Download Conference handbook - University of Bristol

Evolution, Co-operation and Rationality Conference
September 18th-20th
Department of Philosophy, University of Bristol
Burwalls Centre for Continuing Education, Suspension Bridge
Road, University of Bristol
www.bristol.ac.uk/evolution-cooperation
[email protected]
Photograph by Dr. Loren Picco, Department of Physics, University of Bristol. See more at http://www.flickr.com/photos/godofbiscuits/
Conference handbook
In this handbook you will find a list of all the talks and posters included in the conference,
and a timetable. You should also have a map and a name badge – please do wear your name
badge throughout the conference. If you mislay your map or anything else, do please ask Jess
as she will almost certainly have some spares.
Map: This shows the local area, including all the hotels and the road out to Burwalls
(Suspension Bridge Road, square A4).
Food: Coffee, tea and lunch will be provided at Burwalls on all three days. Clifton Village (see
map) has lots of good places to eat on Friday evening (and Saturday evening if you’re not
joining us for the conference dinner at Burwalls).
Tourism: for those of you with partners and/or children wishing to explore Bristol, the
Suspension Bridge itself is probably Bristol’s most significant tourist attraction, but the
famous Zoo, city cathedral and city museum are all within easy reach (again, see the map –
these were provided by ‘Visit Bristol’ so all the tourist places are marked). Bristol Festival (a
celebration of all things Bristol – mainly music, dance and visual art) is in its second year and
will be happening over the same weekend as the conference – see
www.thebristolfestival.org/.
Rushing about: several people asked about good places to go running. The Downs around the
Avon Gorge (see map – just cross the bridge from Burwalls and turn left) are spectacular for
walking or moving at speed. If you prefer something muddier, Leigh Woods can be reached
easily from Burwalls (cross Suspension Bridge Road and turn right almost immediately onto
North Road. Keep going until you find the entrance to the woods on your right – it’s a few
hundred yards).
Internet access: there is no internet access at Burwalls itself. However, those of you staying at
the Rodney or the Clifton should have wifi in your rooms.
TALKS AND POSTERS (chronological order)
Friday 18th September – all sessions in the Lecture Theatre
Alasdair Houston, Biological Sciences, University of Bristol
Natural selection and rational decisions
(Chair: Samir Okasha)
Stuart West, Zoology, University of Oxford
The evolution of co-operation in humans
(Chair: Samir Okasha)
Pete Trimmer, Computer Science, University of Bristol
Bayes, bias and the evolution of decision-making mechanisms
Understanding how animals, including humans, deal with uncertainty is a central issue in the
study of decision-making. We consider a number of simple models which highlight how the
optimal bias of an estimate will depend upon how the estimate was generated. We discuss the
Bayesian mindset with regard to the Ellsberg Paradox, which provides an example of
decision-making between two-types of uncertainty: ‘risk’, in which the probability of an event
is known, and ‘ambiguity’, in which probabilities are unknown. We conclude with an example
where two Bayesian decision-making systems have been combined, resulting in optimal
‘biases’ in each sub-system.
(Chair: Jonathan Grose)
Fabio Paglieri, Institute of Cognitive Sciences and Technologies, Rome
Nobody likes to wait in vain: on the evolutionary roots of impulsivity
People often change their preferences due to the mere passing of time. This seems to violate
rationality, since, absent relevant changes, the subject’s preferences should remain stable over
time. However, the normative validity of this intuition appears debatable. Are we justified in
expecting that the passing of time will not generate relevant changes in the options of
choice? People may look at later options as inherently less certain than earlier ones, and have
no clear intuition on the level of uncertainty connected with delayed outcomes (higher-order
uncertainty). This affects expected utility, in ways compatible with dynamic inconsistency.
Experimental evidence on the connection between time preferences and probabilistic
reasoning are not conclusive. Evolutionary considerations may help solving this checkmate,
by proposing a subtler connection: the uncertainty characteristic of a given environment may
contribute to define optimal levels of impulsivity for decision-making in that environment.
Two main lines of evidence support this hypothesis: studies with non-human primates,
where delay of gratification is suggested to correlate with ecological factors, and social
simulations with neural networks, where artificial agents are evolved under different
conditions and then tested in inter-temporal choice.
(Chair: Jonathan Grose)
Posters will go on display just before the Friday afternoon tea break, outside the
lecture theatre. They will remain on display in this area throughout the conference
and the authors are encouraged to be on hand to answer questions in the extended
tea break on Friday afternoon. The authors are as follows:
Athena Aktipis, Ecology and Evolutionary Biology, University of Arizona
Is self-interest the natural state of the world? The ability to Walk Away promotes the evolution of cooperation
Katrin Fehl, Courant Research Centre Evolution of Social Behaviour, University of
Göttingen
Altruistic enforcement of an egalitarian sharing norm
Karolina Sylwester & Gilbert Roberts, Centre for Behaviour and Evolution, University
of Newcastle
Rational? Altruistic? Or rationally altruistic? Co-operators benefit through reputational-based partner choice
Jack Vromen, Erasmus Institute for Philosophy and Economics, Erasmus University
Is reciprocity-based human cooperation “ultimately” selfish or altruistic? A red herring
Proponents of strong reciprocity (SR) often argue that whereas SR is altruistic, reciprocal
altruism (RA) “ultimately” is selfish. Proponents of RA (such as Burnham and Johnson
2005) sometimes seem to agree. I, by contrast, argue in the paper that this depiction of the
difference between SR and RA is misleading. Although both parties acknowledge that
evolutionary and psychological altruism should be distinguished sharply, it seems that they
are conflated in the above depiction of the difference between SR and RA. Both parties seem
to conclude from the fact that reciprocal altruists initially suffer fitness disadvantages, but
ultimately attain superior fitness (and this is why RA is regarded as evolutionarily selfish) that
reciprocal altruists are willing to incur temporary fitness losses because by doing so they
expect to attain their ultimate goal: realizing superior fitness (and this is why it would be
psychologically selfish). This is a non sequitur, however. As Trivers (1971) already noted, the
psychological mechanism underlying reciprocal altruism might differ considerably from such
a calculating self-serving one. At the other side, it is not clear at all that SR should be
regarded as evolutionarily and psychologically altruistic. Upon closer inspection of models of
how SR could have evolved it appears that if RA is regarded as evolutionarily selfish on the
ground that it ultimately yields superior net fitness benefits, then SR is evolutionarily selfish
too. And in what is perhaps the most often cited study on the proximate mechanisms
underlying SR to date, de Quervain et al. (2004), it is explicitly denied that altruistic
punishments are psychologically altruistic. The paper also argues that the focus on whether
or not SR and RA are ultimately truly altruistic diverts attention from more important issues.
One such issue is empirical: how reliable and robust are costly rewarding and punishing
behavior displayed in anonymous, one-shot games. Another issue is whether it is plausible to
assume that at a more concrete level (than at the abstract level of counting direct and indirect
compensating fitness benefits) processes of group competition and conflict and social
learning played an important role in the actual evolution of (particular forms of) human
cooperation.
(Chair: Christopher Clarke)
Armin Schulz, Philosophy, University of Wisconsin - Madison
Rational Choice Theory and the Argument from Evolution: An Assessment
In this paper, I critically discuss a recent innovation in the debate surrounding the descriptive
accuracy of Rational Choice Theory (RCT): the appeal to evolutionary theory. To do this, I
analyse Gigerenzer et al.’s argument that natural selection favours a decision making
mechanism based on ‘simple heuristics’ over one based on RCT, and try to show that it faces
two major problems. Firstly, it is evidentially implausible, as we lack crucial information
about our past, and secondly, it is methodologically implausible, as it is unable to suggest
genuinely novel phenomena to investigate. Since the same problems also befall many other
evolutionary arguments in this area, I therefore conclude that the evolutionary perspective
has, as yet, contributed little to the debate surrounding RCT.
(Chair: Christopher Clarke)
Peter Hammerstein, Biology, University of Berlin
Biological and Cultural roots of human decision-making
(Chair: Ken Binmore)
(Prof. Binmore’s talk will be followed by drinks on the terrace)
Saturday 19th September – Lecture Theatre or Music Room during parallel sessions
John McNamara, Mathematics, University of Bristol
The importance of individual differences in conflict and the evolution of co-operation
(Chair: Samir Okasha)
Ken Binmore, Philosophy and Economics, University of Bristol; Economics, UCL
Sexual drift
(Chair: Samir Okasha)
Christophe Heintz, Konrad Lorenz Institute for Evolution and Cognition Research
Bounded Machiavellians in action: social heuristics versus social preferences
In my talk, I will analyse the alternative explanations for the cultural diversity of economic
prosocial behaviour and the theories on which they are grounded. The main explanation that
has been put forward in the recent works of Fehr, Gächter, Henrich and others appeals to
variations in social preferences. Against this view, I will defend the hypothesis that variations
in prosocial behaviour as observed in experimental games are due to differences in learned
economic behavioural routines, exploiting and adapted to the specifics of the local
institutions. According to this hypothesis, the social preferences are in fact relatively similar
across cultures; the variation is due to differences in the learned solutions, which can take the
form of ‘rules of thumb’, to common economic or strategic problems and their specific
cultural forms.
(Chair: Cedric Paternotte, Lecture Theatre)
Marco Archetti, Zoology, University of Oxford
Beyond the Prisoner’s Dilemma: optimal contracts and civic duty games from economics to evolutionary biology
I discuss how two ideas developed in economics and in the social sciences can be applied to
the problem of cooperation in evolutionary biology. First, contract theory can explain
cooperation in pairwise interactions without iterations: setting the right costs and rewards
can stabilize mutualism because the possible partners will screen themselves according to
their own interest. Second, social dilemmas can often be described as a volunteer’s dilemma
rather than as a prisoner's dilemma; in this case cheaters can invade and form a stable mixed
equilibrium with co-operators, and it is possible to envisage practical solutions to increase
cooperation.
(Chair: Christopher Clarke, Music Room)
Christine Clavien, Ecology and Evolution, University of Lausanne & Chloe
FitzGerald, Philosophy, University of Manchester
Could the social sciences help make evolutionary game theory useful for industry and policy-makers?
Policy makers and industry need to accurately predict the impact of specific changes on
microeconomic systems. Classical economics is well designed for this, being predictive as
well as explanatory, but there are some well-known problems associated with classical
economics. At first glance, evolutionary game theory looks like a promising alternative, but
when one takes a closer look, it fails to reflect the plasticity of human psychology, is too
general and is more explanatory than predictive. We suggest that evolutionary game theory
could increase its relevance to policy-makers if it focussed on specific microeconomic
systems and worked in closer collaboration with the social sciences. We will provide some
suggestions as to how this could be done.
(Chair: Cedric Paternotte, Lecture Theatre)
Larissa Conradt, Biology and Environmental Science, University of Sussex
Stakeholders and the evolution of ‘fair’ decision-making in animals
Time and again, we face consensus decisions with conspecifics. We have to agree with
partners on dates; with friends on restaurants; with colleagues on collaborative roles, etc.
Social animals face very similar collective decisions. To maintain group cohesion, group
mates have to agree movements and activities. To work efficiently, co-operators have to
agree tasks. Not always do we, or animals, reach consensus easily. There can be conflicts of
interest. Some might be hungry, while others are tired. Nobody in a common enterprise
might want to take on dangerous jobs. Humans solve these conflicts by verbal negotiation,
by agreed rules of preferences aggregating or by referring to established power hierarchies. I
examine collective decision-making in animals.
(Chair: Christopher Clarke, Music Room)
Till Grüne-Yanoff, Social and Moral Philosophy, University of Helsinki
Evolutionary Game Theory, Interpersonal Comparisons and Selection-type Models: A Dilemma
Evolutionary game theory is capable of modelling natural selection both in biological as well
as in social contexts, because it offers causal relations between the interaction of
players and the differential reproduction of strategies. However, all models of these causal
processes require to make interpersonal comparisons of payoffs, at least on the
level of the modeller, but often also on the level of the modelled players. Such interpersonal
preference comparisons can be avoided, if players are reinterpreted to counterfactually
consider alternative strategies as if they weretheir own, and to judge the resulting payoffs as
their own utilities. However, such a reinterpretation does not offer causal relations between
the interaction of players and the differential reproduction of strategies, and hence violates a
necessary condition for modelling natural selection. Avoiding interpersonal comparisons
prevents the modelling of natural selection, and thus, natural selection implies interpersonal
comparisons in evolutionary games. This poses a genuine dilemma, which indicates the
limitations of applying evolutionary game theory in the social sciences.
(Chair: Cedric Paternotte, Lecture Theatre)
Alejandro Rosas Lopez, Philosophy, National Universidad Nacional de Colombia
Punishment sustains co-operation without the second-order free rider problem
I explore a solution to the evolution of ‘altruistic’ punishment where punishment is weakly
altruistic. If defectors change into cooperators in the public goods game, punishers obtain a
profit from punishing. Suppose the profit is greater than the cost of punishment. Then,
punishers are weakly altruistic and play with forgivers n-person Chicken dilemma, where T >
R > S > P represents the preferences over outcomes. Since S > P, it pays a cooperator to
provide the second order public good (punishment) unilaterally. From this perspective, no
conflict arises between the benefit to individuals and the benefit to groups.
(Chair: Christopher Clarke, Music Room)
Werner Güth, Strategic Interaction Group, Max Planck Institute of Economics
Pull, push or both? Indirect evolution in economics and beyond
(Chair: Jonathan Grose)
Gerd Gigerenzer, Centre for Adaptive Behaviour and Cognition, Max Planck Institute
for Human Development
Homo Heuristicus: decision-making in an uncertain world
(Chair: Jonathan Grose)
(Prof. Gigerenzer’s talk will be followed by drinks at the Avon Gorge Hotel and the conference dinner)
Sunday 20th September – all sessions in the Lecture Theatre
Samir Okasha, Philosophy, University of Bristol
Evolution, risk and rational decision
(Chair: Cedric Paternotte)
David H. Wolpert, NASA (with Julian Jamison, Economics, Yale and Michael Harre,
Computer Science, Sydney University)
Schelling formalized: strategic choices of non-rational personas
We introduce a framework that explains non-rationality in non-repeated games. In our
framework a player $i$ adopts a binding “persona” - a temporary utility function - that they
honestly signal before play. By adopting a “non-rational” persona, $i$ may cause changes in
their opponents' behavior that increases $i$'s true utility. We use this framework to explain
experimental data in the Traveler's Dilemma and to show how cooperation can arise in the
Prisoner's Dilemma (PD). We then predict a crowding out phenomenon in the PD. We also
predict a trade-off between the robustness and the benefit of cooperation in the PD.
(Chair: Ellen Clarke)
Jean-Baptiste Andre, Ecology and Evolution, Centre National de la Recherche
Scientifique
The evolution of reciprocity: Social types or social incentives?
The vast majority of human beings regularly engage in reciprocal cooperation with nonrelated conspecifics, and yet the current evolutionary understanding of these behaviors is
insufficient. Intuitively, reciprocity makes sense if past partner’s behavior conveys
information about her future behavior. But it is not straightforward to understand why this
should be an outcome of evolution. Most evolutionary models assume that each individual is
characterized by a stable “social type” (defector, cooperator, reciprocator, etc.). In this case, one's
behavior unavoidably informs about one's social type, which makes it sensible to reciprocate.
In the present paper, after describing the central source of difficulty in the evolutionary
understanding of reciprocity, I put forward an alternative explanation based on, but
significantly modified from, a work by Leimar (1997b). It consists in taking into account the
fact that the payoffs to individuals in social interactions, their incentives, are variable and can
change through time. This offers a solution because one’s behavior then signals one’s
payoffs, which makes it sensible to reciprocate. Even though the overwhelming majority of
evolutionary models implicitly endorse the social types mechanism, I argue in the paper that
the social incentives mechanism is more likely to underlie reciprocity in humans. This leads
to a novel understanding of reciprocity, some of the consequences of which will be
discussed.
(Chair: Ellen Clarke)
Simon Huttegger, Logic and Philosophy of Science, University of California & Rory
Smead, Logic and Philosophy of Science, University of California
Group Selection and the Stag Hunt
We investigate Maynard Smith's haystack model in the context of the Stag Hunt. In the Stag
Hunt, cooperation constitutes a Pareto dominant equilibrium. This accounts for several
differences between our haystack model and Maynard Smith's original treatment. In
particular, we establish conditions under which the population will evolve to the cooperative
equilibrium, and compare the results to when this happens in populations without haystacks.
Brian Skyrms, Logic and Philosophy of Science, University of California
Evolution, information and deception in signalling games
(Chair: Ken Binmore)