UPDATE: I forgot to add a good piece by Steve Pinker which is required reading if you’re being seduced by the idea of group selection. “The false allure of group selection” is published on the Edge website. It’s followed by an online “discussion” involving 23 Edgies.
__________
E. O. Wilson has a new book out, The Meaning of Human Existence, which I’ve mentioned briefly (I haven’t read it). Last year he published another book, The Social Conquest of Earth, which I reviewed in the Times Literary Supplement (TLS; reference below). Since TLS reviews are behind a paywall, but I retain the copyright, I’ve decided to post it here, for I see Wilson was already developing themes in that book that he continues in the new one. In particular, note what I summarize in the second paragraph, which seems like the nucleus of Wilson’s new book.
Be aware that there is one new paper that claims to find group selection in colonial-nesting spiders, so my statement below that there are no examples of the process in nature may be revised someone (but only to the extent that we have one possible example of the process). It’s a complicated paper, and I’ll report on it when I’ve had time to read and digest it.
This is a bit longer than my usual posts, but since readers seem to object to “the fold,” I won’t use that device further except in rare circumstances.
*****
GENES FIRST
Jerry. A. Coyne
The reigning expert on social insects–particularly ants –Edward O. Wilson is best known to the public for his work on the evolution of social behaviour in humans and other animals, and for his unflagging efforts to conserve natural environments and biological diversity. His book Sociobiology: The New Synthesis (1975) was a milestone in applying evolution to sociality, creating a zeitgeist that helped spawn the field of evolutionary psychology.
In The Social Conquest of Earth, Wilson sets out to explain what makes us human, and to answer the fundamental questions of where we come from, what we are and where are we going. He is clear on where the answers lie: not in philosophy or the humanities, and certainly not in religion, which he sees as purveying “unsupportable claims about supernatural causes of reality”. No, the answers must come from biology, since, to Wilson, human nature is essentially a product of evolution. And he sees the most critical aspect of human nature to be our conflicted status as both selfless and selfish creatures. While we may intercept bullets to save our loved ones, co-operate to build houses for the homeless and drop money in a beggar’s cup, we also cheat on our spouses and our taxes, and battle with others for money and status. How can evolution explain these contradictions?
Wilson argues that these conflicting tendencies result from fundamentally different forms of natural selection. Explaining selfishness is simple: it’s the product of traditional Darwinian “individual selection” in which genes that helped our ancestors outcompete other individuals—to eat more food, find better mates, or simply kill each other— would leave more copies than genes promoting self-sacrifice.
But how, then, could altruism evolve? Behaviours that involve sacrificing your life or reproductive ability for others would seem to contravene natural selection, which, after all, involves leaving more descendants than others. An important caveat here is that cooperation is not the same thing as altruism. Many forms of co-operation produce immediate benefits to the co-operator and so can evolve by classic natural selection on individuals. Female lions hunting together, for instance, can kill larger prey and eat more meat per individual than solitary hunters can. A fish that belongs to a school is less liable than a solitary fish to wind up as a predator’s lunch. Many aspects of group living confer such direct benefits, so “selfish genes” don’t invariably produce selfish behaviours.
To explain “true” altruism, in which individuals risk their lives and reproduction for others, evolutionists have suggested two scenarios. Though these are widely accepted by biologists, The Social Conquest of Earth flatly rejects both. The first, “kin selection”, is based on the simple fact that relatives share genes. This means that genes promoting costly altruistic behaviour in their carriers can sometimes spread because they promote the survival of gene copies in relatives. A gene that made me lay down my life to save three brothers, for instance, would leave more copies than an alternative gene favouring self-preservation. The idea of kin selection has been enormously productive in evolutionary biology, explaining not only altruism towards relatives, but behaviours as diverse as parent–offspring conflict, sibling rivalry, spite, animal dispersal, and virulence in disease-causing microbes.
The second process, “reciprocal altruism”, involves the short-term sacrifice of some benefits for the sake of forming longer-term relationships with others who ultimately return greater benefits. If you give your surplus food to others or lend them money in times of hardship, they might return the favour when it’s your turn to be needy. Because such behaviours give net benefits to each partner, they can also evolve via standard natural selection. But you’d expect to see them only in species in which individuals can recognize and remember who is helpful and who is not – or something equivalent such as occupying a stable patch of ground that acts as a proxy for individual recognition. This is indeed the case: reciprocal altruism, involving acts like sharing meat and forming coalitions, is seen in primates like baboons and chimpanzees that live in fairly stable groups.
Having used these ideas in the past to explain social behaviour in animals, in The Social Conquest of Earth Wilson abandons them completely in favour of an older, and largely discredited, evolutionary scenario: “group selection”. According to this view, entire groups of animals compete with other groups for dominance within a species. Under certain conditions this can promote the evolution of traits, like altruism, that are seemingly bad for individuals but good for the group. To the new Wilson, this is the wellspring of our better nature: “Nevertheless an iron rule exists in genetic social evolution. It is that selfish individuals beat altruistic individuals, while groups of altruists beat groups of selfish individuals”. And he sees group selection as a cause of more than just altruism: it’s supposedly responsible for human traits like communication, complex culture, morality, tribalism, notions of duty and honour, and even religion and homosexuality. As the blurb proclaims, “group selection can be the only model for explaining man’s origin and domination”.
Although Wilson pushes this view hard – it’s the book’s centrepiece – he is probably wrong. Most biologists have rejected group selection for two reasons: it doesn’t work well in principle, and, more important, there’s no evidence that it has been of any significance in evolution. For an obvious reason, selection among groups is far less efficient than selection among genes: genes replicate and replace other genes much faster than groups of individuals divide and replace other groups. Evolving all the social traits on Wilson’s list via group selection requires a slow and unrealistic sequence of episodes in which human populations replaced each other, each replacement based on one or a few behaviours. Further, once group selection fixes a disadvantageous trait like altruism within our species, individual selection proceeds to undo it within populations (“selfish individuals beat altruistic individuals”). In other words, altruism that evolved by group selection is unstable and should disappear.
If the better angels of our nature really are based on genes that evolved – rather than on non-genetic aspects of culture – then they are much more likely to have done so by individual than by group selection. This becomes even more plausible with a more detailed look at how human altruism really works. It is preferentially directed towards friends and relatives, there is much concern with reciprocity and one’s own reputation, and psychological studies show that we dislike cheaters who hurt us personally much more strongly than we dislike cheaters who hurt our group. Such behaviours are precisely what you’d expect if altruism evolved by individual rather than group selection. But Wilson ignores these problems.
It is not even clear that altruistic groups of humans would beat non-altruists. Steven Pinker has noted that success of one group over another in the real world is based not on higher frequencies of altruistic individuals, but on matters like harsher discipline, better technology, and more brutal ideology. Indeed, altruistic groups may be more easily defeated because their empathy for the weak makes them susceptible to domination. But the most important problem is this: I know of not one evolved behaviour in any species that is harmful to individuals and their genes but good for their social groups. In the end, Wilson’s invocation of group selection is superfluous.
While the bulk of The Social Conquest of Earth is about human nature, a good chunk deals with Wilson’s personal area of research: the social insects, particularly those “eusocial” ones, like bees and ants, that show a division of labour between “castes” and have a fertile queen whose brood is raised by sterile female workers. The parallel with human culture here is the altruism of the workers, who sacrifice their own reproduction for the sake of their mother’s.
Virtually all evolutionists see eusociality as the product of kin selection. By helping their mother produce fertile brothers and sisters, sterile workers can actually leave more copies of their genes than if they reproduced themselves. This explanation has in fact been tested and confirmed several times. Despite that, Wilson still prefers a combination of individual and group selection, rejecting kin selection not just in this case, but in general: “The foundations of the general theory of inclusive fitness based on the assumptions of kin selection have crumbled, while evidence for it has grown equivocal at best. The beautiful theory never worked well anyway, and now it has collapsed”. He’s wrong on all counts.
Wilson (with two co-authors) proposed his alternative view of sociality in insects in a paper that appeared in the journal Nature in 2010. It was immediately criticized in five published letters signed by 156 authors –including almost every luminary working on the evolution of social behaviour – which emphasized the value of kin selection and the intellectual sterility of Wilson’s group-selection approach. (I was among those critics.) Wilson fails to mention this criticism in The Social Conquest of Earth. For a scientist trying to explain what we know about human evolution, ignoring such serious dissent is not only self-serving but irresponsible.
Wilson’s book, however, is not devoid of merit. There are interesting titbits about biology and anthropology, including fascinating descriptions of how diverse cultures divide up the colour spectrum in similar ways, and how incest taboos, which avert genetically based birth defects, are enforced even by cultures that don’t understand the genetic consequences. Yet the good bits are ultimately scuppered by Wilson’s attempt to feed questionable biological ideas to the public while ignoring the criticisms of his peers. The result is that readers will be seriously misled about human evolution and the evolution of social behaviour as a whole.
It is puzzling that, at the end of a distinguished career, Edward Wilson has chosen to repudiate fertile and long-standing ideas about evolution in favour of alternatives that are deeply flawed. His immense achievements have made his legacy secure, but it will be tarnished by this misguided attempt to explain social behaviour in insects and humans.
____________
Coyne, J. A. 2013. “Genes first” (Review of The Social Conquest of Earth by E. O. Wilson). Times Literary Supplement 4731 (1 Feb. 2013), p. 32.
Interesting review.
Of related interest, this paper on the historical origins of the idea of eusociality & says R.A. Fisher & Leonard Darwin (the latter of whom I am particularly interested in) presaged Nowak et al
R. A. Fisher and Social Insects: The Fisher-Darwin Model of the Evolution of Eusociality
Owen, Robin
Biological Theory, 2014, Vol.9(3), pp.347-356
http://link.springer.com/article/10.1007/s13752-014-0168-9
In fact, when I think about it, that suggests an influence of eugenics on thoughts about eusociality… discuss!
Do I understand that the author is asserting group selection but ignoring many of the most-accepted arguments against it, including those from his own past work? That seems strange. Does he make the claim that kin selection theory has “crumbled” without a detailed proof of said crumbling? I don’t understand how that works. I don’t think the journalist Richard Dawkins would make assertions without addressing prior work or objections head-on. Maybe I’m just confused.
Great review, Jerry. And thanks for putting it here – I didn’t catch it in the TLS before.
Reading this had me recall Dawkins’ Prospect Magazine review of 2012 on the same theme: ‘The descent of Edward Wilson’.
If I remember correctly, Dawkins had a more fundamental problem with group selection: it’s not just that there’s no evidence for group selection, but that groups of organisms do not replicate or mutate. So it seems to me Dawkins rules out group selection because it requires one to forget that the gene is the unit of evolution.
Well, groups of organisms certainly can replicate, and they can also “mutate” in that members of a group can have mutations that differ from those in other groups. If that mutation rate is high (say, for altruism that is individually detrimental), and the replication rate of groups that are more altruistic than others is also very high, you could fix the gene. But then, of course, it will ultimately be replaced by the more “selfish” genes in all groups.
Thank you for the correction and clarification.
I’m sorry but I don’t follow.
Are you saying that groups actually have their own equivalent of genetics?
I always thought that one issue was how weak and fluid groups are, individuals can swap their groups (though not species) and obviously individuals can change themselves.
And since individuals are to some extent ultimately governed by genes, then groups too can or should be rooted in genes (partly, at least).
Am I mistaken?
We were having a discussion of Haidt’s latest book, and he appeals to group selection in humans and I wondered if it would work short term before being overwhelmed by the usual “free rider” problem. Does this make sense? (I realize there’s a lack of evidence to confront, but at least conceptually?)
For a critique of group selection that specifically targets Haidt’s ideas, I’d recommend that you or your group read Steve Pinker’s essay on Edge, “The false allure of group selection.“
Thanks, Jerry, this was helpful to me at least.
I’ve always assumed that the single “killer ap” for group selection would not be the gene for “drop everything and help your buddies [not necessarily cousins].”
Rather it would be “drop everything and help your buddies [not necessarily cousins] kill any random non-buddies”
No, Dawkins clearly states that in group selection as in individual selection what is selected FOR are genes. But selection works ON the carriers of these genes, and the question was if groups, like individuals, show
“something equivalent to surviving or dying, something equivalent to reproducing itself, and that it has something you could call a group phenotype, such that genes might influence its development, and hence their own survival”
What if a behavior or trait is bad for the individuals of a group, yet it is even worse for those individuals who are on competing groups?
You mean like religion and nationalism?
It seems to me that if these are true ‘groups’ (without gene flow between them), then the frequency of the gene should decline in both groups, and it should decline more quickly in the group where it is more harmful. Of course that might indirectly benefit the less-harmed group. Is that what you are getting at?
Excellent review. I like how you distinguish cooperation from altruism – a point often missed. What I just can’t figure out is why Wilson is sticking so fast to his guns. Just last week I was reading this book looking for his explanations. But nowhere in the book or in any do I find him making any real argument. All we get is a wave about that silly discredited gene stuff. If anyone know of a paper where he makes a real argument I would love to read it.
Another question I have is how we would observe group selection. Kin selection is easy to see and make sense from both a genetic and predictive view. But I’m not sure what kind of experiment we would set up to show a situation where ONLY group dynamics contribute to evolution.
I appreciated the altruism vs cooperation bit as well. The interesting thing about eusococial species is that altruism is the mode of existence for the workers, whereas cooperation is the mode for humans, with true altruism being rare. If a soldier or parent makes the ultimate sacrifice for their group, really what they are doing is taking the risk of dying. Perhaps the Kamikaze or suicide bomber is the exception – groups using suicidal actors been less successful overall, and I think Pinker better explains the motivations.
I think you can’t ignore self-awareness in human altruism, per Jerry’s point. The soldier who throws his or her self on a grenade to save comrades is usually making the calculation that s/he is going to die either way. When humans are charitable, it is largely surplus that they share, and in any event we are capable of calculating (if imperfectly) the size of sacrifice we are willing to make. And again, a la Jerry, the calculation includes multiple selfish considerations.
So, yeah, I don’t understand either how for non-eusocial groups how you could test and differentiate group selection and distinguish it from enlightened cooperation. The onus really is on Dr. Wilson to provide that data since there are decades of accumulated data which support other explanations.
Altruism in eusocial castes still eludes me. If you cannot reproduce how is it not selfish to do anything possible to protect and care for your one fertile family member?
And, if a worker bee doesn’t do her job, she dies. There’s got to be some selfishness in that, too, no?
I don’t understand your question. Is it just semantic? Altruistic behaviors are those that appear to directly benefit others at direct expense to one’s own fitness. The fact that such behaviors might indirectly rebound and indirectly benefit the altruistic actor are what is being explained by kin selection. The difference seems to be between direct and indirect benefits and expenses.
The behavior of sterile clones in caring for their reproductively active mother is anything but altruistic. Its their gene’s only hope.
In organisms with haplodiploid sex determination, female workers are better off not reproducing for themselves because they have 3/4 relatedness to their sisters, but only 1/2 relatedness to their female offspring. If anything, they are taking advantage of the queens, who are only 1/2 related to their female offspring.
Right. And taking advantage of the queen.
Death of an individual ant or bee is no more significant, genetically, than death of a skin cell. I tend to think of eusocial individuals as organs or cells.
There have been human political movements that thought of human individuals pretty much this way. Most of them promote this when group survival is threatened.
Every empire, and not a small number of extant nation states, use this trope. Or rather, their 1-percenters do.
> This is a bit longer than my usual posts
…and I read the entire thing 🙂
It was really accessible to a non-biologist like myself. The choice of examples was great which should help me remember kin vs group vs individual (e.g. ‘sacrifice myself for my three brothers’). I bet it takes a long time to find exactly the right ones. Thank you!
Isn’t group selection just kin selection extrapolated to distant relatives? we humans are all family, after all.
No they are fundamentally different. Kin selection depends on the altruistic action benefiting genes that the altruist shares with the beneficiary of the action and also predicts that the extent of the altruism can be predicted from the degree of relatedness (i.e. you are likely to put yourself out more for offspring or siblings than for third cousins twice removed).
Group selection does not require any individual benefit to accrue at all as long as the group benefits.
A great review, with a powerful and easily understood attack on the idea of group selection.
Wilson is not the first great scientist to go off the rails in his old age, Linus Pauling was apparently the same with his ideas about Vitamin C.
Uh-oh. That’s a bad omen: Vitamin C is still thought of as a cure-all by a lot of people.
Vitamin C seems to be a Jekyll and Hyde molecule – it is an anti-oxidant, but it also turns Iron into an oxidant, so you don’t want too much in your body. That is probably why excess Vitamin C in the diet is not absorbed very much. (See “Oxygen” by Nick Lane.)
It is very hard to understand why Wilson has gone down this route. As I understand it the problem lies in the fact that individual level selection inevitably proceeds at a much faster rate than group selection can. Even if the group benefits from altruism, selfish individuals prospering in the shorter term will continually undermine this group benefit. The only realistic way for altruism to evolve in a stable manner is if there is a selfish benefit to the individual (or gene) as the kin selection and reciprocal altruism models allow.
It will be a shame if this spoils Wilson’s scientific legacy which has been fantastic prior to this.
I learned a lot just from your review. I love the way you explain things, making your subject accessible to non-scientists like me.
I like this article for its explanations & I forwarded this as a good primer for various forms of selection.
Sub
I’m very confused how group selection might be operating in what little I have learned about group dynamics in primates. Groups can merge when a stronger and more cunning group takes over, and the males are taking the chance of up to 50% “loser” genes in their offspring with the animals they absorb into the new group. New groups are formed when second-place alphas break away for selfish reasons. I guess weaker groups can starve or otherwise die-out when they are outcompeted, but I don’t see how better fitness, kin selection and simple cooperation aren’t sufficient to explain the dynamic, or if not why there hasn’t been significant contrary effects to necessitate a theory which doesn’t just adjust but literally overturns them.
And just to add, groups splitting and merging do not result in the extirpation of the respective former or absorbed group’s genes. I guess a group that dies out takes its loser genes with it, but, again, I don’t see what’s missing from the conventional view of fitness and selfishness to explain the success of a successful group. If selfishness enhances fitness within the group, what benefit would there be in group selection except as a strategy for the survival of less-fit individuals? And how does that benefit the group? I’m not hearing an argument that validates such a counterintuitive leap.
A blurb about possible group selection in social spiders is here, and can be viewed by clicking on the readcube link. I am not sure if everyone can see it, though. I could copy and paste it here, but i am not sure if that is kosher.
Another very readable article that everyone can see is here.
The idea of social spiders terrifies me. Their scary, sociopathic & now they know how to communicate & cooperate.
In the articles, there are two varieties of spiders in their societies. The ‘warriors’ which go after prey, and the ‘nannies’ that like to snuggle together, and they take care of the young-‘uns. I think the nanny caste sounds kind of cute.
Jerry, good article. Group selection in vertebrates seems pretty unlikely. However, I was under the impression that there is still a good case for group selection in the evolution of reduced virulence of the myxoma virus. Or has that been overturned? I agree that group selection may (I say may) operate in social spiders, but I would like to see the data on spider self-sacrifice, breeding structure that excludes kin selection, low reproductive success of selfish colonies, and guarantees that gene-flow is low between groups. Guess I am condemned to read the article.
I think that group selection could have been a factor in early human evolution, however relying on assumptions that are impossible to prove.
Humans lived in very small bands of a few dozen individuals swapping individuals occasionally to avoid inbreeding as chimps do, particularly when the total human population was less than 10,000.
Evolution occurs faster in small populations than large populations. If there’s any genetic element to cooperation and altruism, then the isolated groups with these traits may do better when conditions are bad than others, and the genes for altruism and cooperation (if they exist) will increase, as less cooperative/altruistic groups die out.
Certainly won’t work in larger populations (as we have now), instead both relying on social factors.
If there’s ever a disaster and humans drop back into early numbers of humans in small isolated groups, then group selection, if it exists, could again become important.
If it exists that is. And I don’t know how it could be distinguished from kin selection.
Group selection relies on there being a genetic determinant, at least partially, for the tendency to act cooperatively and altruistically. And that your survival in a small group relies on survival of your small isolated group is dependent on actions of non-related individuals of your group, again at least partially genetic.
Non-adaptive evolution is fast in small populations because of drift. However, small populations my evolve little because they store less genetic variation (rare alleles will be absent) and, because they are small, the chance of new beneficial mutations will be reduced. Geneticists have shown that you can produce much more change by artificially selecting large population of fruit flies than a small one.
You are right that in a world having many small populations, there would be a better chance for at least one arising with a large fraction of genetically altruistic individuals that could work together to produce new groups comprised of altruistic descendents. However, selfish individuals arising within or joining groups of altruists will be especially favored by being able to parasitize their naively generous neighbors. In the presence of immigration, unless there is policing or social tagging of cheats, selfish behavior will expand and undermine the evolution of altruism. If ‘altruism’ advances because of policing or social control, the mechanism no longer corresponds to the definition of group selection.
In short, as Jerry says, the conditions necessary for the operation of group selection are so restrictive that it is unlikely to be of general importance, although it may on occasion work when conditions are just right. Human evolution does not seem to be one of these Goldilocks occasions.
W. Benson,
I’m not certain that that’s completely correct. Small populations also tend to evolve faster because they tend to occupy small areas exposed to a more uniform selection pressure, as compared to large populations over a wide area exposed to less uniform selection pressures.
Favourable combinations of genes in small populations, if they’re present, allow that population to survive. If they’re not present in another separate small population, then that population tends to die out.
It’s just allopatric speciation and I think the true mechanism of Punctuated Equilibrium which I don’t think has anything to do with rapid genetic change.
Natural selection acts at the level of the gene. And at the level of the population. And hardly at all at intervening levels.
Back when On Human Nature was first published, most ‘social scientists’ were screaming that behavior isn’t genetically based. They aren’t scientists, they don’t understand research, and the papers they write are essays. Humans do both things, the compete and cooperate. Altruism is not separate from cooperation. Each individual has his or her own combination of competition and cooperation. The relationship is very complex and difficult to experiment with. Scientists and people who just write about it are two different groups.
A very good read, thank you.
If incontrovertible evidence of group selection is ever actually demonstrated, will that mean that Robert Ardrey doesn’t have to be an unperson anymore?
Whilst he misled me with group selection, his Social Contract was inspiring and led me on to read Rousseau. It’s been a wild ride since!
What a grand review!
Beautifully lucid and persuasive and excellent comments too.
I wish I’d been able to cite this review when I was trying to explain selection in the comments on this post at The Conversation. That site’s not very link-friendly though.
Jerry and others, What do you think of Christopher Boehm’s book?
“MORAL ORIGINS: The Evolution of Virtue, Altruism and Shame”
For GROKS podcast from Boehm try https://archive.org/details/groks531
Or CARTA video try
http://www.uctv.tv/shows/CARTA-The-Evolution-of-Human-Altruism-Christopher-Boehm-Social-Selection-Versus-the-Notorious-Free-Rider-21281
Thank you, Jerry, this was well-written and very clear.
“The false allure of group selection” and its discussion (in which Jerry takes part) is absolutely brilliant. I’m still digesting, it is not really ‘easy stuff’, since so many ideas are proposed. Again, brilliant, thanks for the link.
I totally agree -as a first conclusion- that the term ‘group selection’ should be shunned, since it means so many different things, but nothing paralleling natural selection, hence confusing the issues.
Sad to see one of my heroes, Ed Wilson, descend to the primitive, Wynne-Edwards and Lorenz type of ‘group selection’ level.
But then, did not the great Wallace end his days in spiritism?
I recently watched a program about an apparently rare instance of a lion killing two cheetahs. It wasn’t fun to watch, however interesting. After reading this post I can’t help but think group selection is certainly not at work within groups of male cheetahs.
To wit: three males, probably brothers, were stalking a single female. They were so aggressively attacking one another in an attempt to establish dominance that by the time they cornered her in small wooded area and began to violently pursue her, they were caught completely unawares by the large male lion that had been tracking them (the purpose for this bizarre hunt was the focus of the program). Unfortunately, the lion rushed into the thicket and killed the female instantly. The three males bolted with one injured in the fracas. Moments later the injured male was killed by a second lion that had showed up on the scene. The two remaining cheetahs took off.
What I found so interesting – more so after reading this post – is that male cheetahs spend a great deal of their time in small groups. As the program stated, this is an effective strategy when you consider that they have evolved into such specialists that their defenses have been diminished in the bargain. So when it comes to their survival, grouping together is an effective strategy. However, this simply sustains their genes, it doesn’t make little cheetahs. When it comes to reproduction, that all-for-one and one-for-all nonsense clearly goes right out the window. This was extremely well-filmed and the violent behavior exhibited by the three males toward their own family members was remarkable. To the untrained eye, this appears to be a notch in the selfish gene belt.
Note that the female was killed too
Great review! So enjoyable to read.