Dawkins the Atomistic Monist: Selfish Gene, Chap 3

In the third chapter, called "Immortal Coils," we are inserted into a debate about the "unit of selection" in natural selection, which Dawkins is kind enough to spare us the details. Mainly, it seems, we don't need to think about other positions (he's "dispatched" one position - group selection - in the first chapter). However, Dawkins' argument "is an argument which some of my most respected colleagues obstinately refuse to agree with [!], so you must forgive me if I seem to labour it!" (26) Actually, he doesn't really labor his point at all. He spends just a few paragraphs justifying his insistence that "genes" are the unit of selection in natural selection.

Dawkins is an advocate of "gene selectionism" (although in the Extended Phenotype he allow for selection at the level of the ogranism, i.e. that they can be explained at multiple levels). This is the notion that the gene is the thing that natural selection works on. The involves an argument about what genes are, and an argument about what is needed in order for natural selection to be the case. What I will show is that Dawkins really begs the question of the unit of selection, and that his is not the only (or the best view) for the "unit of selection."

First, his definition of a gene: "I am using the word gene to mean a genetic unit which is small enough to last for a large number of generations and to be distributed around in the form of many copies" (34). Second, what is needed for natural selection to work: "Natural selection in its most fundamental form means the differential survival of entities. Some entities live and others die but, in order for this selective death to have any impact on the world, an additional condition must be met. Each entity must exist in the form of lots of copies, and at least some of the entities must be potentially capable of surviving - in the form of copies - for a significant period of evolutionary time" (35). Dawkins does ackowledge that "what I have now done is to define the gene in such a way that I cannot really help being right! (35)," but then he goes right back and says, "what I have done is to define a gene as a unit which, to a high degree, approaches the ideal of indivisible particulateness" (35). This ideal is needed because of his assumption of what is needed for natural selection. In philosophical parlance, you might call this a vicious circle, i.e. using your assumptions to define a thing that confirms your assumptions. The fact that Dawkins admits he's doing something like this, and then going on to some pretty unqualified statements, is just bizarre.

But let's think about his assumption of what's needed for natural selection. I'm gonig to examine this from the more specific to the more general, i.e. what's needed qua Darwin's theory (and other scientist's view of what's needed), to Dawkins' general philosophical presuppositions.

1. What's needed for natural selection?
Dawkins says you need a few things: first is differential survival. A lot of scientists, apparently, use the phrase "differential reproductive success" (an updated version of "survival of the fittest"), but Dawkins' doesn't. Instead of "reproduction" he uses the phrase survival of entities, presubably so this leaves open the idea that "did sex, that bizarre perversion of straightforward replication" (46), is really derivative from "straightforward" replication (although, given David Hull's distinction between replicators and interactors, it was never that "straightforward").

Be that as it may, what is needed for natural selection is (a) an entity that displays a variation, (b) the selection of one of these variations, (c) lots of "copies" of the entity. Richard Lewontin calls these three ideas "variation, selection, and transmission" (Sober and Lewontin, 158). This seems like pretty much the same thing, except for (c). Dawkins says that you need something that will replicate itself faithfully, over long generations, something almost "immortal" (which is what he says of genes, 36), instead of individuals, which are "too large and too temporary" to "qualify as a significant unit of natural selection" (36). Genes last a really long time, according to Dawkins, and again, "approaches the ideal of indivisible particulateness" (35). For Dawkins, you need to have some stable substrate underneath all that change, like the Arsitotelean subject, and organisms are too ephermeral.

Now, there are some serious questions with this argument. First of all, why is it that classical Darwinian theory holds that the individual organism is the unit of selection? Clearly this is because despite what the genes do, the organism is what "survives or dies." Selection pressure is on the phenotype, and while the genotype is hugely important, selection just does not happen at that level. What it takes is for an organism to be put out there in an environment, where the pressure is displayed, and this pressure will weed out some individuals and not others (although this is not all - there is random genetic drift, as well as non-adaptive phenotypic traits that eventually get used later on, or not).

Second, Dawkins' claim that genes are "immortal" is not true. If we take genes to be a material entity, they are obviously connected with the life-cycle of the organism. When the organism dies, clearly the genes go with them. Lewontin's definition makes more sense. The "transmission" for him has to do with parents and offspring (and according to Peter Godfrey-Smith, that's really all that's required for transmission - you don't need a stronger "exact-copy" replicator). This makes complete sense. Say you have a population with a certain gene pool. Even though "discrete individuals" (in Dawkins' terminology) are needed for natural selection, as long as there is something that is inherited, there will be enough likeness between the parent and the offspring to "select." Also, remember we're dealing with evolutionary time too - meaning thousands, millions of years. The genes that are passed on, while a hybrid, are finite (since the gene pool is finite), and are even more limited depending on population sizes and locality.

Third, as Stephen Jay Gould says (and he's not the best to bring in here, since the profession row between the two), gene selectionism engages in a "confusion of bookkeeping with causality." Genes certain do record all the changes that selection does, but that doesn't mean genes are the cause of these changes (although they may be - it depends). In other words, there is not necessarily causation when there is correlation.

2. The Monistic Atomist
I haven't really dealt with a lot of substance of this chapter, but what really interests me is Dawkins philosophical perspective. On the one hand, he seems to think that the "ideal" is discrete particulates as the thing that is actually "real," in the sense of what changes. This is an atomist position (even if, as he repeatedly points out, saying what a gene is is a continuum - a unit of genetic material, like a group of DNA on a chromosome, etc), because it sets up what you would need in an explanation as a simple part. In other words, in the order of explanation, Dawkins thinks that we get to a good explanation of what a thing is when we get to its smallest unit.

But inspite of this atomism - which is a form of reductism - Dawkins is also, oddly enough, a monist. This is, in my view, his sense of genes as "immortal coils." The reason they are is because instead of really treating genes as "atoms" (which really do never die - because they always remain the exact same), the material of genes certain does turn over, but it's the "information" that stays the same. This information is a reification - that is, an artifact that is treated as the thing itself. In other words, the "information" (or the coding process) is a result of the material, not the other way around. Dawkins treats it the opposite way. This information because the "one substance," in a strange sense, because its the thing that's hanging around, despite the fact that the material is no longer existence (because the gene is in a hierarchy of the organism) once the organism dies. Genes thus become the one explanation for everything, sort of like other great monistic principles ("Fire" for the Stoics, "Spirit" for vulgarized Hegelians, etc.).

I find both atomism and monist completely problematic. When can you say you've given a full account of a thing? In science, this is when you can give satisfactory explanation for is material and efficient causes, and then predict with relative accuracy (make mathematical models and such) how things would work in the future. The second half I have no idea how to do, and I must rely on scientists to figure out the math (and there has been a lot of criticism on that front against gene selectionism). I do think that, even in my in-expert mind, the Dawkins paradigm would woefully fail the first use of science. In a world as diverse as ours, in the many ways things have to be in the world, reducing them all to the genetic level is such a restriction in one's view of the world it seems laughable. Why would you want to do that? Dawkins this its the "astonishing" truth (ix). But I'm pretty skeptical of thinkers touting their own "discovery" of "astonishing" truths.