Sunday, December 08, 2013

The Gene Supernatural

The neo-Darwinian old guard has come out hard against David Dobbs' (admittedly inflammatory) article in Aeon, "Die, Selfish Gene, Die," which summarizes some recent and not-so-recent objectives objections to gene-centric "modern synthesis" evolutionary theory. Evolutionary biologist Jerry Coyne devoted two posts to demonstrating how "muddled" Dobbs' piece was, and Richard "Selfish Gene" Dawkins himself responded on his site that nothing in Dobbs' article contradicted the theory that he laid out in his landmark book The Selfish Gene. Steven Pinker went so far as to use the opportunity to characterize all science journalists as "congenitally" sensationalist.

But PZ Myers, erstwhile ally of the aforementioned gentleman and scholars in their struggle against theism, has come out with two posts at Phayngula strongly defending Dobbs' basic argument that the gene-centric "modern synthesis" is no longer fully supported by molecular biology.

Five-odd years ago, I was engaged with Myers in a brief but bitter squabble, after I dubbed his "Courtiers Reply" argument the "Lout's Complaint." He replied that I was "clueless" and that he was "proud to be a hooligan." I got a flurry of comments from his readers calling me soft-headed, and then it was over.

When it comes to religion, Myers is still a hooligan. (Sorry, Paul!). But I was extremely pleased to see him strenuously and heterodoxically critique the gene-determinism that has erupted among some of his celebrated colleagues.

Dawkins' "selfish gene" is a beautifully elegant theory. It is near-impossible to argue with the logic of its central premise. To read it is to be utterly convinced that nothing but the gene could ever possibly be the unit of selection, and it is extremely valuable in helping to overthrow popular misconceptions of natural selection, such as that "traits survive for the good of the species."

The problem, as Myers shows with relentless detail in his post, is that genes don't operate in anywhere near the idealized fashion that Dawkins describes. I would propose that this is because in nature, "genes" don't actually exist. We can get a sense of this observing the way that defenders of gene-centric theory alternate between incongruous definitions whenever their theory comes under attack. In Chapter 3 of The Selfish Gene, Dawkins famously defines the gene as any portion of chromosomal material that persists long enough to serve as a unit of heredity. A little bit later he expands on this when he tells us not to worry that a complicated trait (like the mimesis pattern on a butterfly) seems too complex to be controlled by a single gene: we can just redefine the gene as whatever cluster of DNA is responsible for the pattern. (He later proclaims the triumph of his tautology: "What I have done now is define the gene in such a way that I cannot help being right.")

Then, rather astonishingly, he goes on to say that his concept of the gene is not definitive in an absolute yes-or-no way, like an electron or an elephant or a comet, but relative, like size or age. A gene may be more or less gene-y, compared to other genes. Dawkins is very explicit that his mission here is to rescue Mendelian genetics by expanding the concept of the particulate unit of heredity to whatever scale it needs to be for the theory to work.

Perhaps ironically, this is the same slipperiness that gives fits to anti-theists whenever the topic of God's causality comes up. What is the nature of God? Whatever it needs to be to explain the perceivable world around us. What is a (Dawkinsian) gene? Whatever it needs to be to explain the transmission of a corresponding phenotypic trait. Little surprise then, when critics poke holes in the theory, drawing on recent (and not so recent) findings in molecular biology, that Dawkins is able to reply, "Why, my definition of the gene can account for that too!"

Among molecular biologists, the gene was for many years typically defined as the portion of the genetic code (also called a cistron) that carries instructions for the manufacture of an individual enzyme. In Crick's phrase: "DNA makes RNA, RNA makes protein, and protein makes us." The mechanics here are much easier to observe than in the Dawkinsian usage, but even here the definition is not as clear as it would seem. DNA sequences often need a lot of "editing" before they are converted into RNA sequences, and there is in fact no one-to-one correspondence between cistrons and proteins. Some proteins get built from an RNA sequence that has no equivalent in the DNA. This presents some pretty tough challenges for any theory that proposes that heredity is strictly a "genetic" phenomenon, unless we are prepared to count as "genes" any number of factors that are not stored in DNA, and whose manner of hereditary transmission, if it exists at all, is unknown. (Note how Jerry Coyne--who Dawkins calls his "goto guru on population genetics"--bases his entire rebuttal on the notion that regulatory factors "must" reside in the DNA, which seems to indicate that theoretical population genetics has become seriously unmoored from molecular biology).

There's much more to the story: epigenetics, evo-devo, genetic assimilation, and genetic redundancy, much of which you can read about by clicking on the Pharyngula links above. The point I want to make is that we can go one further than David Dobbs and the scientists whose work he summarizes. It's not just that "Selfish Gene" biology is overly gene-centric and deterministic. It's that the central metaphor of that paradigm is based on a spook. The "gene" is barely even a coherent concept, let alone a natural entity that could have causative powers. For a century it has convoluted the way we think about morphology and heredity. If we were feeling especially uncharitable, we might even be tempted to call it ... a Delusion.