Nature, nurture and noise

Share on FacebookShare on Google+Email this to someoneTweet about this on Twitter

Mirrored from

The question of what makes each of us the persons we are has occupied philosophers, writers and daydreamers for millennia but has been open to scientific inquiry over a far shorter time. The answer clearly lies in the brain and, somehow, in how it is “wired” (whether that refers to the amount or type of connections between different brain areas or to differences in how circuits function). But differences in brain wiring could be either innate or due to experience or environmental effects. This has been famously framed, by Galton originally, as a clash of nature versus nurture. The inspiration for this phrase may have come from Shakepseare’s The Tempest, in which Prospero refers to Caliban as a “A devil, a born devil, on whose nature / Nurture can never stick”. That line encapsulates the notion of an innate character that is resistant to extrinsic influences, especially efforts to change aspects of a person’s personality – an idea which anyone who has children may find easy to relate to.

While neurology and neuroscience have offered direct evidence that differences in the brain affect behaviour, it is behavioural genetics that is typically seen as having contributed most directly to the nature-nurture debate. Twin and adoption studies, the mainstays of behavioural genetics, have demonstrated very conclusively that many aspects of personality, behaviour or other psychological traits are highly heritable – that is, a large proportion of the variance in the trait across the population is attributable to differences in genes.

The logic of the twin studies is usually the inverse of the statement above – to look at people who share various proportions of their genes and see how similar they are to each other. These generally show that monozygotic (“identical”) twins are far more similar to each other for most psychological traits than dizogytic (“fraternal”) twins. Also, adoptive children tend to resemble their biological relatives for psychological traits and are hardly more similar to their adoptive family members than they would be to any stranger in the street.

These data have demonstrated unequivocally that variation in genes can affect behaviour in humans. They have also, however, dramatically illustrated the limits of such genetic effects. Monozygotic twins, while much more similar to each other than would be expected for people who don’t share all their genes in common, are nonetheless clearly not identical for most psychological traits. In fact, on average, genetic variance only explains about 50% of the phenotypic variance. What is causing the rest of the variance?

The presumption in the literature has been that it is something in the experience of the individuals – that in fact, the results of these behavioural genetic studies provide the strongest evidence to date for some effect of the environment. I say strongest evidence because in fact there is very little other evidence for direct experience-dependent effects on human psychological traits across the general population. (This is not in any way to deny the impact of factors such as serious abuse in individuals, which can be profound – it is just that, happily, such abuse is rare enough that it does not contribute significantly to the variance across the population).

Certainly, the twin and adoption studies have consistently found only a very modest effect of a shared family environment on the types of traits examined. The argument for interpreting the excess variance as being caused by experience is thus that it is the individual, “non-shared”, experiences that people have that make them different from each other. This has never made any sense to me, I have to confess. Non-shared experiences can make me differ from my twin but shared ones cannot make us more similar? An experience can affect my psychological development, but only if it does not also happen to my twin? Interactions with peers and teachers can have lasting effects on me but interactions with my parents cannot? Not only does this notion not make intuitive sense, there is no evidence for it. It seems far more likely that most non-traumatic experiences, regardless of who they are with, have little long-lasting effect on the kinds of traits that define us as persons. What then can account for the unexplained variance?

The basic problem with the interpretation above is that it limits “innate” influences to “genetic” ones. Just because some trait is not genetic does not mean it is not innate. If we are talking about how the brain gets wired, any number of prenatal environmental factors are known to have large effects. More interestingly, however, and probably a greater source of variance across the population, is intrinsic developmental variation. Wiring the brain is a highly complex procedure, reliant on cellular processes that are, in engineering terms, inherently “noisy”. Running the programme from the same starting point (a specific genotype) does not generate exactly the same output (the phenotype) every time. The effects of this noise are readily apparent at the anatomical level, when examining the impact of specific mutations, for example. In many cases, the phenotypic consequences are quite variable between genetically identical organisms, or even on two sides of the same brain. (If you want to see direct evidence of such developmental variation, take a directly face-on photograph of yourself, cut it in half and make mirror-image copies of the left and right sides. You will be amazed how different the two resultant faces are).

If the way the brain is wired relies not just on the starting genotype, but, to a large extent on chance events during development, then it is reasonable to expect this variation to be manifest in many psychological traits. Such traits may thus be far more innate than behavioural genetics studies alone would suggest. [Note that this does not imply all aspects of a person’s behaviour are innate and resistant to effects of experience – that is obviously a nonsense. It is important to recognize that the kinds of personality traits that have been examined in the studies mentioned above (such as extraversion or neuroticism) are basal characteristics, reflecting, for example, how strongly the brain responds to positive or to negative stimuli. These tendencies, in combination with later experience, influence, but in no way determine, moment-to-moment behaviour].

Happily, in my view, this places intrinsic limits on genetic determinism and the ability to predict many important aspects about a person from their genotype. Not limits based simply on our current knowledge that could one day be overcome – limits due to the inherently variable nature of neural development. Ultimately, what defines us each as persons, is thus dependent on nature, nurture and noise.

For more on this subject, see: Mitchell, K.J. (2007) The genetics of brain wiring; from molecule to mind. PLoS Biology Apr 17;5(4):e113. (Open Access).


  1. I’ve nearly asked about non-genetic biological effects on traits several times here, but never did. I’m glad to see it addressed.

    It always struck me as a bit premature to attribute differences to environment when there are so many biological steps involved in the actual expression of a gene. Random small changes early on can have massive consequences for the entire system, especially in development.

    Genes are the blueprint for a pattern, that does not mean the pattern will always look like its blueprint.

  2. Great post. This is an especially poignant question: “Non-shared experiences can make me differ from my twin but shared ones cannot make us more similar?”

    I would point out that shared environmental effects may be significantly underestimated by twin (and adoption) studies. Take the example of IQ, a psychological trait with a high heritability:
    1. Twins reared together are significantly more similar than twins rared apart, even at middle-age. This comparison is more valid than the usual MZ-DZ comparison because it doesn’t make so many assumptions. (see my post here

    2. The existence of a generational Flynn Effect proves that there are shared environmental influences that change significantly with time. At the very least, you must admit that the shared environment matters between generations.

  3. Clearly there can be environmental effects on behavioural or psychological traits – sometimes quite large (such as from extreme neglect). All the twin studies show is that within a given population at a given time, shared environment effects (usually taken to mean shared family effects) do not contribute much to the overall variance of the trait. (They may contribute something – my reading of the effects on IQ is that any shared environment effects decrease sharply with age).

    Because of the way they are designed, these studies say nothing about differences between populations or between generations, which may be (and commonly are) entirely due to environmental differences. So, for example, a person’s IQ may be affected significantly by both their genes and their nutrition and general health status – however, the differences between people in a population where general nutrition levels are fairly equal will be more due to differences in their genes.

    At the same time, genetic and environmental factors are insufficient to explain all the phenotypic variance – there must be something else adding to the variance and I think there is likely a major role for intrinsic stochastic developmental variation in contributing to phenotypic variance.

  4. I think there is likely a major role for intrinsic stochastic developmental variation in contributing to phenotypic variance.

    This IMO is a very important post and point being made, and of course it bodes poorly for would-be genetic engineers, who would like to re-engineer more intelligent phenotypes out of the available genotypes – in that it may never be possible with any certainty to do this.

  5. Non-shared experiences can make me differ from my twin but shared ones cannot make us more similar?

    I don’t see the problem here. If you have a monozygotic twin then your genes would be identical. If you also had all the same experiences (including pre-natal) then you would be completely identical people. No problem so far?

    Now suppose your genes are still identical, but some of your experiences are not identical. The result is that you are no longer identical people. And it is your non-shared experiences that have made you different! Your remaining shared experiences are still there in the mix, but the answer to the question “why are we different?” is the non-shared experiences. To me this looks like a perfectly good answer to the question.

  6. jb,

    The studies suggest that shared environmental influences do not have any effect on making twins more similar to each other. The key part of the question is why this is the case.. why should shared environmental influences exert *no* influence according to these studies?

  7. Chris said, “It always struck me as a bit premature to attribute differences to environment when there are so many biological steps involved in the actual expression of a gene. Random small changes early on can have massive consequences for the entire system, especially in development.”

    Thank you Chris!

    I believe that I’ve read somewhere that monozygotic twins have many genetic differences, e.g., SNPs and Copy Number Variations?

    If one were to make a chocolate cake and put all the ingredients into a bowl such as the flour, sugar, baking powder, cocoa powder, egg, etc., mix it together thoroughly and evenly pour the batter into two identical baking pans and bake them together in a single oven, one would not end up with two identical cakes.

  8. Thanks for your comment jb. The point from the twin and especially adoption studies is that monozygotic twins who are reared apart are almost as similar to each other as those reared together – a shared family environment is thus NOT making the ones reared together more similar to each other (and the implication is that shared experiences have little effect on the phenotype). It is thus not clear to me how one can argue that the non-shared experiences are making them different from each other – this seems simply like the flip side of the same equation. It is effectively arguing that an experience can have an effect on me only if it is not shared by my twin.

  9. Fraternal twins are more similar in IQ to each other than ordinary siblings. This suggests that twins share an environment, either pre-natal or post-natal (or both), that tends to make them more similar to each other. I’m not sure if this effect persists to adulthood though. Some of it might be due to the Flynn effect.

Leave a Reply