Nature, nurture, environment & culture

I.

There are a few books I’ve read recently that make the case that when it comes to nature versus nurture, nature is pretty much winning. Take this paragraph from Geoffrey Miller’s The Mating Mind (a book I very much recommend):

Our bodies are rich sources of evidence about sexual selection pressures because they are visible, measurable, easily comparable with those of other species, and relatively undistorted by human culture. In recent years much nonsense has been written by post-modern theorists such as Michel Foucault about the "social construction of the body," as if human bodies were the incarnation of cultural norms rather than ancestral sexual preferences.

Or look at basically the whole thesis of Bryan Caplan’s Selfish Reasons to Have More Kids, which is that people who have children don’t need to bother thinking about their parenting methods, because the impact of parenting is basically nil. So please, go forth and multiply, because even if your kids survive on TV dinners each night, it won’t do them any harm in the long run! 

And there are a load of other books making the case that genes really matter, and that it’s misguided to think that environment is as important as genetics when it comes to most traits that we care about.

I agree with a lot of this stuff. Most people underestimate the impact of genes on success, other life outcomes, and human preferences. When we discuss politics, my dad often argues that his successful career as an accountant is due to his self-discipline, and that increasing his taxes would be taking from someone who worked hard and giving to someone who didn’t. When we aren’t talking about politics, he will readily admit (or even boast) he’s naturally extremely good with numbers, and also that he’s able to work much less hard than colleagues because of his intelligence. 

On the left, people often have the mirror view. It simply cannot be the case that some people naturally have a lower IQ than others and that this will have a huge effect on life outcomes, because that would mean that all humans aren’t created equal and income inequality isn’t only the result of greedy capitalists and the like. More generously, the ‘all hail IQ as the supreme indicator of human worth’ position has a number of less-than-credible advocates, and is a pretty nasty view even if the ‘IQ is a strong predictor of life outcomes’ position is correct.

But at the same time, I’ve begun to think that the sort of people I hang around with have begun to overrate nature and underrate the impact of environment. I suspect that’s often because they’re reading the same books that I am, and that a natural reaction to so many people going on about how genes don’t matter is to overcorrect and think they’re basically all that matters. 

But this view is obviously wrong, even if you take seriously the (correct) argument that genes are really important. It’s not that these books make claims that are wrong, you can skim through them and figure out that clearly genes aren’t everything. But I think a load of people have read books like these and come away with the impression that environment and culture barely matter.

II.

Let’s start with Geoffrey Miller on his view that the claim that ‘human bodies [are] the incarnation of cultural norms rather than ancestral sexual preferences’ is clearly wrong. Obviously he’s kinda right, and the claims that human bodies are totally socially constructed are off the mark. But at the same time, cultural norms are pretty important determinants of what is and isn’t sexually attractive, and also determinants of the actual, literal bodies that people have.

Take, for example, the data on the cosmetic surgeries that women are getting. This long read in The Guardian documents the rise of the ‘Brazilian Butt Lift’, a dangerous surgery that costs thousands (or tens of thousands) of dollars and works by taking fat from different parts of the body and injecting it into a woman’s buttocks. 

This is apparently quickly becoming one of the most popular cosmetic surgeries, accompanied by a decline in the number of women getting breast augmentations, presumably as a result of changing fashions. (It might be worth mentioning that when I was younger, I remember that most people considered telling a woman she had a large behind to be an insult, whereas now I suspect it’s more likely to be a compliment). There are other obvious examples: Heroin chic, the highs and lows of facial hair, and so on. 

But okay, the fact that fashions exist, and that people change their bodies (whether through surgery, exercise, or other methods) in response to new fashion trends is pretty obvious. What about the other claims?

The first thing to state is that Caplan’s book actually freely admits that environment is really important in determining outcomes. He just distinguishes ‘environment’ from ‘nurture’ - his claim here is that even though a child’s environment has a big impact on a child’s development, the parenting style a mother and father choose has either no impact or very little impact.

A brief note here: one way to determine the heritability (the degree of variance in a trait that can be explained by genetics within a population) of some characteristic is to use twin studies. Identical twins share 100% of their genes, their upbringing is shared, and they also have some environment that is nonshared (for instance, they might attend different universities). Non-identical twins share 50% of their genes, have a shared environment, and also have a nonshared environment. 

Using some fairly basic statistics, if we have access to a dataset of identical and non-identical twins, we can determine what portion of some characteristic is heritable, what portion is derived from the shared environment, and what portion is derived from the non-shared environment. 

So as an example, when Caplan makes a claim about parenting having no impact on a child’s life expectancy, he is not making the claim that genes are basically all that matters when it comes to life expectancy. The paper he cites here is a Danish twin study that looked at Danish twin pairs born between 1870 and 1900, and found that there was no impact of shared environment on life expectancy. But the proportion of the variance that could be explained by genes was only 26%, meaning that (as Caplan notes) ‘if you live longer than 80 percent of the population, you could expect your separated identical twin to live longer than 58 percent of us’. Environment beats nature here.

(There’s also some evidence that the 26% figure is likely to be overstated, with Ruby et al. (2018) claiming that the ‘true heritability of human longevity for birth cohorts across the 1800s and early 1900s was well below 10%’.)

Another study of interest here is this 2004 adoptee study, also cited by Caplan. Korean-American babies were randomly allocated to adoptive families at birth, and the study determines the extent to which various characteristics are transmitted from adoptive families to adoptees. 

The finding that Caplan quotes is that having a college-educated mother only increases the chance of her adopted child going to university by 7 percentage points, whereas there is a 26 percentage point increase for non-adopted children. But the study also finds that the transmission of drinking and smoking behaviour from parent to child is the same for adoptees as for non-adoptees. So, for smoking and drinking this study seems to indicate that parents do matter, and I also don’t think that a 7 percentage point increase is necessarily a number to sniff at.

The paper also highlights environmental factors that do seem to make a difference to a child’s educational outcomes, even if the total effect isn’t that large. One finding is that ‘Adoptees who are randomly assigned into a single child family have about a 55 percent probability of college graduation versus a 35 percent graduation rate for adoptees assigned into a seven child family’. This correlation remains quite strong after controls, and even if it is the case that there are omitted variables that are doing the work here, it’s still an impact of environment on education.

So, if you take the Caplan view that parenting styles aren’t that important, that’s pretty different to thinking that genes are basically all that matter. And another thing worth mentioning is that there is one parental intervention which clearly does make a difference - and that’s immigrating from one country to another. Given that we know the evidence around migrating on wages and various other outcomes (partially from Caplan’s book Open Borders), we can raise an objection to many of the studies that claim to show heritability, which is that they don’t properly account for culture and shared national environment.

Heritability is a measure of the percentage of the variation of some characteristic that is caused by genes within a population. This means that saying a characteristic is heritable means something different to saying that it is passed down through the genes.

Here’s an example, taken from Sauce and Matzel (2018): the number of fingers on your hand is almost entirely determined by genetics. But if you look at the heritability estimates for ‘number of fingers’, they’re extremely low. Why? Well, the vast majority of variation in the number of human fingers is caused by environmental factors like accidents and prenatal complications. So even though the number of fingers a person has is basically caused by your genes, variation in the number of fingers is mostly caused by environmental factors. So, the heritability estimates come out low. 

Similarly, perhaps there isn’t much variation in a trait across a population, and so the heritability estimate will appear high, but this doesn’t imply that if the environment was totally different, there wouldn’t be huge changes in the trait. The Flynn effect, the substantial increase in measured IQ across the 20th Century, is an example demonstrating that even though the heritability of IQ is very high, environmental changes can make a very substantial difference to IQ scores.

So, there’s my brief correction to the ‘genes are basically all that matter’ crowd. I want to note that I don’t think that the books I cite make the claim that genes are all that matter, but some people seem to have mistakenly come away with the view that environment and culture have very little impact on outcomes, and I think this is pretty obviously wrong.