Brian Skyrms

Evolutionary dynamics of collective action in N-person stag hunt dilemmas

In the animal world, collective action to shelter, protect and nourish requires the cooperation of group members. Among humans, many situations require the cooperation of more than two individuals simultaneously. Most of the relevant literature has focused on an extreme case, the N-person Prisoners Dilemma. Here we introduce a model in which a threshold less than the total group is required to produce benefits, with increasing participation leading to increasing productivity. This model constitutes a generalization of the two-person stag hunt game to an N-person game. Both finite and infinite population models are studied. In infinite populations this leads to a rich dynamics that admits multiple equilibria. Scenarios of defector dominance, pure coordination or coexistence may arise simultaneously. On the other hand, whenever one takes into account that populations are finite and when their size is of the same order of magnitude as the group size, the evolutionary dynamics is profoundly affected: it may ultimately invert the direction of natural selection, compared with the infinite population limit.

Causation, Chance and Credence

A. On the Nature of Probabilistic Causation.- Causality Testing in a Decision Science.- Causal Tendency: A Review.- Intuitions: Good and Not-So-Good.- Response to Salmon.- Regular Associations and Singular Causes.- Eliminating Singular Causes: Reply to Nancy Cartwright.- Reply to Ellery Eells.- Probabilistic Causal Levels.- Probabilistic Causality in Space and Time.- B. Physical Probability, Degree of Belief, and De Finettis Theorem.- Symmetry and Its Discontents.- A Theory of Higher Order Probabilities.- Conditioning, Kinematics, and Ex-changeability.- Ergodic Theory and the Foundations of Probability.- Indexes.

Evolutionary dynamics of Lewis signaling games: signaling systems vs. partial pooling

Transfer of information between senders and receivers, of one kind or another, is essential to all life. David Lewis introduced a game theoretic model of the simplest case, where one sender and one receiver have pure common interest. How hard or easy is it for evolution to achieve information transfer in Lewis signaling?. The answers involve surprising subtleties. We discuss some if these in terms of evolutionary dynamics in both finite and infinite populations, with and without mutation.

Signals, Evolution and the Explanatory Power of Transient Information

Pre‐play signals that cost nothing are sometimes thought to be of no significance in interactions which are not games of pure common interest. We investigate the effect of pre‐play signals in an evolutionary setting for Assurance, or Stag Hunt, games and for a Bargaining game. The evolutionary game with signals is found to have dramatically different dynamics from the same game without signals. Signals change stability properties of equilibria in the base game, create new polymorphic equilibria, and change the basins of attraction of equilibria in the base game. Signals carry information at equilibrium in the case of the new polymorphic equilibria, but transient information is the basis for large changes in the magnitude of basins of attraction of equilibria in the base game. These phenomena exemplify new and important differences between evolutionary game theory and game theory based on rational choice.

Evolution of signalling systems with multiple senders and receivers

Sender–receiver games are simple, tractable models of information transmission. They provide a basic setting for the study the evolution of meaning. It is possible to investigate not only the equilibrium structure of these games but also the dynamics of evolution and learning—with sometimes surprising results. Generalizations of the usual binary game to interactions with multiple senders, multiple receivers or both provide the elements of signalling networks. These can be seen as the loci of information processing, group decisions, and teamwork.

The Prior Propensity Account of Subjunctive Conditionals

I agree with Ernest Adams and Brian Ellis that assertability of uniterated indicative conditionals goes by epistemic conditional probability. It might be thought that subjunctives are mere stylistic variants of indicatives, the counterfactual being used only to convey the extra information that we are in a counterfactual belief state. There are striking examples which argue that this is not always the case.

Co-evolution of pre-play signaling and cooperation

A finite-population dynamic evolutionary model is presented, which shows that increasing the individual capacity of sending pre-play signals (without any pre-defined meaning), opens a route for cooperation. The population dynamics leads individuals to discriminate between different signals and react accordingly to the signals received. The proportion of time that the population spends in different states can be calculated analytically. We show that increasing the number of different signals benefits cooperative strategies, illustrating how cooperators may take profit from a diverse signaling portfolio to forecast future behaviors and avoid being cheated by defectors.

Network formation by reinforcement learning: the long and medium run

Abstract We investigate a simple stochastic model of social network formation by the process of reinforcement learning with discounting of the past. In the limit, for any value of the discounting parameter, small, stable cliques are formed. However, the time it takes to reach the limiting state in which cliques have formed is very sensitive to the discounting parameter. Depending on this value, the limiting result may or may not be a good predictor for realistic observation times.

Updating, supposing, and maxent

ConclusionThe philosophical controversy concerning the logical status of MAXENTmay be in large measure due to the conflation of two distinct logical roles:(1) A general inductive principle for updating subjective probabilities (2)a supposing rule for moving from one chance probability to another.When judged under standards of dynamic coherence appropriate to (1),MAXENT is found wanting. When judged in terms of the logic appro-priate to (2) MAXENT yields for convex closed constraint sets a reason-able selection function with interesting connections with sufficiency andconditioning. Indeed it is just the features of MAXENT which make itappropriate for (2) which make it inappropriate for (1). MAXENT canbe thought of as part of Bayesian logic. But it is part of the logic ofsupposition rather than the logic of induction.

Carnapian Inductive Logic for Markov Chains

Carnap’s Inductive Logic, like most philosophical discussions of induction, is designed for the case of independent trials. To take account of periodicities, and more generally of order, the account must be extended. From both a physical and a probabilistic point of view, the first and fundamental step is to extend Carnap’s inductive logic to the case of finite Markov chains. Kuipers (1988) and Martin (1967) suggest a natural way in which this can be done. The probabilistic character of Carnapian inductive logic(s) for Markov chains and their relationship to Carnap’s inductive logic(s) is discussed at various levels of Bayesian analysis.