Skip to content

Bad news for “intelligence-genes”

Intelligence and its heritability has been a popular topic in scientific communities and public discussions for long. Recent findings give new insight to the debate: one of the largest studies on genetic influence to intelligence and other behavioral traits turned up inconclusive findings, as Nature News reports in a recent article “Smart genes” prove elusive.

Existing literature on candidate gene associations is rich in studies that have been unable to replicate and findings have been based on “wishful thinking and shoddy statistics”. According to an editorial in Behavior Genetics,

it now seems likely that many of the published findings of the last decade are wrong or misleading and have not contributed to real advances in knowledge.

The journal declares it will tighten its publication policy for candidate gene association studies of complex traits: they now recommend direct replication analysis prior publication and rigorous testing of statistical models. By this, they wish to decrease publication of findings brought by mere chance or other kinds of biases.

Especially twin and family studies, which repeatedly have reported a genetic basis for intelligence and behavior, are subjected to critique. The challenge is these studies is the assumption that genetic and environmental effects could be separated, but the case is not so simple.

In contrast to the criticized studies on “candidate genes”, new studies operate with genome-wide association (GWAS) scans of single nucleotide polymorphisms (SNPs) and large samples of persons – the quoted studies included more than 100,000 participants. The GWAS studies to date have not found genome-wide significant SNPs in social-science genetics that replicate consistently, and researchers predict that most reported genetic associations with general intelligence are probably false positives.

A study on genetic variants associated with educational attainment (how many years the individuals were in education) found three variants, but the estimated effect was not remarkable: each variant estimated approximately 1 additional month of schooling. The researchers report that existing claims of candidate genes and their association with complex social science traits were even 100 times larger than in their findings, and they propose their results to be a benchmark for evaluating the existing literature.

 In social-science genetics, researchers must be especially vigilant to avoid misinterpretations. One of the many concerns is that a genetic association will be mischaracterized as “the gene for X,” encouraging misperceptions that genetically influenced phenotypes are immune to environmental intervention and misperceptions that individual SNPs have large effects (which our evidence contradicts).

In the follow-up study, the education-associated SNPs were related with SNPs associated with cognitive performance. Three variants associated with both were found, but again the effect was low: even if a person had both copies of all the variants, she would score an average of 1.8 higher points in an IQ test (when most of the population scores 85-115; hence, 1.8 points is not a big number). The study is cautious with these kinds of results: with such small effect sizes, the risk of false positives is increased.

The origins of intelligence, thus, remain twisty. Bearing in mind that there is not even a unified definition for the very concept and its operationalization, the situation is somewhat expected.  Are IQ-tests the best measure? What about different kinds of intelligence? Do things such as logical and social intelligence and creativity come together or can they be exclusionary? Can there be a bulk term “intelligence” to describe the “best thing to be”? Many answers of course exist, but simplified would be to say that the question was easy or solved.

But at least this seems to be an easy conclusion: the “intelligence-gene” has little scientific evidence. Many genetic variants are related to cognitive ability, but with small effects. Genetic studies on heritability of complex traits tell us about the variation of a certain trait in a population – not necessarily the reasons why some people have the trait. Heritability does not mean controllable.  What we might call intelligence is a polygenic and multifactorial question: genes of course affect it, but the outcome is the result of all sorts of environmental questions including ecological, physical, cultural, and socioeconomic issues. Genes and environment are a matter of circular causation.  

An interesting question of its own is that what lies behind the “wishful thinking” in studies in genetics and intelligence. Why it is a thing to be hoped for that intelligence would be determined by genetics? What does it mean to hope that intelligence would be genetically determined?

At least it means that intelligence would be inborn, and the division between the intelligent and non-intelligent would be a solid one. This idea indeed has historical roots: “The use of flawed science to help present a case for accepting the status quo is not new. Attempts to link intelligence to social hierarchies were made throughout the nineteenth and twentieth centuries.” For example the Bell Curve went on making straightforward connections about the genetic origin of intelligence between individuals and groups, including race. As the scientific validity of that and similar studies have been shown to be highly questionable, the motivation for creating such findings is in need of explanation. Why is the idea preferable that some people, or groups, would be more intelligent than others? That others would, biologically, be less intelligent than others? Moralism, discrimination, and the aim of cementing power structures and social hierarchies should not be hidden into flawed science and purpose-oriented interpretations.

Straightforward conclusions about motives behind optimism of finding specific genetic origins of intelligence are of course not preferable. However, digging up and discussing assumptions and ideologies is worthwhile.

The bioethical literature suggesting genetic enhancement of intelligence of course could have another answer: that the genetic origin would bring the possibility for making us all very intelligent. However, considering the implausibility of the “gene for intelligence”, this is likely to remain as a thought-experiment. If there is a big concern about bringing possibilities for becoming more intelligent, then, taking account the extensive data on increasing “intelligence” with environmental means, there is already now the possibility for fierce campaigning, for example, for access to high-quality education throughout the population.

Share on

6 Comment on this post

  1. There is of course an equal kind of wishful thinking or even ideology behind hoping to show that there is no strong genetic basis for intelligence: then maybe those individual differences are all due to environment and/or individual choice, far more amenable for intervention than genes. The history of intelligence is replete with examples of people on both sides showing plenty of motivated cognition.

    If intelligence can be increased using environmental means, it is no cause for ethical complacency. There is often a hidden assumption in people’s intuitions that “soft” interventions are cheap and fair. It might turn out that education is more expensive than genetics or smart drugs would ever be (it requires ongoing support by highly skilled professionals rather than something mass-produced), or that it has a multiplicative effect that preferentially benefits the cognitively and motivationally better off.

  2. I have a feeling (and only a feeling) that the elusive ‘intelligence gene’ only (partially) determines the hardware necessary for intelligent software to be developed, to use the perhaps overused hardware-software analogy. The hardware for intelligence (a certain pattern of neural connections, say) can facilitate the ease with which the software of intelligence is actualized – a process predominantly due to environment and socialization.

    And here’s the rub: the potential for intelligence (i.e. hardware) may be genetically determined, but it is only the actualization of intelligence that can be observed and tested (through IQ tests or more holistic approaches that capture the scope of however ‘intelligence’ is defined).

    The challenge for genetic researchers seems to be how to identify or operationalize the potential / hardware for intelligence: how is the potential for (higher) intelligence observed in an individual who has not fully (or to an above-average level) actualized that potential? And what good would pinpointing this potential-for-intelligence gene be when the capability for intelligence, i.e. the societal-level valuation for every individual’s freedom to actually become intelligent, is no longer regarded as such?

    1. The hardware/software issue is discussed in the field, but usually discussed as potential vs. realized intelligence. One of the main findings is that this indeed links to socioeconomic status. Heredities of IQ are higher for middle-class people than for lower-class people. The reason is likely that the chance to fully realize one’s potential requires stimulating experiences, parental interactions, healthy food, less stress and disease, and education – the lower SES people have IQs bounded by these more or less random factors rather than the upper limits set by their genetics, while the middle-class people are more or less bumping the ceiling: they have already had as stimulating and healthy childhoods as is possible.

      If this ceiling effect is the true explanation then we might be lucky in that we actually know roughly how much resources needs to go into childhood to help the worst off. But even if we could marshall those resources (and they are not just money) there would now be a fair bit of genetic influence – remember that the OP was about the difficulty of finding individual genes, but nobody disputes that IQ is fairly strongly heritable (about to the same degree as height, which also shows the SES/ceiling effect).

  3. Thank you for this.

    There is nothing particularly new in these findings it is just that they are not particularly fashionable. Nature and nurture theories come in and out of fashion; the former, having recently had a good run because of the human genome project, is now beginning to run out of steam as the claims made by some researchers and the growing band of futurologist are found to be wishful thinking and/or a grab for funding. (Your “wishful thinking” question needs a fuller answer than this but that would take us into the history of genetics)

    We know that good education and social conditions are a benefit to individuals regardless of their intellectual abilities and are therefore good for society as a whole. There is no evidence that enhancing “intelligence” by genetic engineering or, as Anders suggests, smart drugs will have any long term benefits for individuals or society. There are numerous reasons for this, not least of which is that we do not know what intelligence is. The often unquestioned use of IQ as a measure of intelligence is a worrying return of an old ill-fitting fashion that should only be used as a rough guide to identify some conditions. The notion that we should devote much research to increasing IQ by genetic engineering and/or use smart drugs is bad science and irresponsible. Irreversible genetic engineering (“enhancement” is usually irreversible) should only be considered as an extreme and last resort to very well defined severely disabling or potentially fatal genetic conditions. All drugs have side effects and making people dependant upon these drugs to live their lives on equal terms to other users would be a form of mass addiction (very good for Big Pharma but little benefit to the addicts). Social engineering on such a massive scale can never be justified and we can but hope that we are entering a period where some of the bad science can be rejected and we take a more balanced view. I agree with Michel that there are interesting lines of research open, but please let us get back to the science and leave the wishful thinking to sci-fi authors.

    1. I disagree with the argument that enhancement will not benefit society because we do not know what intelligence is.

      First, there are many enhancements that deal with well-understood things like memory or attention that clearly could have beneficial effects.

      Second, just because something is vaguely characterized it is neither irrelevant or impossible to affect: just consider “wealth” and “happiness”, two things we do not have strict definitions of, yet know some ways to improve and generally would like to improve across society.

      And third, we do have fairly compelling data suggesting that societies or individuals that score highly on IQ do have better outcomes. Childhood IQ does predict health outcomes, longevity, income, risk of being a victim of crime and so on; adult IQ protects against a bundle of bad things like divorce, of course improves education and income, and correlates with holding long-term plans and altruism. Economists have found clear GDP effects, probably due to network effects of more smart people getting better education and vice versa (the causality issue is fascinating; untangling what causes what in virtuous circels is a fair challenge).

      I feel many of the comments on the IQ/genetics topics on this blog over the past few months are attacking strawmen views of both the state of psychometric and behavioral genetic research, as well as the reasoning in enhancement ethics. Maybe I am just grumpy. But thinking that intelligence research will/ought to disappear just because there are much conceptual and ideological stupidity around it, and that enhancement research/application is science fiction, that is in itself a rather blinkered view of what is going on.

  4. 1) How much memory, how much attention? I am questioning whether it is possible to increase memory and attention by genetic engineering and/or drugs in a way that benefits individuals and society. For sure there may be uses for drugs that increase memory, but it cannot be assumed that they would be beneficial if taken by most of the population. If in the far of distance future a drug is produced that has very little or no side effects you might, if you are still living, be onto something more than moonshine.

    2) The analogy you make with wealth and happiness is not relevant to this problem because you are not directly trying to increase them by genetic engineering and/or drugs. If you did plan to do this I would object on similar grounds to that which I have in the case of intelligence. I would like to see knowledge, understand and intelligence increase, as I would general wealth and happiness, but not by your means because I am not convinced it will work.

    3) You identify the problem with this arguments. People who have been through a “good” education and come from middle class or above background are more likely to score well on an IQ test. True. But that does not make it a reliable objective measure of intelligence. It is a measure of the cognitive skills of that class of people. When I was being taught how to program computers in the 60s I was also taught how to significantly increase my IQ because a high IQ was expected in that class of people (the 11+ also increased IQ). Of course I was not anymore intelligent just better at doing IQ tests. As I say, it should only be used as a rough guide to some conditions.

    I am not saying that research in these areas should be stopped. What I am objecting to is the extraordinary hype and absurd claims that are made by those that preach enhancement. You cannot say that those of us that do not share your faith are attacking strawmen. Johanna’s post cites further evidence of hype and bad science. I am not convinced by the so-called science that underpins the technological fix to higher intelligence and the good life. I am not making ridiculous claims about how it will be the duty of parents to genetically enhances their offsprings’ intelligence or that it will be our duty to become ethically enhanced (whatever that means). I am not saying that if you disagree with me you might have to be forced to agree me (I am threatened with compulsion if I do not do my duty). Nor indeed am I wasting my time writing about how to keep a superintelligence in a box. (Sorry, I am sure Nick Bostrom means well, but surely it is a little premature to start worrying how an AI is going to use its ‘supersocial skills’ to charm its way out of a box in order to immediately take over the world.) I could go on and fill pages but I am leaving that for another place.

    I repeat, I do not want to stop research although I would like to see a little more realism and a lot less hype. I would like to see those engaged in enhancement ethics more engaged in serious philosophy. We are all too often presented with statements like ’We can’t think about this or that because we are not enhanced or superintelligences – we must wait and see what the transhumans and superintelligences think (not that we will understand very much about what they think)’. We have not been convinced by this argument for the last few hundred years when it was used to make the mind of God unknowable and it does no better within this context.

    I know I am grumpy and I know what has made me that way.

Comments are closed.