A Cooperative Species: Human Reciprocity and Its Evolution [Hardcover]

I was really looking forward to reading this book. One of the questions that keep haunting me is: how is it possible that the same species that David Livingston Smith called (correctly) "The Most Dangerous Animal" (dangerous for other members of its own species, that is !) is at the same time one of the most cooperative species of the world, surpassed only by eusocial insects and maybe naked moles? This book, I hoped, would give me some hint to solve this conundrum. And it did.

But first a warning : When the book finally arrived, I leafed through it - and was tempted to send it back immediately. Mathematical formulas and equations, lots of, crawling like little black spiders on every second page! Math makes me sick. I haven't got any mathematical education beyond the rule of three (and I'm not proud of it, believe me), so I tackled the book with more trepidation than hope. Unfortunately, the style also lived up to my worst fears: hardcore scientific prose you normally expect in journals like "Behavioral Ecology and Sociobiology" or "Journal of Economic Theory". I never read these publications, a trait I share with the majority of Amazon customers, I guess.
It's not a book for somebody with a diploma in, say, philosophy or literature, who just happens to be interested in the question "Why are humans such a cooperative species?". It's a book written by two experts for their fellow experts, and unless readers are well versed in economic or game theory they will have to content themselves with reading for gist.

So I just kept skipping the parts with the math and tried to make sense of the rest. And now for the good news: The rest does make sense. It gave me some hints to look for an answer to the maddening ambivalence of human nature I mentioned above. The hint, in a nutshell, is something like: Take confrontation and cooperation as two sides of the same coin. The term Bowles/Gintis coined for this ambivalence is "parochial altruism".

There are in principle two ways of explaining human altruism.
(A) Altruism is only skin-deep. Selfishness always lurks behind nice appearances. For instance Trivers' "reciprocal altruism", a misnomer, because no genuine altruism is involved. People, according to this approach, expect (sub-consciously) to be repaid sooner or later. If somebody helps a total stranger, with no prospect of being repaid, this theory explains it as a kind of "Big Mistake". Because humans spent most of their history living in small groups based on kinship the human mind even today, in an anonymous situation, acts as if it still were in the pleistocene, helping not some stranger but a relative or some guy he or she will probably meet again sooner or later. Scratch my back, and I'll scratch yours. See you. Genuine altruism today therefore would be a kind of misfiring in a situation unknown to the old mental moduls that still govern our behavior. Bowles/Gintis show that this "Big-Mistake-Hypothesis" cannot be true because even in the Pleistocene humans had plenty of contact with strangers. The picture of small and closed bands (based on kinship) of stone-age people is a myth.

(B) Bowles/Gintis argue that altruism can be genuine. But if altruism is defined as a behavior that increases the fitness of the recipient and that is costly to those practicing altruism - how can it spread ? Because evolution is a game that dispenses with rules other than natural laws, being nice has to pay off, or it will disappear. What, according to Bowles/Gintis, is the mecanism that assures the success of altruism ? Their answer is: group-selection. The whole argument of the book, as far as I can see, hinges on the question whether group-selection, or multi-level-selection as it is often called today, works or doesn'nt. That's the point where the mathematical models come in and where I drop out.

The prerequisite for group-selection to be effective is that selection within groups is reduced compared to selection between groups. The picture that emerges from the model is that groups who managed to reduce internal strife and competition by inventing and promoting cultural "leveling mechanisms" will act as units of selection and will outcompete groups with more selfish members. The competition, suppressed or reduced within groups, will increase between groups. Humans developed a special kind of groupishness: Being unconditionally altruistic towards their own people, being even eager to punish freeriders of the own group, even if the punishment is costly for themselves, while at the same time acting xenophobically towards other groups. The parochial quality of unconditional (not reciprocal !) cooperation is the necessary condition of genuine altruism to prevail.

Unfortunately, I'm in no position to decide whether the scientific arguments forwarded by Bowles and Gintis, based on mathematical models, hold true or not. On this question, I declare myself incompetent. But I'd say that everybody who is interested in human evolution should read this book, even without competence in higher math, because the idea is fascinating that in human history it was groups rather than individuals who were selected for or against. And that it was and still is culture that formed groups in a way as to making them act as units of selection. I think that culture and its group-forming force is the part of the picture that Dawkins and the gene-centered view of evolution missed ("memetics" doesn't explain anything).

Just one final remark: Herbert Gintis is one of the top reviewers of Amazon, and it's always a pleasure to read his commentaries. They are written in a clear, accessible style. If only this book was written a bit more like his reviews. My wish would be : keep the jargon to the journals. Books like these are for interested layfolks. The math may be essential for the argument; the jargon certainly isn't. Therefore only four points.

The book is pretty complicated but then also tells the whole argument for why Dawkins et al were wrong about the evolution of the human altruism (human ultra-sociality) through kin selection and reciprocity - and how the evolution of ultra-sociality really happened through multi-level selection and gene-culture co-evolution.

These people really know what is happening both in economics, antropology and evolutionary biology today, they strongly contribute to the development of both and they can convincingly make their point.

The discussion about the evolution of cooperation has certainly not yet come to a final conclusion. But on this stage it looks Bowles and Gintis are right when claiming:

"The view that early humans lived in worlds with little contact outside one's family--Dawkins' ideal conditions for self-interested cooperation to flourish--is difficult to square with what is known about the Late Pleistocene and early Holocene. Like Jean-Jacques Rousseau's philosophers, Dawkins, Huxley, and other biologists seem to have jumped on a faulty time machine, and have journeyed to an imaginary ancestral world...

We will see that neither the likely size of groups, nor the degree of genetic relatedness within groups, nor the typical demography of foraging bands is favorable to the view that Late Pleistocene human cooperation can be adequately explained by kin-based altruism or reciprocal altruism...

But our joint work with Stefany Moreno Gamez and Jon Wilkins of the Santa Fe Institute indicates that if ancestral groups indeed had been small and closed, the degree of differentiation among groups predicted by the standard model of equilibrium differentiation would be substantially greater than what we observe (Moreno Gamez et al. 2011). The genetic data thus are more consistent with ancestral groups being of considerable size and with ever-changing composition."

One issue came to my mind. Normally the efficiency of the market economy (with several companies offering alternative products) is said to be due to competition allocating productive resources to their most highly-valued uses. But maybe the efficiency is also due to competition causing the employees of the competing companies to act like a hunter-gather groups. The competition causes higher efficiency via the mechanism of heightened free-rider punishment by gossip etc.:

"A final experiment provides a possible link between group conflict and the evolution of cooperation based not on the fact that altruists, if parochial, are willing fighters but that group conflict stimulates altruistic punishment of free-riding fellow group members. Sääksvuori, Puurtinen and their coauthors (2011) implemented a series of eight-person public goods games with and without a punishment option and in which the payoffs of members of the groups playing these games either depended on the outcome of group competition or were independent of the performance of any other group.

...

In the treatments with group competition, the groups with the larger contribution to the public good won a prize that was twice the group difference in the level of public contributions. Group competition greatly heightened the punishment of shirking group members where this was possible, so that groups with the punishment option prevailed over groups without it."