When I was in graduate school at Duke University in the early 1980s I remember a young professor visiting from Michigan State who gave a talk about group selection. The professor was of course David Sloan Wilson, because at the time he was the only academic willing to stick his neck out for group selection. I thought his ideas were eminently sensible and was surprised to learn that none of the other graduate students agreed with me. Group selection was not particularly close to my research interests then (I was working on population movements of insects), but I kept following it. What I heard during the 1980s and 1990s was a relentless drumbeat against group selection. As a result of books by such luminaries as G. C. Williams and Richard Dawkins, evolutionary biology at the time thoroughly repudiated this idea.
Later, during the 1990s I became interested in scientific study of history (so much that I eventually switched from biology to historical social science) and discovered that the theory of group selection, which by then became the theory of multilevel selection, could provide a very productive conceptual framework for the study of the evolution of complex societies. It turned out that multilevel selection yielded predictions that could actually be tested with historical data. What was particularly exciting was that theoretical predictions, time and again, yielded novel insights that were, amazingly enough, supported empirically (when I started working on historical dynamics I did not realize how much data there are to test different theories).
Meanwhile, there was a glacial, but also tectonic shift (sorry about mixing metaphors) in social evolution, especially in human social evolution. Gradually an increasing number of researchers came over to the view that cultural and genetic group selection provides a very viable theoretical framework for the study of evolution of human sociality. Colleagues like Pete Richerson and Rob Boyd, Sam Bowles and Herb Gintis, Elliott Sober, Joe Henrich, and many others became seriously engaged with models and empirical analyses framed within the multilevel selection paradigm.
The signal event in this tectonic shift was the “defection” of E. O. Wilson to the group selection side in 2007. The two Wilsons co-authored a Quarterly Review in which they proposed the dictum, “Selfishness beats altruism within groups. Altruistic groups beat selfish groups. Everything else is commentary.”
A month ago we ran a workshop at NIMBioS (the National Institute for Mathematical and Biological Synthesis; I will be reporting on it later this week). The workshop was attended by a truly multidisciplinary group of more than 40 scientists. What was really surprising was that the idea of multilevel selection was something all of us could agree upon.
I believe that we are in a Kuhnian paradigm shift, and I fully expect that multilevel selection will become the reigning paradigm in the next 5-10 years. But the transition is not going to happen without pain.
This paradigm shift is associated with a remarkable degree of acrimony (although I did not experience at first hand other paradigm shifts, so perhaps they all are similar in this respect). The case in hand is the exchange over the weekend between David Wilson and Jerry Coyne. It started with a post by Wilson When Richard Dawkins is not an Evolutionist. Coyne shot back with David Sloan Wilson loses it again, and Wilson blasted away with Pugilistic Science.
In the process, Wilson intimated that Dawkins is not really a scientist, while Coyne suggested that “Wilson is totally over the waterfall.” At the end of his post, Coyne added: “I recently did a podcast interview for the Evolution: This View of Life site. Had I known that the biology part of the site was run by Wilson, and is used largely to promote his own views about religion and group selection, I would not have done it.” Now I don’t want to paint Jerry as a blackguard, all parties of this debate have transgressed over the norms of polite discourse, and in a different age this could easily lead to a dawn encounter with matched weapons.
But let’s return to the substantive issue, that of group, or better multilevel selection. Specifically, not whether it is prevalent in the animal world, but its role in the human social evolution. I am of course a partisan in this debate, but I don’t understand the extreme position taken by Richard Dawkins and Jerry Coyne, who deny any role for multilevel selection. If we want to understand the evolution of large-scale complex societies, what is the alternative? How could human ultrasociality, our ability to cooperate in huge groups of genetically unrelated individuals, evolve? The two theories discussed in The Selfish Gene, reciprocity and kin selection, fail utterly on both conceptual and empirical grounds, as has been abundantly demonstrated by the likes of Richerson and Boyd.
To make this question even more specific, we know that humans are capable of sacrificing their lives for the sake of huge groups consisting of unrelated strangers. As an example, think of the Southerners and Northerners volunteering for the Confederate and Union armies during the American Civil War. It is a particularly well-documented case, because those volunteers were literate, and they left behind thousands of letters explaining their motivations. Hundreds of thousands of them died on the Civil War battlefields. So how can you possibly explain such remarkable properties of human sociality, other than with a group selectionist model?
This is not a rhetorical question – I genuinely would like to see what theoretical alternatives there are, so that we can start figuring out how to test them empirically. So let’s hear them.