Evolution for Everyone

In their book Darwinism Evolving, David J. Depew and Bruce H. Weber make the interesting point that pre-Darwinian notions did not come to an abrupt halt with the advent of Darwin’s theory. Instead, they often became repackaged in superficially Darwinian terms.

That certainly applies to notions of adaptation in nature and human society. Before Darwin, most people regarded nature as the creation of a benign God. Of course it must be adaptive, from top to bottom! Human society must also be part of God’s plan, however inscrutable. As the Anglican Bishop Joseph Butler (1692-1752) put it, “It is as manifest that we were made for society and to promote the happiness of it, as that we were intended to take care of our own life and health and private good.”

Against this background, the original problem identified by Darwin and his partial solution (see T&R II) must have been disorienting. He was suggesting that adaptations might be restricted to individual organisms and that society might merely reflect their conflicts of interest. His partial solution meant that adaptations might exist above the level of individual organisms, but only if special conditions are met. In modern terms, adaptation at level x of the biological hierarchy requires a corresponding process of selection at the same level and tends to be undermined by selection at lower levels.

After Darwin, many biologists continued to assume that adaptations evolve at all levels — for the good of the individual, group, species, or ecosystem — without requiring special conditions. Social theorists continued to portray human society as like a single organism in the functional integration of its parts. When the problem identified by Darwin was acknowledged, it was common to assume that higher-level selection easily trumped lower-level selection. Today, these assumptions are labeled naïve group selectionism.

Naïve group selectionism is not dead. If you think that diseases evolve to avoid killing their hosts, that animals evolve to manage their population size, that ecosystems evolve to efficiently recycle nutrients, that nature left undisturbed achieves a harmonious balance, that earth’s entire biota qualifies as a single organism (the Gaia hypothesis), or that human society can be compared to a single organism, including technology leading to a single global brain, then you are a naïve group selectionist. You might be right, but you are not paying sufficient attention to Darwin’s sobering message that special conditions are required.

Everyone who teaches evolution knows that a large fraction of students start out as naïve group selectionists who are likely to utter phrases such as “for the good of the species” unbidden — even as the wrong answer on the final exam. There is something about adaptation above the level of the individual that just seems right, even when your teacher tells you it’s wrong.

A few biologists who followed in Darwin’s footsteps were not naïve, especially the three founders of population genetics theory, Ronald Fisher, J.B.S. Haldane, and Sewall Wright. Like Darwin, they had a lot on their minds. In their effort to place all aspects of evolutionary theory on a mathematical foundation, group selection was just one problem among many. Nevertheless, each clearly recognized that most social adaptations are locally disadvantageous. Either they pass out of existence, or they evolve on the strength of a selective advantage at a larger scale. Fisher, Haldane and Wright sketched a few models to illustrate the point but group selection did not occupy center stage and mathematical models were a foreign language to most biologists in any case.

To illustrate the influence of naïve group selectionism in the middle of the 20th century, here is the final paragraph of the most influential ecology textbook of the period, Principles of Animal Ecology (1949), authored by W.C. Allee, A.E. Emerson, O. Park, T. Park, and K.T. Schmidt, affectionately known as the great AEPPS.

The probability of survival of individual living things, or of populations, increases with the degree to which they harmoniously adjust themselves to each other and their environment. This principle is basic to the concept of the balance of nature, orders the subject matter of ecology and evolution, underlies organismic and developmental biology, and is the foundation for all sociology.

This passage, and the whole textbook, is suffused with the notion that nature evolves to be adaptive from top to bottom.

Enter George C. Williams, a tall man with the craggy features of Abe Lincoln or the statues on Easter Island. He was not mathematically trained but had learned the lessons of population genetics as a graduate student at the University of California at Berkeley before accepting a postdoctoral position at the University of Chicago in the late 1950’s. There he attended a seminar by Alfred Emerson, one of the great AEPPS, a termite biologist who regarded all of nature as like a termite colony. As George recalls the event, “if this was evolutionary biology, then I wanted to do something else — like car insurance.” George left the lecture muttering “Something must be done…”

A great reckoning was about to take place.

To be continued.


  1. #1 daedalus2u
    October 25, 2009

    Very nice discussion of naïve group selection; there is an analogous naïve conceptualization of physiology which uses the term coined by Walter Bradford Cannon, homeostasis.

    Many analogies can be drawn between the naïve physiology of homeostasis as typified by “to keep things constant” and naïve group selection as typified by “for the good of the species, ecosystem, biome”.


    I think naïve physiology and naïve group selection both fail because they don’t consider the path by which the final state occurs. They observe the final state, and then try to fit details to that final state in a “top-down” manner (extrapolating backwards from the final state), rather than trying to build a path from the “bottom-up” from more fundamental processes known to actually happen.

    In a way, the idea of homeostasis evolving “for the good of the organism” is like the idea of individuals evolving “for the good of the species”. Complex physiological pathways can only evolve a little bit at a time, essentially one gene at a time. Similarly the gene pool of a species can only change one organism at a time.

  2. #2 InfuriatedSciTeacher
    October 25, 2009

    Nice treatment of naive group selection. One of my issues with group selection in general is that it provokes ideas such as the Gaia hypothesis without any real attempt to verify that hypothesis with empirical evidence. The adverse consequence doesn’t negate the possibility that group selection does in fact take place in some instances, but pure theory without solid empiricism behind it falls short. Hopefully I’m not becoming tiresome, as I am actually attempting to have a rational discussion about this rather than being a troll. And I’m more inclined to accept evidence in organisms that are truly social, wherein the phenotypic trait examined can be considered an emergent property as opposed to something that’s clearly reducible to a single gene. Headed in that direction, perhaps?

  3. #3 InfuriatedSciTeacher
    October 25, 2009

    In reading the articles suggested by Mike in TR1, I find that I have perhaps held an inaccurate (archaic?) view of group selection theory. Multi-level selection, in the sense that individual fitness can be increased by characteristics of the group is certainly plausible if examined on a case by case basis. I can also see from the 2007 article written with E.O. Wilson that I’m not raising any concerns that are, in any sense, new. If properly evidenced, multi-level selection would also seem to provide an explanation for speciation in absence of geographic isolation, presuming a low enough level of genetic drift between neighbouring populations or differing in-group selection pressures (altruists would fare poorly in a selfish group, yet a group of altruists could evolve enough difference within their own population for two related but separate species to occur with different allelic frequencies for other traits).

  4. #4 David Sloan Wilson
    October 26, 2009

    I’m glad to see a lively and informed discussion develop. I will reply to the comments at regular intervals, although not necessarily daily. I will post the installments at daily intervals to keep up momentum.

  5. #5 piker
    October 26, 2009

    “(altruists would fare poorly in a selfish group, yet a group of altruists could evolve enough difference within their own population for two related but separate species to occur with different allelic frequencies for other traits)”
    Assuming that altruism is truly genetic and not a strategic option directed by circumstance – in which case a group of altruists would likely cooperate themselves to an early grave.

  6. #6 abb3w
    October 26, 2009

    One difficulty is that most models use a smooth approximation to what is ultimately a discrete process. This is ultimately inherited from the transition from statistical mechanics to thermodynamics. While the mass may be approximated as moving smoothly from a distance, the individual particles travel by discrete drunkard’s walks.

    David Sloan Wilson: In modern terms, adaptation at level x of the biological hierarchy requires a corresponding process of selection at the same level and tends to be undermined by selection at lower levels.

    I think “at the same level or higher” might be more exact. EG: gene and cell adaptations happening as a result of selection at the level of the organism.

  7. #7 piker
    October 26, 2009

    Even if “altruism” were to be a gene adaptation at any level, it would have to be a malleable trait, or suffer consequences not contemplated in the artificial models where malleability is not made a part of the algorithm.

  8. #8 abb3w
    October 27, 2009

    Would you be so kind, piker, as to express decidable criteria for “malleable” for those unfamiliar with your local terminology?

  9. #9 piker
    October 27, 2009

    My local terminology? Funny how that seems to rhyme with yokel. Just google malleable trait for a range of decidable criteria, but in the mean time consider “amenable to learning” as one such.

  10. #10 Alan Tabor
    February 25, 2010

    This is a question for DSW. I’m not sure if comments to older posts actually get seen but it’s worth a shot.

    First off, I’ve very much enjoyed and appreciated reading your work here and in Unto Others, the Quarterly Review article with EO Wilson, etc. Thanks.

    I’m reading _Group Selection_ edited by Geo. Williams and published in 1971 but consisting primarily of reprinted articles collected from the mid-60s including some authored both by Maynard Smith and Wynne-Edwards.

    In a 1964 Nature article, Wynne-Edwards explicitly defends his work in terms of the competition between groups and seems quite aware of what’s needed to make it work when it crosses purposes with individual selection. (I’d be happy to upload a block of his text if that would be useful.)

    So what’s naive about naive group selection theory? The construct seems consistently expressed through the last 40 years. Is it that acceptance pre-60’s was too facile and the naivete was in the scientific community and not the theory itself? Was it the lack of adequate mathematical underpinning? Or what?

    I’m having trouble connecting the pieces.

    Thanks much in advance.

New comments have been disabled.