Sociobiology 3: Kin selection and pluralist explanations

[The third in a series on a recent paper by David Sloan Wilson and Edward O. Wilson. Post 1; Post 2]

In presenting a group selectionist account of sociobiology, Wilson and Wilson argue that alternatives such as kin selection are not really alternatives.

Kin selection and multiplayer games

Attacking the third leg of the tripod, W&W argue that even genic or individual selectionist accounts implicit include the existence of groups as a factor in evolution. This resolves, as far as I can see, to the view that equivalence classes in genetic selection form a group: if a population is composed of alleles A and a, the individuals with one of these alleles socially interact with their conallelics, as well as with the other group.

Second, interactions between trait groups are implicitly group selectionist even if the traits are seen as borne by individuals, and population is structured accordingly. Third, these equivalence classes evolve via interactions between trait groups - that is, there needs to be between-group selection for the population to evolve. I don't quite know what to make of this argument. If I read it aright, it seems vapid. Gene selection accounts of course involve interactions between N individuals, but the mere fact they are equivalence classes doesn't make them selective players as groups.

They claim, possibly with justification given their redescription of the notion of a group, that Hamiltonian inclusive fitness is a kind of group selection, and that Hamilton himself accepted this. But is their definition of a group anything new? That individuals interact in a population, and that fitness is determined by one's neighbours in part, is nothing much new. It's called density dependent selection. For those who don't know it, it works like this:

Suppose you have the eponymous altruists and cheaters in a small population. If altruists outnumber cheaters to a certain ratio (which depends on the payoff matrix) then the occasional cheater will not lower the altruists' payoffs (fitness) noticeably, but the cheater does very well indeed. But in a population of mostly cheaters, the altruist does very badly while the cheaters do less well than in the other population. Now make these populations subpopulations of a larger one... W&W say:

All of these models obey the following simple rule, regardless of the value of N, the duration of the groups, or other aspects of population structure: Selfishness beats altruism within single groups. Altruistic groups beat selfish groups. The main exception to this rule involves models that result in multiple local equilibria, which are internally stable by definition. In this case, group selection can favor the local equilibria that function best at the group level, a phenomenon sometimes called “equilibrium selection”...

Is this group selection? I think that the answer depends solely on one's terminology, and not on any theoretical difference between individual selectionists and group selectionists. Population geneticists using "standard" models accommodate this just fine. It seems to me that the debate has moved from being substantively theoretical to being a matter of words. W&W seem to agree. They say "They [the different models] differ, however, in how they partition selection into component vectors along the way. The frameworks are largely intertranslatable and broadly overlap in the kinds of traits and population structures that they consider." So the next move of W&W strikes me as somewhat unnecessary.

Pluralism



They argue next that the "selfish gene" approach of Williams and Dawkins rejected selection above the individual. I think this is false - Williams himself in 1966 identified a case of group selection of the segregation distorter t-allele in mice, based on a study done by Lewontin and others. But thinking this, they assert

When Williams and others rejected group selection, they were rejecting the possibility that adaptations evolve above the level of individual organisms. This is not a matter of perspective, but a fundamental biological claim. If true, it is every bit as momentous as it appeared to be in the 1960s. If false, then its retraction is equally momentous.

So they discuss three examples of "pluralism", or attempts to give multiple levels of selection. The first is Hamilton's inclusive fitness model of kin selection, the second is Kerr and Godfrey Smith's (two philosophers) distinction between collective and contextual fitness. The third is the failure of the Price equation to properly classify fitness bearers. I will leave you to read this discussion for yourself.

The point of this, however, is that W&W claim that we need a core set of criteria that applies to all cases of putative selection. In effect, they reject pluralist perspectives in favour of a monism, which can be summarised thus:

1. A distinction between group and individual fitness. In short, they deny that these are intertranslateable.

2. Locally or individually disadvantageous traits can only evolve through group selection.

3. There is no general reason to think group selection is a weak or inactive force in evolution. So we must evaluate each case on its own merits.

4. Traits that evolve throughout a population may yet be the outcome of group selection. They say

If we are merely interested in whether a given trait evolves, then it is not necessary to examine levels of selection, and multiple perspectives can be useful. If we want to address the particular biological issues associated with multilevel selection, then we are required to examine the appropriate information and the perspectives converge with each other.

They then appeal to Maynard Smith and Szathmáry's "transitions" concept, which is the next post. Then I will (finally) give my own thoughts on this whole matter.