Genetic and Cultural Evolution of Cooperation. Peter Hammerstein, ed. MIT Press, Cambridge, MA, 2003. 450 pp., illus. $45.00 (ISBN 0262083264 cloth).
The question of how cooperation evolves is fundamentally a question of mathematical biology or sociology. Recent theory about cooperation has been overwhelmingly mathematical, and virtually all of the 40 contributors to the 90th Dahlem Workshop report, Genetic and Cultural Evolution of Cooperation, are in some sense mathematicians. Yet arguably the most striking feature of this four-part book is a virtual absence of mathematics. What could be the reason for that? Could it simply reflect that publishers like to sell books, and equations reduce sales dramatically? Could it just mean that brainstorming at a workshop is more conducive to talking than to doing? Or is it perhaps a sober reflection of how little we truly understand about the evolution of cooperation, despite decades of concerted effort?
In a timely article, Michael Reed (2004) has reminded us that evolution makes mathematical biology hard for many reasons, including that a priori reasoning is frequently misleading, and that different species may accomplish the same task by different mechanisms, astounding in their variety. For example, that reciprocity can sustain cooperation has been demonstrated ad nauseam in theory; but there is surprisingly little evidence of such “reciprocal altruism” in nature, as editor Peter Hammerstein—lamenting the amount of energy invested “in the publishing of toy models with limited applicability”—protests in his personal contribution to part 1 of the volume. Indeed, as Joan Silk stresses, keeping score—and how can there be reciprocity without scorekeeping?—is a proven detriment to friendships.
To be sure, reciprocity is not the only mechanism for cooperation that theoreticians have considered. There is first of all relatedness, but this book addresses only cooperation among unrelated individuals. Can such cooperation be sustained in humans, if not by reciprocal altruism, then by “costly signaling,” “indirect reciprocity,”’ or “genetic group selection”? The usual suspects are trotted out by Ernst Fehr and Joseph Henrich, but soon dismissed as unlikely perpetrators of “strong reciprocity”—defined to arise when one is willing to incur long-term net costs from helping another in response to a kindness. Of the known suspects, only “cultural group selection” survives pretrial scrutiny, to be revisited in part 4 by Peter Richerson, Robert Boyd, and Henrich.
The evidence for such strong reciprocity, now seemingly overwhelming, is effectively evidence that animals— especially humans—are much more cooperative than game theory has predicted. So what have the models been neglecting? One good answer is psychological mechanisms. Daniel Fessler and Kevin Haley argue that “our emotions, long disparaged as both a reflection of our animality and the source of our irrationality, are… exactly the opposite, namely, the keys to our complexity, efficacy, and remarkable ability to cooperate,” and they discuss the “thirteen emotions that seem to have the greatest impact”: anger, contempt, envy, guilt, gratitude, righteousness, romantic love, pride, shame, moral approbation or outrage, admiration, elevation, and mirth. Their view is echoed by Edward Hagen, who argues that depression, far from being a mental illness, may be an adaptive emotional strategy. These and other themes (e.g., reputation) coalesce in the final chapter of part 1, a group report by Richard McElreath and 10 others on cognitive and emotional mechanisms that may sustain cooperation.
The flavor of part 2, on markets and exploitation in mutualism and symbiosis, is markedly different from that of part 1. The emphasis switches from humans to other organisms, and from intraspecific to interspecific cooperation. Samuel Bowles and Hammerstein discuss market theory in relation to biology; Redouan Bshary, Ronald Noë, and Judith Bronstein highlight the importance of model systems; Olof Leimar and Richard Connor discuss “pseudoreciprocity”; and Carl Bergstrom and Michael Lachmann argue that slowly evolving species are likely to gain a disproportionate fraction of the surplus generated through mutualism. A group report by Bergstrom and 10 others concludes part 2.
In part 3, the focus switches all the way from the interspecific to the intercellular. Rolf Hoekstra discusses mechanisms that prevent or promote genomic parasitism; Eörs Szathmáry and Lewis Wolpert discuss the transition from single cells to multicellular organisms; conflict mediation during this transition is Richard Michod's topic; Neil Blackstone and Thomas Kirkwood discuss the capacity of programmed death to restrain the selfish replicatory potential of individual cells in multicellular groups; and Lachmann and seven others intertwine these strands in a group report on cooperation and conflict in the evolution of genomes, cells, and multicellular organisms. Part 3 is arguably somewhat detached from the rest of the book. However, even if its mechanisms operate only at the level of the cell, it may yet offer lessons at the level of whole organisms. Certainly, there is nothing in the group report to suggest that the level of cooperation within organisms is any more perfect than the minimum level of cooperation among humans that our planet's future well-being now clearly and urgently requires.
Part 4 of the book, on cooperation in human societies, picks up a thread left dangling at the end of part 1. Richerson, Boyd, and Henrich propose that group selection on cultural variation is at the heart of human cooperation (though they acknowledge a role for other mechanisms); Peyton Young describes how social norms can coalesce from the decentralized, uncoordinated choices of many interacting individuals; Eric Smith emphasizes the importance of language's role in human cooperation; Bowles and Herbert Gintis discuss a special mechanism for human cooperation that stresses the role of gene–culture coevolution in group dynamics; and Henrich and eight others conclude with a group report.
Although this book identifies numerous gaps in both empirical and theoretical knowledge—indeed, that is precisely its strength—it is arguable that the most critical lacunae concern the modeling of psychological factors. In this regard, the book is the tip of an iceberg, because its emphasis on the role of emotions in cooperation reflects a rapidly growing consensus that “psychological traits must be incorporated into our model, however difficult that may be” (Nesse 2001, p. 162): Emotions are capable of sustaining commitments that may otherwise seem irrational, and commitments are often crucial to cooperation, as discussed more fully elsewhere (see, for example, Frank 2001). Furthermore, the need to incorporate psychological factors transcends cooperation and commitment. For example, that the most egalitarian developed nations—as opposed to the richest—enjoy the best health has now been firmly established by Wilkinson (1996), who wrote that “although economics is far from exclusively asocial… rational choice theory has grossly underestimated human social needs and [that] their satisfaction should often take precendence [over] demands to maximise individual consumption. There is a missing social economy of well-being” (Wilkinson 1996, p. 109). Some have attempted to address this issue within the confines of top-down, neoclassical economic analysis by adding a social-capital term to the standard utility function (see, e.g., Becker and Murphy 2000); but if the linkages between health and social status, or between cooperation and commitment, are primarily psychosocial— as Wilkinson or Fessler and Haley (and others) have so convincingly argued—then the standard economic approach will soon yield diminishing returns. A much more radical, bottom-up approach is needed.
Hammerstein and his collaborators are among the pioneers of the requisite “behavioral game theory” (Camerer 2003). Aware not only of the many discrepancies between game theory's predictions and the kind of cooperation observed in nature but also of the tremendous variety of mechanisms that could in principle sustain cooperation in different contexts, they are effectively going back to the drawing board. And I for one would like to think that that is why this monograph contains so little mathematics: The precise structure of models that failed yesterday need no longer be relevant to models that may succeed tomorrow.
The book is a refreshingly candid portrait of what we know and—more important—don't yet know about the evolution of cooperation. Honest disagreements abound (e.g., between Smith and Boyd and Richerson over the importance of group selection in the evolution of human cooperation), yet there is at least a universal consensus that, as Silk puts it, “As always, more data and better models are needed.” The group reports are splendid state-of-the-art summaries emphasizing open questions, unsolved problems, and directions for future research; and all agree that the challenges are considerable.
Which brings us full circle to Reed, who has warned against doing mathematical biology to satisfy a desire to find universal structural relationships, because “you'll be disappointed” (Reed 2004, p. 339). Is it therefore a strategic error even to seek a general theory of cooperation? That remains an open question, but if you want insightful perspectives on most of the relevant issues, I heartily recommend that you read this book. And even if now is not the time for a general theory, we mustn't forget that now is not forever. After all, when all is said and done, who will disagree with W. G. Runciman's categorical assertion that “our social behaviour is as reliably patterned as our individual behaviour is unmanageably diverse” (Runciman 2000, p. 88)?