20 June 2012

Group (selection) therapy

ResearchBlogging.orgCan natural selection act on groups of individuals?

This is a decades-old argument in evolutionary theory that many thought was settle late in the twentieth century, but it has recently reared its head again.

One of the lead proponents of resurrecting group selection has been E. O. Wilson. You can read reports of Wilson’s recent presentations here and here. Steven Pinker recently jumped into the fray.

Almost everything that I read on both sides of this issue is highly conceptual stuff. Models are formulated. Reasons are advanced. It’s all very abstract and with a lot of jazz hands. It’s been frustrating for me, and keep in mind, I like this stuff. I’m co-teaching a graduate class in evolution right now. I can’t imagine what a non-specialist makes of this stuff.

Largely absent in these discussions are terms like “hypothesis,” “prediction,” and “test.”

What I would love to see would be for those in this argument to say, “Here is an organism with that routinely lives in groups and benefits from being in a group. If group selection is occurring, we would expect to see this much more reproductive success than you could account for by individual success alone.”

If not reproductive success, gene frequencies, or some other measures.

And I would love to see these experiments made for animals other than humans. There are too many pitfalls that can trap even the wariest researcher when thinking about humans. Meerkats might be a good case study, if this interview with Tim Clutton-Brock is anything to go by. He rejects kin selection as the main explanation for cooperation in meerkats.

I started off thinking that the whole thing ran through kinship. Kinship is obviously there but these group-related benefits are obviously fantastically important, and my suspicion is that in most of the highly co-operative mammals, things like naked mole rats and African wild dogs, the same pattern holds. They're related to each other, but there are lots of other social mammals that are related to each other which aren’t advanced co-operators. And the unusual thing of the advanced co-operators is that they live in habitats and they live in ways where they’re really dependant on other members of the groups.

Note, though, that “benefiting from being in a group” – which Clutton-Brock is describing – is not the same as “group selection.”

Those who are advocating that group selection is a major evolutionary force need to articulate a research program.

The value of having a research program is often underestimated. It’s not enough to be right in principle; you have to suggest experiments that you can do to test them. A lot of experiments.

For example, there are a lot of different ideas out there for how to define species. But the biological species concept (species are non-interbreeding) suggested a research program while other species concepts did not (Coyne and Orr’s Speciation). The discovery of a major impact at the K-T boundary ( Alvarez et al. 1980)suggested a research program. Eldredge and Gould’s punctuated equilibrium (1972) suggested a research program, and lots of productive science arose from that. Gould and Lewontin’s Bauplan (1979) didn’t readily suggest experiments, and languished.

Wilson developed a research program when he published Sociobiology (1975). That book (perhaps along with The Selfish Gene; Dawkins 1976) galvanized animal behaviour. Ethologists probably spent a good 15 years or so exploring and testing ideas arising from those books.

I have only vague ideas about what a group selection research program would look like. But then, it’s not my job to spell out the research to do: that burden lies with those who are advocating group selection is important in evolution.

Additional, 24 June 2012: The Guardian has noticed. They frame this as a “Let’s watch the famous guys fight,” though, which is disappointing.

Additional, 25 June 2012: Jerry Coyne is disappointed by The Guardian piece.


Alvarez LW, Alvarez W, Asaro F, Michel HV. 1980. Extraterrestrial cause for the cretaceous-tertiary extinction. Science 208: 1095-1108. doi: 10.1126/science.208.4448.1095

Coyne JA, Orr HA. 2004. Speciation. Sunderland, MA, Sinauer Associates.

Dawkins R. 1976. The Selfish Gene. Oxford, Oxford University Press.

Eldredge N, Gould SJ. 1972. Punctuated equilibria: An alternative to phyletic gradualism. Models in Paleobiology. TJM Schopf (ed.), pp. 82-115. San Francisco, Freeman, Cooper and Company.

Gould SJ, Lewontin RC. 1979. The spandrels of San Marco and the Panglossian paradigm: A critique of the adaptationist programme. Proceedings of the Royal Society of London. Series B. Biological sciences 205(1161): 581-598. DOI: 10.1098/rspb.1979.0086

Wilson EO. 1975. Sociobiology: The New Synthesis. New York, John Wiley.

Photo by Digimist on Flickr; used under a Creative Commons license.


miko said...

As useful and essential as mathematical modeling is, it can easily stray too far from biology to be relevant. The Nowak paper is the evo equivalent of the computational neuroscience models we're always supposed to be interested in even though they bear no resemblance to the organization or functional properties of nervous systems and neurons. Mathematicians are by nature not empirical.

I say let them have their cubicles and Mountain Dew and you-are-your-connectome and uploading and singularities. It's only an issue once in a while when they convince someone like E.O Wilson to take them seriously.

Bjørn Østman said...

Zen, to what extent do you think the research program has to be suggested by those who theorize? What if others suggest hypotheses and test them? Is the problem that it gets too decentralized?

Patrick said...

There has been SOME empirical work on the issue. Goodnight and Stevens (1997; Am Nat 150:S59-S79) reviewed experimental and observational studies of group selection in great detail (I'll note that some of these studies have been criticized by Coyne and others). Nonetheless, check out the papers citing G&S 1997 for a few more examples.

I agree that more empiricism is needed, but I also think that theory is useful in its own right. Evolutionary dynamics are not always intuitable without the help of mathematical models. Plus, population genetic approaches that could quantify multi-level selection in natural populations are inherently mathematical. E.g., Bijma et al (2007; Genetics 175:277-288) developed a mathematical framework for dealing with responses to multi-level selection with a strong emphasis on testable predictions.

Regardless of the above points, my answer to your initial question is an emphatic ‘yes!’ if for no other reason than the ambiguous nature of ‘individuality’. What constitutes an individual (vs. a group of individuals) in corals, slime moulds, plants that reproduce both asexually and sexually, or the Portuguese man o' war? Similarly, the evolution of multicellularity (multiple independent times, I believe) seems to me to REQUIRE a transition from cell-level selection to selection at the level of groups of cells.

Zen Faulkes said...

Bjørn: Well, they do say that the burden of proof is on the claimant. ;)

I don't think it's the sole responsibility of the theorists to lay out the entire research program. I think the secret to the examples I listed was that the kinds of experiments to do were sort of obvious.

In the case of group selection, I would say the theorists advocating that group selection happens have a little more responsibility than usual to suggest predictions and tests. First, the idea has been controversial. Second, it is tricky, and careful theorists could help experimentalists avoid pitfalls. I used the Clutton-Brock example, because I could easily see how some people could think the situation he is describing implies group selection. It may be - but there's not enough evidence from that quote.

Patrick: True, and I neglected to mention some of those published examples, because I am a lazy blogger. (Also, I sold a book that had references to one study that was very convincing to me.) I see a new paper on "multilevel selection" came out yesterday that I might review here.

My impression is that the campaign to get group selection back on the agenda of evolutionary biology started in earnest maybe... five years ago? with people like David Sloan Wilson and E.O. Wilson.

In that time, the ratio in discussions has been skewed towards pontification and not empiricism. But my impression could be wrong, as I don't follow the literature on social animals incredibly closely. Maybe there, things are bubbling over with new experiments.

trashbird1240 said...

There may be a pseudo-political reason that no one has produced any studies on this, other than theory. Some people (who might sit on boards of funding agencies) turn bright red and steam comes out of their ears when people mention group selection, or even "multi-level selection." They show you cartoon lemmings and that passes for refutation of a concept that, in theory, has no real problems. As a theorist, I see most of the problems people have with group selection as driven by passion, not logic. Considering the career stature of the people who proposed group selection as an explanation of eusociality, I'd say that supports my point. I doubt anyone could get a postdoc doing this.

Michael said...

I wonder if research couldn't focus on some emergent behaviors from groups, something the group does that isn't just an additive property of the members - this may be why Wilson, who works with ants, is such a proponent of the idea. He must have seen a lot of colony behaviors that seem to benefit the group to the possible detriment of the individuals, but which select cooperative individuals in extreme circumstances. Think the formation of "living rafts" versus the individuals fleeing above flood levels, maybe...