When half of a quarter is all, or at least mostly

« previous post | next post »

Steve Connor, "Nature rather than nurture governs intelligent behaviour in primates, scientists discover", The Independent 7/10/2014:

The vexed question of whether intelligence is inherited from birth or acquired through education seems to have been answered – for chimpanzees at least.

Scientists have found that being a smart primate is down to genes rather than upbringing, suggesting that nature rather than nurture governs intelligent behaviour in our closest living relatives.

The next sentence backs off from "down to genes rather than upbringing" to "mostly governed by its genes rather than its environmental background":

Researchers have argued for many decades over the genetic basis of human intelligence – based on IQ tests and studies of identical twins reared apart – but now a study has found that a chimpanzee’s “cognitive ability” is mostly governed by its genes rather than its environmental background.

And later in the same article, it becomes "about half of the variation in this ability is down to genetic factors":

The study involved behavioural tests on 99 captive chimps aged between 9 and 54 years. The performance of each ape was measured on a series of standardised cognitive tests for primates and the researchers concluded that about half of the variation in this ability is down to genetic factors – about the same or a little less than humans – and half down to non-genetic influences. […]

“We found that some but not all cognitive traits were significantly heritable in chimpanzees,” the researchers said in their scientific paper.

If we read the research report under discussion– William Hopkins et al., "Chimpanzee Intelligence Is Heritable", Current Biology I[n Press, July 2014] — we find the estimated genetic role diminishing somewhat further.

The authors gave 99 chimps 13 tasks from the Primate Cognition Test Battery, and used principal components analysis to select four components which collectively accounted for 54.2% of the variance in test battery scores:

They then used "Sequential Oligogenic Linkage Analysis Routines" (background here) to estimate the heritability of the PCA components:

Thus Component 1 (23.6% of test variance) was significantly heritable — h2 = 0.538. The symbol  h2  is used to denote "narrow-sense heritability", which is the ratio between the variance due to average effects of alleles, and the phenotypic variance as a whole:

$$h^2 = \frac{Var(A)}{Var(P)}$$

In other words, about half of the variance in a PCA component accounting for about a quarter of the variance in test results was accounted for by genetic variation.

Component 3 (10.8% of test variance) was also significantly heritable, with h2  = 0.335. Thus about a third of the variance in a PCA component accounting for about a tenth of the variance in test results was accounted for by genetic variation.

The genetic relationships of components 2 (11.7 of test-score variance) and 4 (8.2% of test-score variance) were not statistically significant.

Other media coverage pushes the Nature-Over-Nurture meme even harder. Thus Kukil Bora, "Chimpanzees’ Intelligence Hugely Dependent On Genes They Inherit From Parents", International Business Times 7/11/2014:

The level of intelligence in chimpanzees, just like in humans, varies from one individual to another, and the differences in their intellect are attributed to the genes they inherit from their parents, according to a new study.

And a second article from the same publication — Lydia Smith, "Nature Over Nurture: Chimpanzees' Intelligence Comes from Genes not Upbringing", International Business Times 7/11/2014:

Chimpanzees' intelligence is down to genes rather than upbringing, which suggests that nature rather than nurture governs smart behaviour in our closest living cousins.  

Scientists have long pondered the basis of human intelligence, based on IQ tests and studies of identical twins brought up apart, but a new study has found that the "cognitive ability" of primates is dependent on inheritance rather than their environment.

Or Rober Preidt, "Like Humans, Chimps' Smarts May Rely on Genes", US News and World Report 7/10/2014:

Nature, not nurture, may play the bigger role in the innate intelligence of individual chimpanzees, a new study finds.

Or "Genes More Important than Environment in Determining Intelligence in Chimpanzees", Nature World News 7/11/2014 (with this cute picture):

Genes, not environment, are important in determining intelligence quotient of chimpanzees, Georgia State University researchers find.

The role of genes in human intelligence has been the subject of several research papers. In humans, environmental factors such as education, socioeconomic status are known to affect intelligence quotient.

The latest study shows that in chimpanzee, which is closely related to human, genes play a far more important role than the environment. Simply put: when it comes to chimps, smartness runs in the family!

At this point, you might want to remind yourself that "genes play a far more important role than the environment" is a translation of "about half of the variance in a PCA component accounting for about a quarter of the variance in test results was accounted for by genetic variation".

Also, "Chimpanzee intelligence largely determined by genes: study", Xinhua News 7/11/2014:

Some chimpanzees are smarter than others, and their intelligence differences may largely be explained by the genes they inherit, a U.S. study said Thursday.

And Mark Prigg, "Smarter than the average ape: Chimpanzee intelligence is determined by their genes not their environment, researchers say", Daily Mail 7/10/2014:

A chimpanzee’s intelligence is largely determined by the genes they inherit from their parents, reveals a new study.

"Chimpanzee intelligence determined by genes", phys.org 7/10/2014; "Nature is more important than nurture in chimpanzee intelligence, say researchers", Daily Digest News 7/10/2014; "Chimpanzee Intelligence Depends on Genes, Finds New Study", sci-news.com 7/11/2014; "Chimpanzee IQ Mostly Determined By Genes", Brevard Times 7/11/2014; "Chimps' Intelligence is Determined by their Genes", Science World Report 7/11/2014; etc. etc. etc.

If you're interested in the rather vexed logical, statistical, and political questions involved in this discussion, you should read Cosma Shalizi, "…In Different Voices", 6/22/2007; "Those voices again", 6/24/2007; "Yet More on the Heritability and Malleability of IQ", 9/27/2007; "g, a Statistical Myth", 10/18/2007.

Cosma starts by discussing the fact that among humans, accent (as in how you, like, talk) is indubitably heritable; later he consider the fact that zip codes (or more generally, latitude and longitude) are also heritable; and he closes with a discussion of what you're really doing when you apply principal components analysis to psychological test results.

The end of the next-to-last post in the list above:

Do I really believe that the heritability of IQ is zero? Well, I hope by this point I've persuaded you that's not a well-posed question. What I hope you really want to ask is something like: Do I think there are currently any genetic variations which, holding environment fixed to within some reasonable norms for prosperous, democratic, industrial or post-industrial societies, would tend to lead to differences in IQ? There my answer is "yes, of course". I've mentioned phenylketonuria and hypothyroidism already, and many other in-born errors of metabolism also lead to cognitive deficits, including lower IQ, at least in certain environments. More interestingly, conditions like Williams's Syndrome, Downs's Syndrome, etc., are genetically caused, and lead to reasonably predictable patterns of cognitive deficits, affecting different abilities in different ways. In many of these cases, it seems very likely (but is not yet established) that these variants cause problems with the signaling pathways which set how gene expression responds to environmental cues. Manipulating those signaling pathways during the right time windows would change what kind of mind the organism has later. The fact that different genetic disorders lead to different patterns of cognitive deficits, rather than just generally making people duller all around, suggests ways of disentangling which genes are relevant to which abilities through which molecular mechanisms. (Cf.) At a popular level, I've still not run across a better description of way the regulation of gene expression couples genotypes and environments during mental development than Gary Marcus's writings, but if you want details there is a whole rapidly-growing field of molecular developmental neurobiology (as I'm not-infrequently reminded).

I suspect this answer will still not satisfy some people, who really want to know about differences between people who do not have significant developmental disorders. Here, my honest answer would be that I presently have no evidence one way or the other. If you put a gun to my head and asked me to guess, and I couldn't tell what answer you wanted to hear, I'd say that my suspicion is that there are, mostly on the strength of analogy to other areas of biology where we know much more. I would then — cautiously, because you have a gun to my head — suggest that you read, say, Dobzhansky on the distinction between "human equality" and "genetic identity", and ask why it is so important to you that IQ be heritable and unchangeable.


  1. Michael Cargal said,

    July 10, 2014 @ 10:43 pm

    The reason it is important that IQ be heritable is that the more heritable IQ is in a culture, the more egalitarian the culture is. If reported IQ scores were 100% heritable, then nobody would be prevented by the environment from having as high an IQ as nature will allow. If they were 100% environmental, the fact that some people are smarter than others means they had greater opportunities, and differences in intelligence would be an indication of inequality, and to that extent an indictment of the culture. Good liberals such as I should work to equalize environments (medicaid, head start) in the hopes that IQ will become progressively more heritable as environmental influences are mitigated and equalized.

  2. Patrick M. Dennis, MD said,

    July 11, 2014 @ 6:58 am

    Ha, Michael C! Nice work standing the conservative argument on its head. The Jensonist would of course counter that the "fact" that differences in IQ are heritable precisely absolves the environment (i.e., society, culture) from either culpability for maximizing them or any responsibility to attempt to minimize them.

  3. H Stephen Straight said,

    July 11, 2014 @ 10:06 am

    Mark, I am surprised that you did not mention that one half of one quarter is (actually just) a bit.

  4. JW Mason said,

    July 11, 2014 @ 12:41 pm

    If reported IQ scores were 100% heritable, then nobody would be prevented by the environment from having as high an IQ as nature will allow.

    Similarly, a disease which is 100% heritable is utterly impervious to treatment of any kind. Wait, what?

  5. KevinM said,

    July 11, 2014 @ 1:03 pm

    90% of the game is half mental. Yogi Berra.

  6. J. W. Brewer said,

    July 11, 2014 @ 1:04 pm

    A practical example a la Michael Cargal is that in the United States and Western Europe adult height has been moving closer to 100% heritable – a few centuries back it was affected more dramatically by childhood environment but (at least according to Robert Fogel) as the countries got more prosperous they reached the level where even kids at the poorer end of the income distribution mostly got the minimum level of nutrition in their formative years needed to not have their growth stunted – once you are over a certain key threshold in childhood diet you are (absent more specific injury or disease along the way) highly likely to hit the adult height you are genetically predisposed to, without anything additional a rich family can provide by way of environmental stimulus likely to add to that. That we have gotten to that near-100%-hereditary point seems like it ought to be a sign of progress.

  7. Rubrick said,

    July 11, 2014 @ 4:28 pm

    Presence of the words "nature" and "nurture" in an article correlates strongly with lack of intelligence.

  8. the other Mark P said,

    July 11, 2014 @ 5:17 pm

    and ask why it is so important to you that IQ be heritable and unchangeable.

    Because we believe what we see with our eyes, not what academics tell us?

    My kids have exactly the same academic strengths and weaknesses as my wife and I. They're not just the same "intelligence", they're the same sort of intelligence. This despite my strong attempts when raising them to stress all sorts of other knowledge and thinking. They just shrugged that off and became what they would become. All around you see peoples' children who show the same pattern — families with amazing musical ability, say, despite the parents not stressing that at home.

    It's far too strong to be merely upbringing. As we see when people who stress academic success in their kids, and whose kids strive hard as a result, run up against the kids' inherent weaknesses. You can't force brains.

    Now I can't prove it, of course, to an academic's level of "proof". But the reason people believe in heritable intelligence is because they see it around them all the time.

    That the current strand in liberal academic theory is to deny this obvious fact is just something to endured until the tide turns and we go back to it being heritable again.

  9. Alfonso said,

    July 11, 2014 @ 6:28 pm

    @other mark, You keep seeing "heritable" traits deriving from nurture because you keep forgetting that the primary method organisms, including humans, use to gain experience and learn behaviors is the "observe and test" method, not the "listen to what's being explained to you and then try to implement that" method. Your kids ended up like you and your wife because, as their parents, you were key models in their observations, and your reactions, whether or not you consciously intended them to, trended them towards acting like you.

  10. Yosemite Semite said,

    July 11, 2014 @ 11:55 pm

    This week (7/11/14) the radio program Radiolab, heard on National Public Radio, aired a segment on morality, featuring a report on the response of experimental human subjects given difficult moral decisions to make, during which the experimenters took MRIs of their brains. In the discussion of the results, one of the experimenters hazards the opinion that, based on the types and brain locations of the MRI results that he sees, morality is deeply embedded in older portion of the brain. He even somewhat jokingly calls it "the inner chimp." The Radiolab team also goes off to talk with Frans de Waal, of the Yerkes Primate Center, about morality among chimps. I cannot do justice to the segment, because I'm merely reporting from listening to the segment once through earlier this evening, without the slightest inkling that it would have any relevance to anything else, and the podcast of the report is not yet up. It may not have any strong bearing on the thrust of this post — apart from the validity of the contents of the segment — but I offer it up as something possibly to consider. http://www.radiolab.org/story/91508-morality/

  11. Yosemite Semite said,

    July 11, 2014 @ 11:57 pm

    BTW, both the MRI researcher and Frans de Waal posit a strong genetic component for morality. Which is why I suggested it might have any bearing on this post in the first place.

  12. Pflaumbaum said,

    July 12, 2014 @ 7:14 am

    @ the other Mark P

    The problem is that beliefs which seem like 'obvious facts' that we 'see around us all the time' frequently turn out to be false. As shown many a time on this site with respect to widespread beliefs about language.

  13. P said,

    July 12, 2014 @ 12:56 pm

    The average test-retest correlation in the study was about 0.56 (see Supplemental Information). If we assume that the two testing sessions were parallel in the psychometric sense, it means that almost half of the variance in the tests consists of random measurement error. Therefore, it makes a lot of sense to estimate the heritability of the principal components which mostly consist of reliable variance, rather the observed test scores, much of which are just random noise.

  14. P said,

    July 12, 2014 @ 1:00 pm

    Okay, I misread the Supplemental Information. There appears to be no information on the reliability of the tests. Even so, I would guess that the reliability of the chimpanzee IQ tests is probably not terribly high.

  15. P said,

    July 12, 2014 @ 1:15 pm

    I guess you could use the total communality of the principal components as an estimate of reliability. It is 54 percent, so my original point stands.

  16. Jonathan said,

    July 13, 2014 @ 11:41 am

    With respect to Cosma Shalizi, one criticism of his argument against the g-factor is that he doesn't explain why different cognitive batteries correlate so well in the first place. He argues simply that a general statistical factor is bound to appear whenever different tests correlate in their results, as if this disposes of the matter, but even if he is right that there is no cause to reify this statistical g factor and posit some "general intelligence" module in the brain (which in fact no one is doing), we still have to explain why the tests correlate.

    While he is certainly very smart, he is not a psychometrician and hasn't contributed original research to the field, as far as I know, and the consensus among experts in IQ research has been that there is such a thing as general intelligence and that it is significantly heritable. See the famous open letter "Mainstream Science on Intelligence" by Linda Gottfredson, published in 1994.


  17. Jonathan said,

    July 13, 2014 @ 11:45 am

    And I think the question of heritability of intelligence is extremely important and socially relevant, precisely because we are so concerned with equal outcomes across different groups. If we didn't care that, say, non-Asian minorities and women were underrepresented in STEM fields and the tech industry, the issue would be purely academic, but we are obsessed with leveling these differences, which means that we need to approach the question honestly and with an open mind, since the answer will determine whether we are indeed faced with socially conditioned inequality that can be remedied by appropriate policies, or whether we are looking at fixed genetic differences that cannot be engineered away, except perhaps by eugenics.

  18. Keith M Ellis said,

    July 13, 2014 @ 1:08 pm

    "With respect to Cosma Shalizi, one criticism of his argument against the g-factor is that he doesn't explain why different cognitive batteries correlate so well in the first place."

    Did you read g, a Statistical Myth?

    First, he points out that, given everyone assumes that g exists, and given that intelligence tests exist in a narrow context of what they're useful for, then of course they correlate with each other because any that don't will inevitably be discarded.

    Second, and much more to the point and damningly, his airtight and meticulous argument shows that even if these cognitive tests constitute unbiased data, g will be produced by these methods whether it exists as a causal single factor, or not. It's a mathematical inevitability.

    "While he is certainly very smart, he is not a psychometrician and hasn't contributed original research to the field, as far as I know, and the consensus among experts in IQ research has been that there is such a thing as general intelligence and that it is significantly heritable."

    But he is a well-regarded statistician and he is criticizing the mistaken conclusions drawn from an application of statistical methods.

  19. Richard Mongler said,

    July 13, 2014 @ 1:27 pm

    @the other Mark P – I don't think that particular quote was to deny that possibility (even if denying some form of "proof" leaves you open to a lifetime's worth of confounding variables) – he certainly doesn't; not when he spends a great deal of space trotting off conditions that we'd say suffice if the initial question was purely about intelligence sans its specific associations. We've far more to learn from specifically answering the question with our *ends*, and I'd argue it's the theme relating that (in particular) to his digression around the area of a fundamental human worth and its role in relation to innate capability (is LL the place for a discussion over axiomatic *attitude*, though?) –

    – but your interest in *specific* intelligences is just as interesting a can of worms – I can think of a counter-correlating factor: the fact that not all "stressing" is the same, and what should seem like low-pressure "encouragement" ends up crashing against assumed expectation, assumed destiny, the assumption that rejecting "encouragement" can be interpreted as inability. To undo all of that we'd have to unpack and specify, and that gets us much closer to academic rigor…

  20. Jonathan said,

    July 13, 2014 @ 2:06 pm

    @Keith Ellis:

    I did read the article. It is certainly persuasive if you know statistics but happen to know little to nothing about psychometrics and how its research is carried out, in which case you would discover that his claim that IQ researchers exclude tests whose results do not correlate with the other is simply false. The correlations are empirically real, not just artifacts of bias.

    The second point in your summary I already dealt with: the point is not whether g can be explained by a single neurological cause or whether it is controlled by several neurological factors; rather, the point is that g exists and explains many behavioral outcomes in people.

    To the extent that Shalizi's criticism is scientifically valid, it is that we still don't understand how g is instantiated in the brain, since it differs so markedly from the modular model of the mind that cognitive science has had such success in elucidating otherwise. Similarly, while g is heritable, we don't know what the genes are that control it, just as we don't understand what genes control language, even though we are pretty sure the ability to acquire language is heritable. So he is right to point to gaps in our ability to explain the existence of g, but he is not right to deny that g really exists.

    I've based my response on the more detailed rebuttal of Shalizi's article here.

  21. Jonathan said,

    July 13, 2014 @ 2:07 pm

    Sorry, here is the article:


  22. Kathrin P said,

    July 14, 2014 @ 2:03 am

    So if I got this right, then the paper behind the Independent article studied how far the chimpanzees' performance on cognitive tests could be explained by their genetic heritage.

    Then there is a leap where someone (was it just the journalists?) said this meant there was evidence for nature over nurture in humans.

    But — isn't nurture, specifically in its effects on intelligence (real or otherwise), very different between chimpanzees and humans? I doubt that the quality of kindergarden care, or the number of books in the parents' home, is a big factor in chimpanzee child-rearing.

    What I think the study shows is that there is a genetic component to the ability to perform cognitive tasks. It cannot quantify the effect of that genetic component compared to education and upbringing because those components are not adequately represented in the test setup.

  23. Matthew McIrvin said,

    July 14, 2014 @ 9:54 am

    @Michael Cargal: Actually, I think your argument would work if you'd said that the more genetically heritable IQ is, the more egalitarian the culture is.

    As it is, you're still making the common confusion of heritability with genetic basis. In a society with low to nonexistent social mobility, a hereditary aristocracy and a permanent peasant class, IQ would probably be extremely heritable–that is, your IQ would be determined to a great degree by your parents' IQ–but it would not be an egalitarian society, and without more information we wouldn't be able to say much of anything about the effect of genes on IQ.

    Scientific inegalitarians sometimes insist that the society they live in is already so egalitarian that not just individual, but even ethnic and racial differences in IQ are now primarily down to ineradicable genetic differences. It's a form of the just-world hypothesis.

  24. J. W. Brewer said,

    July 14, 2014 @ 12:38 pm

    Matthew McIrvin: I think hierarchical societies with limited social mobility often have folk tales (or plays by Plautus, or short stories by P.G. Wodehouse . . .) in which the stupid/foolish child of wealth and privilege is bailed out of a jam by the extremely smart/clever slave/servant/peasant. One might think there were generally some empirical observations underlying the popularity of that sort of plotline. Elites in societies like that do not tend to justify their privileged position in the great chain of being by appeals to g/IQ/standardized-test-scores/etc.

    The current elite in the U.S. (and possibly other Western nations) likes to think of itself as a "meritocracy" for a meaning of "merit" that correlates fairly strongly to measured IQ (because of the standardized-test sorting mechanisms that are used for elite-college admissions and the like). That makes the question of just where this "merit" comes from exceedingly fraught, both politically and psychologically. Indeed, in a less "meritocratic" society, we'd probably agonize over it less. To go back to my prior example of height, because there are only a few corners of our present social arrangements where being at the very high end of the height distribution is unusually valuable/prestigious, we don't spend a lot of time worrying about why e.g. so few Asian-Americans seem to be as tall as Jeremy Lin (who at 6' 3" is not particularly tall by the standards of his specialized profession) and what nature/nurture factors have combined to contribute to underrepresentation of Asian-Americans among the extremely tall.

  25. P said,

    July 14, 2014 @ 3:14 pm

    @Matthew McIrvin

    'Heritable' is a technical term that refers exclusively to genetic effects, not social/environmental effects even if the latter are passed from one generation to next. Behavioral genetic methods make it possible to differentiate genetic from non-genetic inheritance. Non-genetic inheritance is called the shared environment, and can be estimated in twin and adoption studies. If a twin study of IQ heritability was conducted in your hypothetical non-mobile society, it would show that the heritability of IQ is low and the shared environmental effects high.

RSS feed for comments on this post