indicates that intelligence could be bred for, just as much as rat maze-running could be bred for.

I believe it WAS bred for among Jews, and that smarter Jews were higher status and had more offspring that survived. Especially in cities. And I'm now suspecting that complex city life has more often bred for intelligence over rural life, and especially over hunter-gatherer life.

Also, with the strong pro-abortion, anti-family prejudices of so many mostly city-oriented Feminists, I suspect city women of having fewer children than rural, family oriented women. So smart women being childless is tending to, in the present US situation, promote less smart women having more children, and winning the evolution race.

(i) Abraham didn't exist: he's a character in fairy tale.
(ii) I don't think many people argue that Jews are particularly clever - it's specifically Ashkenazi Jews who do indeed appear to be.
(iii) One explanation, favoured by Greg Cochrane, is that it's the result of selection pressures on European Jews over the last thousand years or so. He point outs that nobody in antiquity thought Jews clever. Greeks, yes; Jews, no.

OK reading this and the comments too fast, let me rephrase the ideas here:

1. Intelligence is impacted by many genes, it is harder for evolution to keep an eye on lots of genes versus just one.

2. Diminishing returns implies a convergence of intelligence...at least from genes. If you have some group that has been breed to have high IQ's, that population is actually at an evolutionary disadvantage. They have a lot less freedom to have a mutation in a few random genes hence the high IQ cannot be maintained or a lot of energy must be spent policing a lot of genes for consistency.

2.1 Diminishing returns cuts off IQ advances. The cost of more IQ via genes goes up fast and the rewards that you can get for that additional IQ start to diminish fast.

3. The question then isn't whether there are significant IQ gaps between groups but whether it makes sense that they are large and caused by genetics. It would seem the genetic side of that answer is no.

4. It's not at all clear IQ is an evolutionary win. For many species, a dumb brain that works on rules of thumb takes less energy and works enough for evolutionary success in the billions of years. Human intelligence only seems to have appeared once and odds are for for many species the gains for additional IQ was not worth the additional energy needed to feed and maintain an energy hungry big brain.

5. On the human level IQ, I suspect, is a mixed bag. A high IQ may in many cases mean a person can think himself into a lot of unnecessary trouble. Is being able to reinvent a wheel very helpful to the economy when we already know about wheels?

Your comments are pretty weak beer, Boonton. Did you read any of the more penetrating comments above or did you just jump in?

Might be censored by now:


Was supposed to be a response to Boonton.

Tried to go to it but net blocked the URL as possibly malicious. Could you summarize the arguments you found worthwhile that it presented in respect to my comment?

I doubt there are any malicious scripts running there, but I bet some blacklists may have the site listed as a "hate site".

The key argument is that a country's GDP can be predicted from average IQ, the hypothesis being that those with IQ of 106+ are a limiting factor in bringing up functional businesses and governmental institutions.

After reading your post with better care, I realize there is no actual contradiction between "smart fraction theory" and your argument, you claim is that above a certain level, IQ ceases to have an evolutionary advantage. This is not contradicted by a claim that a country's GDP/capita increases exponentially with the fraction of people with IQ above 106.

OK why would average IQ determine a country's GDP? Perhaps it's the other way around. A higher GDP demands the development of and use of IQ. Did the USSR dominate Chess because Russians have some mysterious 'chess gene' or did they dominate it because their culture was set up in a way to reward and cultivate chess playing while our culture may cultivate professional basketball and football players.

My understanding of Murray's whole IQ argument is that in theory even if you evened out all environmental factors, you would still find IQ variation presumably caused by inheritance. But no one has evened out the environment and a lot of IQ points remain to be added by environment.

Now personally I'm skeptical of this IQ=GDP argument. First it doesn't make sense in relation to diminishing returns. Second it seems to me a lot of innovations actually facilitate production via lower IQ and knowledge. The assembly line meant you didn't need a skilled craftsman but a man to do repetitive motions that could be picked up by a high school dropout in less than a day's training. Today an uber driver can navigate anywhere whereas before the veteran cabbie had decades knowledge of the city streets with all the tricks to avoid traffic and problems.

It seems to me the production function is geared towards rewarding not ever more esoteric ideas an inventions but removing IQ as a barrier. Rather than have people ring up the price of meals in their head, or memorize all the menu prices, just put pictures of the food on the register and let them push what the customer ordered.

In the future I'm even less sure IQ matters. With AI I suspect the pseudo-monopoly rents high IQ professionals collect are in danger just as a 'professional chess player' doesn't seem like a very stable job after IBM Watson.

Regression towards mean. That should explain all to the fellow economists reading.

"But for some traits, the “goal” of that program is to hit a species-optimal set point."

That's a ridiculous statement. What is optimal can be radically different in different times and places. Natural selection does not operate at the species level in any meaningful sense of the term.

Intelligence can vary between groups within a species. Within selection-imposed limits, and with consequences.

On the upside...



"The moral of the story is: populations can have genetic diseases if they also provide a useful advantage to carriers. And if those genetic diseases are limited to a single group, we expect them to provide a useful advantage for that group, but not others. Might the Jewish genetic diseases provide some advantage? And why would that advantage be limited to Jews?

Most of the Jewish genetic diseases cluster into two biological systems – the sphingolipid system and the DNA repair system. This is suspicious. It suggests that they’re not just random. They’re doing something specific. Both of these systems are related to neural growth and neural branching. Might they be doing something to the brain?

Gaucher’s disease, one of the Ashkenazi genetic diseases, appears to increase IQ. CHH obtained a list of all of the Gaucher’s patients in Israel. They were about 15 times more likely than the Israeli average to be in high-IQ occupations like scientist or engineer; CHH calculate the probability that this is a coincidence to be 4×10^-19."


On the downside...

Hanson: Darwin’s Unfinished Symphony


"Laland’s research group has for decades studied the origins of human cultural evolution. They’ve learned a lot. In particular they attribute humanity’s unique ability to accumulate culture over a long time to our very high reliability in transferring practices. Humans achieve such high reliability both by being smart, and by our unusual ability to teach, i.e., changing our behavior to make it easier for others to copy our practices. Just how high a reliability is required is shown by the example of Tasmania, where several thousand isolated humans slowly lost many skills and tools over thousands of years. It seems even human level intelligence and teaching isn’t good enough if your population is only a few thousand."


No limits?

Super-Intelligent Humans Are Coming



Yes limits?

Chollet: The impossibility of intelligence explosion


The fact that IQ is so incendiary a topic among laypeople gives a clue as to why we should be wary and aware of underlying agendas.
Laypeople like to assume IQ is some summary judgment about people's overall rank and ability like some Darwinian super trait that determines success.
Which would require us to ignore everything we know about human nature and history.
Successful outcomes in both the personal and societal level depend on an optimal balance of many complex factors of which raw mental computing power is a small part.
It is asserted for example that IQ correlates with success in business. But so do a lot of things, like being tall, or being attractive or gregarious with good social skills.
The assumption that we should desire higher IQ either on an individual or societal level isn't well supported.

It is pretty well supported. See Garrett Jones Hive Mind. And on an individual level, IQ is the #1 predictor of success in most areas, I think you would be hard-pressed to find someone who wanted to be less smart.

We don't see this empirically.
The Person with the highest IQ doesn't always win the job, or contract.
There are obviously many other factors to success.

The thing is, most things depend on a large number of traits, not just one thing.

Variation in IQ accounts for something like 25-40% of variation in income, depending on the particular study.

That isn't everything, but it would still be the single largest factor.

All I know is that when I get the early version of brain augmentation in 2029, I better be converging with the augmentation Robin Hanson gets, not diverging even further away. The latter would suck.

It's all relative.

As a disclaimer, I am not a geneticist or social scientist. A single IQ test seems to be a poor measurement of intelligence. Rigor? Quality control for repeatability and reproducibility? How do you calibrate for intelligence? Are results comparable between independent tests? How are outliers treated? Is normal distribution the right model?

Mitchell writes, “In particular, if there is a lot of variation in environmental factors that affect a trait within a given population, then the heritability will be lower, because proportionally more of the variance in phenotype will be due to environmental variation.” Then I would conclude that if variance is low then heritability should rise. Education, nutrition, environment, security, awareness of health factors have greatly improved in my short lifetime. Access to knowledge has never been more accessible. I just perused a discussion between geneticists. How cool is that? My intuition is that we are enhancing IQ in a more collective manner. This is where the median’s actual proximity to the mean would be instructive.

A devils advocate would say:

White parents provide good intelligence training and black parents good athletic training.

Majority white schools get all the good teachers and majority black schools get all the good coaches.

Rich people are able to deliver trains for their children that overcomes the natural tendency for their children to be less motivated to make more money because the "can rely on the old man's money".

Black parents and majority black schools must be better at creating an environment conducive to creating in music and helping people in an emergency.

Does that explain most of it?

I hope he is right on racial differences, that would be a good thing.

Does this mean CRISPR technology is unlikely to make babies smarter anytime soon.

The idea that intelligence has been maximized is extremely dubious for several reasons.

The first, and largest, reason is that we see a ton of variation in intelligence within human populations, and it is very significant – higher IQ has a big positive influence on your success in life in modern society, and there’s little reason to believe it did not have a large influence on success in more primitive societies. The biggest reason is that higher IQ makes you better able to deal with social situations, as well as learn abilities faster and more quickly and deal with more varied situations more easily.

The fact that we see major differences in how well people deal in modern-day society, and thus, see a big difference in IQ, despite the fact that IQ is such a positive trait, suggests that IQ is nowhere near maximized.

This is true both within and between groups – there is quite a lot of variation within and between all groups we’ve measured IQ for, though I’m not sure if there are large samples for some populations.

The second issue is that the supposed limitations on intelligence themselves can evolve. Humans have evolved to give birth to ever-larger headed babies, and babies with ever-longer vulnerable periods. And the thing is, these traits, too, could well vary between groups. Indeed, we know that there are differences in average skull sizes – Asians have slightly larger skulls than whites do, with whites having considerably larger skulls than blacks. Aboriginal Australians, according to at least one study, may have the smallest skulls of all. The fact that there are these differences could well suggest evolution has acted to allow white and Asian mothers to give birth to larger-skulled babies. Indeed, the fact that black women in the US have higher maternal mortality rates might well be due to these adaptations being absent from them – black women in the US are actually generally black-white crossbreeds, and they may not have all inherited the traits necessary to give birth to larger-skulled babies that come with having white genes. Or maybe black populations in general have higher maternal mortality rates for women, resulting in a decreased number of large-skulled babies relative to white and Asian populations – indeed, sub-Saharan African countries have the worst maternal mortality rates in the world. And while these countries are indeed very poor, it is possible that this could in part be a biological difference as well. Incomplete evolution of secondary required traits could prevent upwards motion on baby head size.

Another issue is rate of maturation. And we know from a number of studies that in terms of physical maturity, black children mature faster than white children. This is, again, a potential point of divergence – longer maturation = longer period of time to grow that bigger, more developed brain.

The fact that these secondary traits vary across population suggests that there are several axes by which a “maximized” intelligence could further increase in some populations and not others, because of various secondary mutations which enable for the primary trait (intelligence) to shift further upwards.

And indeed, we’ve seen evolution select for taller humans even over historical timespans. We know that height positively correlates with intelligence, and it is entirely possible that the increased height not only increased brain size, but also led to larger women who could give birth to larger children.

As for the dysgenic argument – I think this is fatally flawed .The problem is that human intelligence has already gone up a bunch. If being a polygenic trait prevented this, then it would have happened long before now. There’s no reason to believe that we have coincidentally achieved some sort of maximum; this downward pressure would have always existed on this trait, as it would on any other polygenic trait. And indeed, we see from another polygenic trait (height) that over time, members of a genus tend to get larger across many, many kinds of animals (Cope’s Rule) – which would indicate that such polygenic traits can indeed increase despite deleterious mutations trying to pull them down.

None of this means that there are necessarily genetic differences driving differences in intelligence between human population groups - we likely won't know the answer to that until we understand human genetics. But the argument he put forward is extremely untenable in my eyes.

Comments for this post are closed