« Open Access Biology | Gene Expression Front Page | Genetics of Hair Color (again) »
December 23, 2003

Nature "vs." Nurture (again!?!?!)

Earlier today "Guessedworker" commented on my previous post that offered a reductionistic hypothesis for intra-racial attraction. He seemed to think that I was offering a "Blank Slate" explanation and was being "Oedipal." In an IM chat with a friend, the same issues of relationshp to Freud came up. This really surprised me and was frustrating, since both are regular readers of GNXP, and I began to wonder how unclear I was being.

The thing is, I never think of Freudianism when I propose anything, because I conceive of it as just another religion! I don't even notice that I might sound Freudian because I think that paradigm is pretty unscientific, so confusions of terminology don't register in my mind. So I apologize for the navel-gazing and inability to look beyond my own horizons. My hypothesis was based on this study. I admitted that it was provisional and I was more illustrating a way to think about a problem than offering an air-tight solution to it. Physical attractiveness is such a multi-faceted topic that I approach it with trepidition, though I do broach it in a crass fashion now and then (this is mostly to entertain and lighten the atmosphere).

But back to the general issue, was I being a "Blank Slater"? I pointed to language acquisition as a rebuttal to that assertion. There is a general consensus that language is something that has a deep biological root. Even the neural connectionists agree on that (though they reduce the biological element to very atomic units that "emerge" over time into the complexity that we see [hear]). At this point, a Pinkerian-Chomskyian model seems ascendent, proposing a "Universal Grammar" and "Language Instinct." So is what language you speak determined by your genes? No!. It is determined by your peers! It is a case where the genes set parameters and constraints, and various factors in your environment come together to "crystallize" your linguistic skills and proficiencies.

Moving even further into generalities, this sort of typology, "Nature vs. Nurture," needs to really be discarded. If there is something you take away from the blog, that should be essential. I didn't give the most glowing review to Matt Ridley's new book Nature via Nurture, because I thought that the central thesis was exceedingly banal. But was it? Our society still uses the term "Nature vs. Nurture" as if the two variables are always at cross-purposes. Ideas like the extended phenotype, which have been around for 20 years have obviously not made it into the public consciousness (the idea that genes can influence your choices of environment, enriching your genetic predispositions and heightening relative genetic fitness, etc.). For a specific example, look at the study of the MAOA gene. While those with high levels of MAOA tended not to become abusers even if abused, what one must note is that even those with low levels of MAOA did not become abusive if not subjected to abuse. This is not a case of environment and genes being at odds, rather, the problem (in this case) is the intersection between gene and environment.

On this blog we tend to focus on genetics. There are many reasons that I do this, speaking for myself.

1) People tend to be less comfortable with biology than they are with psychology and sociology or anthropology, so the default explanations tend to be the latter, and this can be distorted to the most bizarre lengths (ie; the idea that gender is totally socially constructed).

2) I have a strong interest in science, I want to share that and meet like minded people.

3) New technologies and methods are making the genotype a very rich field for study and pointing toward therapy of many diseases that have been relativelly immune to talk therapy or other "environmental" redresses.

4) There are technical issues I'd like to bring to the fore. As an example, I want to make more well known the technical definition of heritability to the lay public. In sum, when we say that a trait is 50% heritable, that does not mean that everyone has a 50% genetic component in phenotype X. Rather, 50% of the variation of a trait within the population is due to genetic variation.

Let me be clear, that does mean I think that environmental factors are irrelevant. Dogmatic hereditarians (who are usually not scientifically trained in any way) try and find genes or biological causes for everything. The on-off tendency in human nature seems to evince itself in the press too, this is not a trait to be found among a narrow sect on the margins. There are constant stories about "genes" for a broad behavioral category. The journalists, let alone the public, seem to believe that we will find the seat of intelligence, anger, anti-sociality and a host other traits in one or two genes. This is almost always not so. Most of these traits that are being "explained" are:

1) Polygenic (many genes impact them).
2) Often additive to the point where one gene contributes to less than 10% of the genetic heritability, and the genetic factor might be the cause of less than 50% of the variation within the population!.
3) Very different genes might lead to the same behavior in the population (follows from 1, but I wanted to be explicit).

The caveats keep on coming.

Solutions? From us:

1) I'll be clearer and assume less.
2) I'll be more cautious.
3) I'll highlight non-genetic factors in phenotype X more strongly so that people don't cherry pick (can't).
4) I'll try to think of the non-intuitive population genetic foundations that underlay some of the principles, which I might take as givens.
5) Instead of linking to definitions of things, I'll try and define them in the text, so that those who forget to double-check the link will know what I'm talking about in the paragraphs that follow and not misunderstand.

From you:

1) Whenever you think "Nature vs. Nature," do a "search & replace" with "vs. or and or times" (in other words, there are more operations than subtraction implied, addition in polygenic traits and multiplication in the gene-environment interaction for instance might be appropriate analogs).
2) Outside of highly lethal diseases like Cystic Fibrosis, abandon the simple dominant-recessive Mendellian Model in your head. If it's a complex behavior, keep in mind that dozens of genes are likely tweaking the overall phenotype, that gene-environment interaction can amplify the magnitude of the expressed phenotype, and that epistasis, the interaction of different genes in the genome itself in producing a phenotype, might also occur.
3) Realize that probabilites are important, genetics is statistical, that the importance of various factors shifts from person to person, dependent on the variables themselves (environment, specific genes, etc.) and we are usually speaking of whole populations.
4) Framing is important, English both under-expresses and over-expresses points (usually depending on personal context). This is partially sloppiness noted above, but the problems are magnified by the filters readers use.

I could go on. The main point is that this isn't particle physics. A neutron has a mass that equals a proton + an electron. A positron has a positive charge. The zoo of particle physics has precise definitions.

When we speak of "heritability" it can get confusing since there are two main definitions (and the common sense default one tends to creep in without much noise). Many of the models we present or paradigms we are using are first aproximations and rough fits of highly complex phenomena. If we saturated our text with the real number of caveats, notations, etc. that come to mind, you might as well go read PubMed.


Posted by razib at 02:17 PM