Laws of Cooperation

Teaming up with science writer Highfield to address lay readers, Nowak argues for the fundamental importance of cooperation throughout life's history. Teaming up with science writer Highfield to address lay readers, Nowak argues for the fundamental importance of cooperation throughout life's history.

T he greatest challenge to science is to explain the prevalence of improbable, stable structure in the universe. Entropy mercilessly ensures that through time disorder is bound to increase. Discovering general mechanisms that favor local order is in some sense the most important game that all theorists in the natural sciences have been playing since the mid-19th century. In biology, theorists have focused on the battle between the dissipative tendencies of selfi shness and the structuring capabilities of cooperation. Theoretical biologists-like theoretical physicists working in the abiotic domain before them-have sought natural laws that explain when and where cooperative order, and consequently adaptive complexity of various kinds, might come into existence.
In SuperCooperators, mathematical biologist Martin Nowak and New Scientist editor Roger Highfi eld present for nonspecialists recent progress in understanding how cooperation evolves from the competitive interactions inherent in Darwinian dynamics. They argue that we now know enough about the mathematics of adaptation to conceive of cooperation as a fundamental evolutionary principle and not merely as a surprising feature of biology in some way at odds with evolutionary processes.
We currently think of mutation, competition, and selection as fundamental properties of all evolutionary processes at different levels of explanation. Nowak and Highfield suggest that cooperation should be elevated to a similar status. They devote much of the book to building their case, presenting fi ve candidate mechanisms (repetition, reputation, space, multilevel selection, and kin selection) that, when key inequalities are satisfi ed, allow cooperative confi gurations to outnumber selfi sh ones.
Many readers will turn to the book for further insights into the ongoing argument that pits Nowak and his colleagues against pro-ponents of Hamiltonian kin selection. I have been rather perplexed about the source of so much uproar (beyond the obvious sociology of science and the desire that precedence be acknowledged), and the book should go some way toward clarifying the nature of the debate for outsiders (albeit with a participant's perspective). Two chapters, "Kin selectionnepotism" and "The lord of the ants," most explicitly address the dispute.
Nowak's essential (and rather simple) claim is that all forms of cooperation can be understood in terms of individual-level selection operating in hierarchically structured populations. The source of structure can be kinship, spatial constraints on mobility, or mechanisms of individual recognition and attraction. Historically, kin selection has suffered from a few ambiguities. As originally conceived by Bill Hamilton, kinship implied genetic relatedness. It was soon recognized that this was too restrictive, and the approach was expanded to include multiple forms of assortativity. Hamilton also introduced the concept of inclusive fi tness, which is where the substantive argument arises. Inclusive fi tness calculations are often articulated in terms of coeffi cients of relatedness (not assortativity in general), which moreover only make sense under rather limited conditions of weak selection. Nowak and colleagues, along with several population geneticists before them, suggest that we model individual selection and hierarchical population structure explicitly and do away with the phenomenological ambiguities of inclusive fi tness arithmetic. I suspect that this is what will eventually happen, because theoretical science ultimately favors mathematical clarity and effi ciency. The book includes ideas that might reconcile the ground-breaking work of Hamilton and subsequent developments with the structured population dynamics offered by Nowak, Corina Tarnita, and Edward Wilson ( 1). Since the 1960s, when Hamilton wrote down his famous rule for the fi xation of cooperates in terms of the ratio of benefi ts to costs and the coeffi cient of relatedness (rb > c), there has been a proliferation of analogous inequalities. Nowak describes several, including a graph theoretic version, b/c > k (where k is the average number of nodes to which a focal node is connected in a network), and a group selection form, b/c > 1 + n/g (where g is the number of groups and n the number of individuals per group). All of these recent inequalities explain the fi xation of cooperation in terms of two classes of separable mechanisms: reproductive or growth terms (functions of time) and population structure terms (functions of space). Hamilton was the fi rst to intuit the form of these evolutionary laws, and Nowak and colleagues have been pinning down their fundamental origins and meaning.
In the penultimate chapter, Nowak suggests unifi cation in the form of one parameter, his sigma-the structure coeffi cient. Sigma measures the rate of like interactions in an evolutionary game (e.g., cooperators encountering cooperators or defectors encountering defectors). When sigma exceeds unity, cooperation invades; when below, selfishness. Sigma captures in one scalar value essential features of assortative interactions, and hence it points toward the essential processes underlying the origin of cooperation.

David Krakauer
The Classic and complex case. Cooperation in leaf-cutter ants (here Atta cephalotes, from Costa Rica) involves not only colony members but also the fungi they farm and bacteria they use to keep their crop disease-free.
www.sciencemag.org SCIENCE VOL 332 29 APRIL 2011 I n The Blame Game, Christopher Hood identifi es one of the most common gripes that citizens have about bureaucracy and government, namely, that no one in either accepts responsibility for making mistakes of omission or commission. In this brief and often illuminating book, Hood (a political scientist at Oxford University) explores the diverse and insidious ways in which ducking blame manifests in public life.
Hood's book follows in a spirit similar to other powerful approaches to politics by posing a single maximand (here, dodging blame) and inferring propositions that extend from this by-necessity unrealistic assumption. David Mayhew famously postulated that members of Congress must be single-minded seekers of reelection, for any other goal must be subordinate to seeking and retaining offi ce ( 1). They may seek power, or have policy goals, or have aspirations for higher offi ce, but all of these goals are for naught without election to offi ce in the fi rst place. With relevance to bureaucracy, Hood's blame dodging is akin to William Niskanen's budget-maximizing

John Brehm
The reviewer is at the Department of Political Science, University of Chicago, 5828 South University Avenue, Chicago, IL 60637, USA. E-mail: jjbrehm@uchicago.edu bureaucrat, James March's slack-maximizing organization member, or Herbert Simon's satisfi cer (2)(3)(4). The weaknesses of such singleminded approaches are palpable: they require grotesquely simplified characterizations of the behavior of the actors (legislators, bureaucrats, organization members) they study.
Of course, all of these actors have other goals, some of which may provide better explanations of particular aspects of the actors' behavior. Yet the strength of singleminded approaches is that by providing clear assumptions about the maximands, authors can deduce specifi c, falsifi able propositions about behavior. Their methods of deduction might be formal (as with Niskanen), discursive (as with Mayhew), or a hybrid (as with Simon and March), but from each readers would learn to expect something: that the acquisition of a greater budget smothers programmatic goals; that taking a position on controversial matters of politics matters less than claiming credit, which is in turn less frequent than simple advancement of name recognition; or that bureaucrats become less efficient in the long term by seeking resources that permit rapid response in the medium term.
What are the specifi c propositions that dodging blame implies? In his discussions, Hood identifi es four main classes of actors: top leaders, street-level bureaucrats, intermediaries, and "all the different individuals with whom governments and public service providers deal." (Hood has the habit of colloquializing many of the names of the actors: Top leaders are variously "generals" or "bananas," and those directly serving the public are the "poor bloody infantry" or "frontline troops," but none of these, often military, terms are really apt.) All these actors have a repertoire of ways to accomplish the same essential task: Leaders defl ect blame toward subordinates, other agencies, or entities entirely outside of government; the front line shoves blame toward their leaders, other agencies, or outside entities; middle management might pass responsibility up, down, sideways, or out; "civil society" (the least coherent of the groups of actors) pushes sideways or outward. In short, all directions of blame shifting are possible-hardly the precision desired from a one-motivation model. Some of these directions are rather dysfunctional. Top leaders who blame subordinates may fi nd themselves bereft of essential support in subsequent crises, and subordinates who blame their chiefs may fi nd themselves called on the carpet before too long.
Actors may avoid blame by ignoring present problems until public attention shifts elsewhere, persuading the public that the problem is not as bad as it seems (or could be), diverting attention to another problem, or even apologizing for the problem in the fi rst place. One wonders what is left over, especially if apologies count as blame ducking-perhaps the degree to which the actor dodges blame has to do with the very subjective notion of how sincere the apology is? Hood approvingly mentions President Obama's apology and acceptance of blame after the failed attempt to appoint former senator Tom Daschle to the post of Secretary of Health and Human Services ("I screwed up"). Yet it is far from obvious that Obama had any particular role to play in the failed appointment other than being in charge of an administration making its fi rst appointment mistake. The book contains multiple fi gures that resemble extensive form games and graphs displaying situations that suggest indifference curves, but Hood uses these graphical devices to illustrate rather than to press an argument. Nor does he provide much in the way of direct quantitative evidence for his principal claims despite occasionally wishing for more useful indices. Instead, he supports his argument with qualitative evidence culled from stories of government or bureaucratic malfeasance. The anecdotes make for great story telling, but are they convincing data?
By the end of this short book, the reader may reach the rather unsatisfying conclusion that what we experience sporadically actually occurs regularly, for reasons that apply only occasionally. Bureaucracies may or may not be functional places where routine processing happens in a fair and appropriate way, but The Blame Game provides little guidance as to why politicians and government offi cials accept blame or how a polity might more effectively encourage appropriate responsibility.

The Blame Game
Spin, Bureaucracy, and Self-Preservation in Government I have little doubt that SuperCooperators will generate great interest and debate and meet with its fair share of dissension. Nowak, a very successful and plainspoken scientist, has a disinclination to compromise. His style has always been to forge ahead with alacrity, and at times this has led to an absence of historical considerations that offends those working on related problems. But he and Highfi eld should be commended for communicating in an accessible language many fascinating insights into the nature of cooperation. Their readers will fi nd an informative survey of the rich mathematical models and theories that defi ne a lawlike realm of biological and cultural phenomena.