Friday, 13 July 2012

The problem with group selection

Steve Pinker has written a very welcome critique of group selection over at the Edge. His verdict is damning: "the concept of group selection is making a shambles of the application of evolutionary biology to human psychology". I concur -- mostly because it has never been clear what problem ‘group selection’ is supposed to solve.

Evolutionary theory has provided us with numerous explanations of social, cooperative and altruistic behaviour -- including kin selection, mutualism, reciprocity, conflict resolution and costly signalling. These theories already seem more than capable of explaining most of what humans get up to; and their full potential has yet to be realised. Good old-fashioned mutualism, for example, has received relatively little attention, but provides the most straightforward explanation of human groupishness.

Mutualism has to do with situations in which individuals do better working together than they could do alone, and in which the opportunity to free-ride is not available. Darwin provides a typically charming example: "Hamadryas baboons turn over stones to find insects, &c.; and when they come to a large one, as many as can stand round, turn it over together and share the booty" (Darwin, 1871). The point is that, but for the concerted effort of all the baboons, there would be no "booty" to share. No doubt a group of baboons that coordinate their activities in this way will do better than a group of baboons that do not. But nothing is added by calling this ‘group selection’. And much is lost -- an appreciation of the specific selection pressures and adaptations responsible for the behaviour.

The situation is no different if the challenge that the baboons face is not a heavy rock, but a rival group of hostile baboons. Certainly, the presence of rival group A can re-align the interests of members of group B, such that they no longer face a dilemma over whether to cooperate: it's a case of "We must all hang together, or assuredly we shall all hang separately", or what has been called "the soldier's dilemma" (Clutton-Brock, 2009). Humans have no doubt been subjected to hundreds of thousands of years of lethal intergroup competition, with the ensuing ‘coalitionary arms race’ selecting exquisite adaptations for forming and maintaining groups. But again, nothing is added, and much is lost, by calling this battle of the mutualists ‘group selection’.

Of course, the situation is more interesting if individuals are sacrificing themselves for the benefit of their groups -- that is, if they are being altruistic. Here we have a problem in need of explanation. But ‘group selection’ is not so much an alternative explanation of altruism, as an alternative way of accounting for the effects explained by existing theories. And it is not clear what problem is solved if, instead of dividing up selection pressures ‘vertically’ according to adaptive problems and their solutions, we divide them up ‘horizontally’ according to their net effects at different arbitrary ‘levels’.

At best, ‘group selection’ refers to a quirk of a fission-fusion population structure that can result, by chance, in the positive assortment of altruists. But although this can happen in theory, there are few if any actual examples of it happening in practice. As such, group selection should be given the attention that it deserves -- that is, very little.

Clutton-Brock, T. H. (2009). Cooperation between non-kin in animal societies. Nature, 462, 51-57. 
Darwin, C. (1871). The Descent of Man and Selection in Relation to Sex. London: John Murray.

No comments: