“What Looks Like Cooperation is Really Selfishness”


Three years late, Scientific American is reacting to a paper on evolution of cooperation  [2012] with a hint that something is rotten in the state of Denmark.  This is what my wife used to call an FGIO, a faint glimpse into the obvious.  The paper was by interlopers from physics, mathematically adept senior luminaries.  The message of their mathematics was that if you put together a group of selfish individuals what evolves is (I’m sure you’ll be shocked to hear this) a selfish community.  A non-communal community.  The authors use the colorful word “extortion” to describe the dysfunctional social relations.

The reason this is important is that the standard evolutionary theory of cooperation says that global cooperation can emerge from behaviors that are individually selfish.  It can’t.

(We may be tempted to dismiss the whole topic as esoteric or peripheral, but in fact it pokes a hole in a framework of evolutionary thinking that has become sclerotic, and is overdue for rethinking from the ground up.  At stake is the way that we think about evolution, and by extension, the way we understand purposes and mechanisms in all of biology.  Aging is one particular case in point.)


Diverse forms of Altruism in Nature

Human soldiers offer up their lives all too readily in wars of questionable legitimacy.  Amoebas stop competing among themselves and form a “fruiting body” that behaves like a single organism.  Worker bees spend their entire lives serving a colony whose only reproduction is by a queen bee that may not even be a blood relation to the worker.  And yeast cells will commit suicide to digest themselves as food for their cousins when they detect that the local region has run out of food.  Fish of two different species work together to hunt more effectively.

When we look at the biosphere, the ubiquity of grand alliances is striking.  Everywhere, we see extreme examples of individuals sacrificing their lives and their legacy for the good of the community.  Not infrequently, the sacrifice is complete, in that the individual leaves no offspring at all; and yet the benefit to the group seems modest and indirect. What theory could account for the mismatch between individual cost and communal benefit?

In stark contrast to this broad observation, we have the dominant version of evolutionary theory that purports to explain how all this came about.  This is the “selfish gene” model.  The great majority of papers published in evolution today are still framed within this model.  This mode of thinking has infected and distorted scientific understanding of aging, which is why I became interested in the problem originally, and why this subject belongs in an aging blog.

Of course, adherents of the dominant school feel responsible to provide some account of cooperation, but they look at the world with tunnel vision.  There are just two narrow windows through which the broad spectrum of cooperative phenomena are permitted to be viewed.  One is kin selection: a gene that causes altruistic behavior in its bearer may survive if the altruism is directed toward close relatives who are likely to have copies of the same gene.  Second is Tit for Tat reciprocal altruism.  Higher animals that recognize each other as individuals may learn who they can trust, and on the basis of a network of pairings, broad cooperation can spread through a group.  It is reciprocal altruism that has been called into question by the 2012 paper referenced above, by William Press and Freeman Dyson.


Hamilton’s Contributions

I find it curious that both these windows were opened by William D Hamilton, early in his career, before Hamilton became a convert to group selection.  He died untimely of malaria after a trip to Africa observing and collectins specimens in 2000.

Kin selection is understood in terms of Hamilton’s Rule, which says C < rB.  In order for an altruistic behavior to evolve, the cost to the altruist (C) must be less than the benefit to the recipient (B) multiplied by their mutual relatedness (r). Relatedness r is ½ between siblings or for parent and child, and it goes rapidly down from there.  “I wouldn’t jump into the river to save my brother,” quipped J.B.S. Haldane, “but I’d lay down my life for two brothers or eight cousins.”

Reciprocal altruism is understood in terms of the “Tit for Tat” strategy, made famous in apaper by Hamilton with Robert Axelrod in 1981.  Based on a simulated tournament implemented with the crude computer facilities available at the time, Hamilton and Axelrod concluded that the most successful selfish strategy could be a basis for global cooperation.  This was validation of the “invisible hand” that Adam Smith had described two centuries earlier.  Emergent order in a system where everyone was only looking after himself.  Emergent cooperation in a system where everyone is behaving in a manner wisely selfish.

“What looks like cooperation is really selfishness” is a theme that pervades the evolutionary literature, in guises both direct and quite subtle.  It was the basis of the sociobiology movement a generation ago, and it thrives today in fields as diverse as evolutionary ecology and the medical biology of aging.


Origin of the Current Crisis

This way of thinking took it on the chin in 2012 when Dyson and Press published their paper, demonstrating a form of selfishness that Axelrod and Hamilton had not considered.  Their diabolical strategy took evolutionary learning into account, and gradually trained opponents over time to be submissive, using a purely mathematical form of “extortion”.

This is but the latest scream in the chorus of voices telling us that the “selfish gene” version of evolutionary theory doesn’t work.  The strength of cooperative groups is that they win out in competition against other cooperative groups.  The idea that the power of a cooperative group is just a lucky side-effect of selfish behavior self-organizing never was credible, and finally it may be biting the dust.  The truth is that the “selfish gene” version of evolution has outlived its usefulness. It is long past due for an overhaul.  Evolution does not work “one gene at a time”.  Cooperative relationships have evolved not by accident, but because cooperative groups are very successful competitors, and they can spread and take over areas where individuals are unorganized and selfish.


What does Selfish Gene Theory Leave Out?

Now that we see the truth, we can go back and ask about the Selfish Gene theory, what is the reasoning that supports it, and what is the flaw in that reasoning?

There are two lines of reasoning relied on by Selfish Gene adherents:

  1. All mutations originally appear in just one individual.  If a mutation for altruism appeared in one individual surrounded by selfish behavior everywhere, that individual would be left in the dust. Everyone around her would accept her generosity, thank you very much, and then use the advantage she had given them to out-compete her and leave her behind.  Genes for altruism could never spread in a selfish community.
  2. Conversely, in a community where everyone was cooperating, if a gene for selfishness arose in one individual, she would be able to benefit from the help given by all the others around her without herself bearing any of the cost of helping others.  Her star would rise, her progeny would prosper, and her selfish gene would spread through the community, poisoning it from the inside out.

This is a compelling perspective, and not obviously wrong.  However, it does not describe the world that we see.  It predicts that cooperation should be rare in nature, when in fact we find cooperation wherever we look, and not just among close kin.

So we know that the argument for the Selfish Gene must have a catch somewhere…what can the problem be?

In my theoretical work, I have offered a general answer to this question that I derived from the work of Michael Gilpin in the 1970s.  It goes like this:

All animals depend on an ecosystem.  There is a common reservoir of food in the form of living plants or animals on the next trophic level down.  The community of animals dependent on the same food stock is tied together in their fate by a need to conserve that resource.  It is all too easy to eat everything in sight, to deplete the food species and to use all that extra food energy to reproduce like crazy.  But the collective consequence of this behavior is rapid disaster.  Once the food species is gone, it takes a long time to grow back.  The next generation will starve, and the ecosystem will take a long time to recover, if it ever does.

There is a powerful tendency for population dynamics to fluctuate wildly, leading to extinction.  The extinction is rapid, and can occur in a single generation.  Therefore, it constitutes a very potent force of natural selection, one that can easily counterbalance selection for pure selfishness, and defeat the Selfish Gene.

So when Selfish Gene adherents think about a gene for selfishness taking over a community, what they’re leaving out is that the community would likely be destroyed in the process–the community would die out before the selfish gene could become dominant.  This is true of the easiest and most powerful form of selfishness, which is overconsumption.  And computer simulations show that once this most powerful form of selfishness is tamed, it ties communities together in a way that makes it easier for other forms of cooperation to evolve.

This is the mother of all cooperation, the glue that binds communities and makes selfishness a dead end.  Selection for population homeostasis opens a door that permits all other forms of cooperation to evolve.


The Bottom Line

Let me conclude with this bold statement of my radical new thesis:  If it looks like cooperation, it probably is.


This article originally appeared on Josh’s blog Aging Matters here.

[Editor’s note, see also http://www.nature.com/news/physicists-suggest-selfishness-can-pay-1.11254]