Games, Groups, and the Global Good

297 Pages·2009·3.16 MB·English

by  Steven A. Frank (auth.), Simon A. Levin (eds.)

Springer Series in Game Theory
Simon A. Levin
Editor
Games, Groups,
and the Global Good
123
Editor
Professor Simon A. Levin
Department of Ecology and Evolutionary Biology
Princeton University
Eno Hall
Princeton, NJ 08544-1003
USA
Preface
No problem is more central to understanding biological organization than explain-
ing cooperation. Indeed, the puzzles posed by extreme forms of cooperation were
acknowledged by Darwin as challenges to his theories, and delayed his publication
of the Origin of Species for 20 years. Today, we have learned a great deal about
the evolution of cooperation, from quorum sensing in bacteria to coalitions among
humans. Nevertheless, deep questions remain. How is cooperation maintained in
large groups, where individuals help others they have never met, or whose identi-
ties are unknown to the helpers? Why will individuals apparently sacrifice their own
welfare to sustain community norms, through charitable behaviors or punishment of
norm offenders? How are institutions, from social norms to civil and religious laws,
maintained? How do moral systems arise, and how are they maintained? These ques-
tions are central to understanding how societies maintain robustness, and they also
are key to achieving a sustainable future for humanity.
Much of the formal theory of cooperation can be embedded within the theory of
games, the origins of which are usually traced to John von Neumann’s “Zur Theo-
rie der Gesellschaftsspiele,” (On the Theory of Parlor Games) published in 1928 in
Mathematische Annalen, 100, pp. 295–300. Actually, Emile Borel published sev-
eral papers that laid the foundations for game theory 7 years earlier, but it was
von Neumann who really began to develop a comprehensive theory, culminating
in his 1944 Princeton book with Oskar Morgenstern, Theory of Games and Eco-
nomic Behavior. Von Neumann died in 1958; and on the 50th anniversary of his
death, a symposium was organized at Princeton University with the sponsorship of
the John Templeton Foundation to revisit progress in the theory of cooperation, and
particularly to investigate what new advances in game theory might be stimulated
by considering the broader question of the establishment of moral systems and the
regulation of public goods. Most of the participants in that symposium then devel-
oped their papers into longer contributions for this book, and other authors were
invited to complete the story.
Societies cannot exist without cooperation, or without norms, customs, laws, and
other institutions that sustain cooperation. These provide collective benefits that
maintain the groups, and provide them advantages in conflict with other groups.
One of the great challenges facing humanity is in discovering whether those col-
lective benefits can be extended to the global level, without the tribal conflicts that
v
vi Preface
threaten our survival. First we must understand the dynamics of groups : : : how
they form, how they are maintained, and how they interact with their constituents
and with other groups. What are the origins of the normative practices that sustain
these groups, as well as the second-order systems of reward and punishment that
sustain the norms themselves? What are the developmental and evolutionary trajec-
tories that lead to the development of moral systems, from informal arrangements,
to societies, religions, and other formal institutions with formal rules and practices?
This book examines how such approaches can be extended to consider the
broader questions that cross scales of organization, from individuals to cooperatives
to societies. By expanding traditional approaches, can we explain how heuristics,
like concepts of fairness, arise, and how they become formalized into the ethical
principles embraced by a society? How do those ethical systems themselves co-
evolve with the societies and other institutions? What maintains the robustness of
social contracts? This book brings together essays by a diverse group of schol-
ars, providing a broad perspective on these questions, and hopefully suggesting
promising new directions for game theory.
I am pleased to acknowledge the contributions of my co-organizers – Mary Ann
Meyers, Martin Nowak and Steven Brams – in the organization of the conference,
and in the development of many of the ideas in this Preface. I am also grateful
to the authors of the chapters in this book, and to the other participants in the
symposium – Robert Axelrod, Freeman Dyson and Ehud Kalai – for their intel-
lectual contributions and discussions at the meeting. Terry Guthrie was tireless in
organizing the conference, and Sandi Milburn was brilliant in helping this book
come to fruition. I am especially grateful to Mary Ann Meyers and the Templeton
Foundation for their inspiration, collaboration and support in creating the confer-
ence, and in the production of this book. Without Mary Ann, this book could never
have appeared.
Princeton University Simon Levin
December 29, 2008
Foreword
An iconic event in popular accounts of the history of ideas about evolution and
the origin of species is the Huxley–Wilberforce debate in 1860. The Bishop of
Oxford, “Soapy Sam” Wilberforce, was a gifted, though markedly self-satisfied,
orator. Speaking first, he unwisely mocked Huxley by asking “Is it through grand-
mother or grandfather that you descend from a monkey?” This opened the door to
Huxley’s memorable riposte: “Would I rather have a miserable ape for a grandfather,
or a man highly endowed by nature and possessed of great means and influence,
and yet who employs these faculties and that influence for the mere purpose of
introducing ridicule into a grave scientific discussion – I unhesitatingly affirm my
preference for the ape”. After that, the meeting dissolved into Pythonesque chaos.
First, Fitzroy (who had captained the Beagle, gone on to become the Governor of
New Zealand, but by 1860 was mentally unstable, and soon afterwards committed
suicide) came raving down the aisle, brandishing a bible, and demanding that all
return to “The Book”. The meeting ended as Lady Brewster – in the words of one
of Darwin’s biographers, DeBeer – “employing an idiom now lost, expressed her
sense of intellectual crisis by fainting”.
What most accounts of these events overlook, however, is the fact that Wilber-
force, had he possessed an all-encompassing knowledge of the science of his day,
could have won the debate. The Darwin–Wallace theory of evolution, at that time,
had three huge problems.
The first problem concerned the time available for evolutionary processes to oper-
ate. Fifty years were to elapse before the first glimmers of awareness of weak and
strong nuclear forces were to appear. Of the four fundamental forces recognized by
today’s physics, only gravitational and electromagnetic (“chemical”) forces were
known in Darwin’s day. But if the sun’s energy source was gravitational, it could
not have been burning for more than about 20 million years. And chemical fuels
would give an even shorter life. A different calculation showed that it could not
have taken more than roughly 20–40 million years for the earth to cool from molten
rock to its present temperature. These two calculations meant that either the earth
was at most a few tens of millions of years old, or that Victorian physics was fun-
damentally deficient. Faced with these arguments, Darwin removed all numerical
references to geological time spans in the third and later editions of the Origin of
Species, and you will look in vain for any explicit chronology in the later Descent
vii
viii Foreword
of Man. Of course, the subsequent discovery of nuclear forces showed Victorian
physics was indeed inadequate: the sun burns nuclear fuel; and the heat generated
by the decay of radioactive elements inside the earth invalidates simplistic calcula-
tions about cooling rates. We now understand that evolutionary processes on earth
have all the time they need.
The second problem stemmed from the conventional wisdom of the day, namely
that inheritance worked by a blending of maternal and paternal characters. The
essentials of this issue can be grasped by considering a trait (such as height or
weight) that can be described by a single variable. Suppose the mother departs
from the population average in this respect by an amount x and the father by an
amount y. Then, under a scheme of blending inheritance, the progeny will depart
1
from the mean by =2.x C y/. Suppose further that, in the parental generation, the
statistical scatter of the variable about its mean value is characterized by a variance
2 2 2 2
; that is, the expectation values of x andy are both . It is then straight-
forward to show that, with blending inheritance, the variance of this trait in the
next generation is halved. But persisting variability is the raw stuff upon which
natural selection works to produce descent with modification; it was critical to
Darwin’s ideas. This fundamental difficulty was pointed out to him, most notably
by the engineer Fleeming Jenkin. He acknowledged it as a problem, but – given the
observed persistence of variability in natural populations – he simply put it aside.
The resolution of this major difficulty lies, of course, in the fact that genes are inher-
ited in particulate Mendelian fashion, not by “blending”. And, as shown in 1908
independently by Hardy and by Weinberg, under Mendelian inheritance variability
remains unchanged from generation to generation, unless perturbed by factors such
as selection, mutation, statistical drift, or nonrandom mating.
In short, these first two of Darwin’s three truly major difficulties have been
entirely swept away by advances in our understanding of the natural world.
Darwin’s third major unsolved problem, which he himself arguably saw as the
most important, is not yet solved. This problem was, and still is, explaining how
cooperative behavior among animals evolved. The present volume is devoted to
recent advances toward a solution.
At first glance, the answer seems easy. You pay some small cost to gather a
much larger cooperative benefit. For example, a prairie dog takes a personal risk
in giving an alarm call, but all the colony benefits and, by taking turns as alarm
giver, each individual’s group benefit exceeds the occasional risk. But any such
arrangement is immediately vulnerable to cheats who enjoy the benefits without
paying the risk-taking dues. In evolutionary terms, such risk-avoiding cheats have
a selective advantage. Today we would say their enhanced probability of survival,
and consequent greater reproductive success, means their uncooperative behavior is
more represented in the next generation (possibly via their genes, or alternatively by
teaching their offspring – Dawkin’s memes). It is thus unclear how such observed
cooperative phenomena can arise, or if it does, how it can be maintained.
Following work on “kin selection” by Hamilton and others, a century after
Darwin, we now understand how such cooperative associations can evolve and per-
sist in relatively small groups of sufficiently closely related individuals. This would
Foreword ix
seem to solve the problem for many non-human groups of animals, which are indeed
found in such small kin groups. In particular, haplo-diploid systems of genetic inher-
itance, where siblings share more genes than do parents with their offspring, further
facilitate such kin selection, which can help explain apparent altruism among some
social insects.
During the hundred-thousand years and more when humans existed as small
bands of hunter-gatherers, such considerations of kin selection could well have
promoted cooperative behavior. But for large aggregations of essentially unrelated
individuals, as developed once agriculture appeared some ten millennia ago and
cities began, the origin of cooperative associations – with group benefits which
exceed the “cost of membership” – remains almost as puzzling today as it was for
Darwin.
Nor is this some abstract, academic problem for evolutionary biologists. The past
150 years have seen the human population increase sevenfold, and the ecological
footprint of the average individual also increase sevenfold, for an overall 50-fold rise
in our impacts on the planet. And these impacts are still increasing. There are con-
sequently huge and global problems – climate change, loss of biological diversity,
pressure on water supplies, and much else – which demand globally cooperative
solutions. These problems are further compounded by the fact that nations must
cooperate, but – given past history – in equitable proportions.
These problems have recently received an increasing amount of attention in
the scholarly literature, employing a variety of metaphors: the Tragedy of the
Commons; the Free-Rider problem; the Prisoner’s Dilemma; and others. These
metaphors are allied to artificial games in which the subjects (usually undergradu-
ates) trade small sums of money to test limits to altruism and tolerance of cheating.
Given Research Councils’ resources, the sums involved in these experiments are
necessarily relatively small, and I think it likely that some of the results would be
significantly different with larger stakes. Moreover essentially none of this work,
either theoretical or experimental, deals with situations where the costs and benefits
vary among the players, as it commonly does in the real world (actions to ameliorate
future climate change provide one striking example).
This book on Games, Groups, and the Global Good thus brings together a diverse
collection of evolutionary biologists and economists, reviewing recent advances and
speculating on the way forward.
Early chapters set the stage, with Frank surveying the biological foundations of
the problem and Nowak setting out a possible taxonomy of distinct mechanisms
whereby cooperation may evolve in the absence of kin selection. Bowles and Gintis
sketch a view based more on economic thinking about wider aspects of choice
in situations involving other people. Maskin gives a beautifully clear account of
evolutionary stable strategies in games where two players repeatedly interact, with
symmetric pay-offs, but with the realistic additional feature that random mistakes
can occur. Skyrms concludes this overview by suggesting it may be better to deal
with ever-changing networks of interacting individuals.
The second broad section of the book considers cooperation and group formation.
Here we largely move away from models and games, into what might be called a
x Foreword
more humanistic idiom: Flack and Krakauer on the origins of moral systems; Levin
on games, norms and societies; D. S. Wilson on prosociality and the evolution of
institutions; Johnson on supernatural punishment and cooperation; Hare on moral
motivation; and Appiah on explaining religion. I particularly like Levin’s discus-
sion of how social institutions and “norms” may arise, and how they can develop
over time. He sees evolutionary strategies evolving as diffuse responses to collec-
tions of situations, which in turn points to “new territory for game theorists”. I also
enjoyed Johnson’s answer to the often-asked question: if punishing cheats promotes
cooperative behavior, who takes on the costly job of punishment? He suggests that
inventing supernatural entities who are ultimately responsible can be a good answer,
and he elaborates this with a correspondingly new framework for game theoretic
models.
The third and final set of chapters address practical issues involving coopera-
tion and problems of the commons: Ostrom on building trust to solve commons
dilemmas; Brams and Kilgour on how democracy resolves conflict in difficult
games; O’Neill on the duty to apologize as part of a normative regime; concluding
with Sugden on team reasoning and market relationships. Specifically, Brams and
Kilgour suggest that democratic processes can “stabilize cooperative outcomes” by
giving voters a clear choice between a cooperative outcome and the inferior conse-
quences of failing to cooperate. It is, however, far from clear to me that democratic
processes work in this way in the real world, as distinct from pious idealizations of
it. Apart from anything else, extensive work on “the theory of voting”, dating back
200 years to Condorcet (and later Dodson, aka Lewis Carroll), show for instance
that the result of democratic choices among three alternatives – A, B, C – can under
certain circumstances result in any one of the three emerging as the winner, depend-
ing on how the voting system (or sequence of choices) is organized; Ken Arrow’s
Nobel Prize in economics was awarded essentially for his independent rediscovery
of this fascinating phenomenon. O’Neill’s discussion of the role of explicit apology
is most intriguing, and several contemporary illustrations of his thesis come to mind
(e.g. the recent and very positive outcome of Australia’s new Prime Minister, Kevin
Rudd, saying “sorry” to the Aboriginal population). Sugden’s chapter was presum-
ably written before financial markets collapsed, around the world. He notes that
“economics cannot represent the idea that paradigm market relationships : : : have
moral content”. It can be argued that recent events suggest that economic theory
should explicitly represent such ideas, and do so in clearly enforceable ways!
My personal musings about how cooperative human societies evolved – which
have some parallels with Johnson’s thoughts about “supernatural punishment and
cooperation” – are both less academic and analytic, and more gloomy. Once we
move out of the mists of pre-history, we find stories of dreamtime, creation myths,
ceremonies and initiation rites, spirits and gods, with a unifying theme that all seek
simultaneously to help explain the external world and also to provide a “stabilization
matrix” for a cohesive society. There are, moreover, some striking and unexplained
similarities in belief systems and rituals from different times and places. Conscience,
a simple word for a complex concept which helps foster behavior in accord with
society’s professed norms, has been memorably defined by H. L. Mencken as “the