In the spatialized Prisoner's Dilemma, players compete against their immediate neighbors and adopt a neighbor's strategy should it prove locally superior. Fields of strategies evolve in the manner of cellular automata (Nowak and May, 1993; Mar and St. Denis, 1993a,b; Grim 1995, 1996). Often a question arises as to what the eventual outcome of an initial spatial configuration of strategies will be: Will a single strategy prove triumphant in the sense of progressively conquering more and more territory without opposition, or (...) will an equilibrium of some small number of strategies emerge? Here it is shown, for finite configurations of Prisoner's Dilemma strategies embedded in a given infinite background, that such questions are formally undecidable: there is no algorithm or effective procedure which, given a specification of a finite configuration, will in all cases tell us whether that configuration will or will not result in progressive conquest by a single strategy when embedded in the given field. The proof introduces undecidability into decision theory in three steps: by (1) outlining a class of abstract machines with familiar undecidability results, by (2) modelling these machines within a particular family of cellular automata, carrying over undecidability results for these, and finally by (3) showing that spatial configurations of Prisoner's Dilemma strategies will take the form of such cellular automata. (shrink)

Most current modeling for evolution of communication still underplays or ignores the role of local action in spatialized environments: the fact that it is immediate neighbors with which one tends to communicate, and from whom one learns strategies or conventions of communication. Only now are the lessons of spatialization being learned in a related field: game-theoretic models for cooperation. In work on altruism, on the other hand, the role of spatial organization has long been recognized under the term ‘viscosity’.Here we (...) offer some simple simulations that dramatize the importance of spatialization for studies of both cooperation and communication, in each case contrasting a model dynamics in which strategy change proceeds globally, and a spatialized model dynamics in which interaction and strategy change both operate purely locally. Local action in a spatialized model clearly favors the emergence of cooperation. In the case of communication, spatialized models allow communication to arise and flourish where the global dynamics more typical in the literature make it impossible.Simulations make a dramatic case for spatialized modeling, but analysis proves difficult. In a final section we outline some of the surprises of spatial dynamics but also some of the complexity facing attempts at deeper analysis. (shrink)

The goal of philosophy of information is to understand what information is, how it operates, and how to put it to work. But unlike ‘information’ in the technical sense of information theory, what we are interested in is meaningful information. To understand the nature and dynamics of information in this sense we have to understand meaning. What we offer here are simple computational models that show emergence of meaning and information transfer in randomized arrays of neural nets. These we take (...) to be formal instantiations of a tradition of theories of meaning as use. What they offer, we propose, is a glimpse into the origin and dynamics of at least simple forms of meaning and information transfer as properties inherent in behavioral coordination across a community. (shrink)

‘The problem with simulations is that they are doomed to succeed.’ So runs a common criticism of simulations—that they can be used to ‘prove’ anything and are thus of little or no scientific value. While this particular objection represents a minority view, especially among those who work with simulations in a scientific context, it raises a difficult question: what standards should we use to differentiate a simulation that fails from one that succeeds? In this paper we build on a structural (...) analysis of simulation developed in previous work to provide an evaluative account of the variety of ways in which simulations do fail. We expand the structural analysis in terms of the relationship between a simulation and its real-world target emphasizing the important role of aspects intended to correspond and also those specifically intended not to correspond to reality. The result is an outline both of the ways in which simulations can fail and the scientific importance of those various forms of failure. (shrink)

We extend previous work on cooperation to some related questions regarding the evolution of simple forms of communication. The evolution of cooperation within the iterated Prisoner's Dilemma has been shown to follow different patterns, with significantly different outcomes, depending on whether the features of the model are classically perfect or stochastically imperfect (Axelrod 1980a, 1980b, 1984, 1985; Axelrod and Hamilton, 1981; Nowak and Sigmund, 1990, 1992; Sigmund 1993). Our results here show that the same holds for communication. Within a simple (...) model, the evolution of communication seems to require a stochastically imperfect world. (shrink)

To understand the meaning of evolutionary equilibria, it is necessary to comprehend the ramifications of the evolutionary model. For instance, a full appreciation of Axelrod's The Evolution of Cooperation requires that we identify assumptions under which conditionally cooperative strategies, like Tit For Tat, are and are not evolutionarily stable. And more generally, when does stability fail? To resolve these questions we re-examine the very foundations of the evolutionary model. The results of this paper can be analytically separated into three parts. (...) The first part is conceptual: it identifies the evolutionary model's assumptions and shows how different assumptions imply different types of evolutionary stability. The second part is deductive: it establishes necessary and sufficient conditions for the types of evolutionary stability identified in the first part, and demonstrates in which games these kinds of stability can (and cannot) be attained. The third and final part is applied: it relates the general findings (which are independent of the specific payoffs of any particular evolutionary game) to the issue of the evolutionary stability of cooperation. Results on cooperation appear throughout the paper as they both exemplify and motivate the general results. These results essentially explain when cooperation is and is not stable, and why, thus shedding new light on the meaning and applicability of Axelrod's widely known claims. (shrink)

Because the label "computing and philosophy" can seem like an ad hoc attempt to tie computing to philosophy, it is important to explain why it is not, what it studies (or does) and how it differs from research in, say, "computing and history," or "computing and biology". The American Association for History and Computing is "dedicated to the reasonable and productive marriage of history and computer technology for teaching, researching and representing history through scholarship and public history" (http://theaahc.org). More pervasive, (...) work in computing and biology enjoys the convenient name of "bioinformatics...the science of using information to understand biology..., a subset of the larger field of computational biology, the application of quantitative analytical techniques in modeling biological systems" (http://oreilly.com/catalog/bioskills/chapter/ ch01.html). The recent venture of the Association for Computing Machinery and the Institute of Electrical and Electronics Engineers to publish the Transactions on Computational Biology and Bioinformatics (TCBB) bears witness to the reach of computing and biology and underscores its objective. TCBB intends to report "archival research results related to the algorithmic, mathematical, statistical, and computational methods that are central in bioinformatics and computational biology; the development and testing of effective computer programs in bioinformatics; the development and optimization of biological databases; and important biological results that are obtained from the use of these methods, programs, and databases" (http://tcbb.acm.org). In the case of "computing and history" and "bioinformatics," each discipline stands in a particular relationship to computers that raises questions unique to itself. But both are devoted to the development of computational tools to aid discovery.. (shrink)

What is it for a sound or gesture to have a meaning, and how does it come to have one? In this paper, a range of simulations are used to extend the tradition of theories of meaning as use. The authors work throughout with large spatialized arrays of sessile individuals in an environment of wandering food sources and predators. Individuals gain points by feeding and lose points when they are hit by a predator and are not hiding. They can also (...) make sounds heard by immediate neighbours in the array, and can respond to sounds from immediate neighbours. No inherent meaning for these sounds is built into the simulation; under what circumstances they are sent, if any, and what the response to them is, if any, vary initially with the strategies randomized across the array. These sounds do take on a specific function for communities of individuals, however, with any of three forms of strategy change: direct imitation of strategies of successful neighbours, a localized genetic algorithm in which strategies are ‘crossed’ with those of successful neighbours, and neural net training on the behaviour of successful neighbours. Starting from an array randomized across a large number of strategies, and using any of these modes of strategy change, communities of ‘communicators’ emerge. Within these evolving communities the sounds heard from immediate neighbours, initially arbitrary across the array, come to be used for very specific communicative functions. ‘Communicators’ make a particular sound on feeding and respond to that same sound from neighbours by opening their mouths; they make a different sound when hit with a predator and respond to that sound by hiding. Robustly and persistently, even in simple computer models of communities of self-interested agents, something suggestively like signalling emerges and spreads. Keywords: meaning, communication, genetic algorithms, neural networks. (shrink)

We work with a large spatialized array of individuals in an environment of drifting food sources and predators. The behavior of each individual is generated by its simple neural net; individuals are capable of making one of two sounds and are capable of responding to sounds from their immediate neighbors by opening their mouths or hiding. An individual whose mouth is open in the presence of food is “fed” and gains points; an individual who fails to hide when a predator (...) is present is “hurt” by losing points. Opening mouths, hiding, and making sounds each exact an energy cost. There is no direct evolutionary gain for acts of cooperation or “successful communication” per se. In such an environment we start with a spatialized array of neural nets with randomized weights. Using standard learning algorithms, our individuals “train up” on the behavior of successful neighbors at regular intervals. Given that simple setup, will a community of neural nets evolve a simple language for signaling the presence of food and predators? With important qualifications, the answer is “yes.” In a simple spatial environment, pursuing individualistic gains and using partial training on successful neighbors, randomized neural nets can learn to communicate. (shrink)

How do conventions of communication emerge? How do sounds or gestures take on a semantic meaning, and how do pragmatic conventions emerge regarding the passing of adequate, reliable, and relevant information? My colleagues and I have attempted in earlier work to extend spatialized game theory to questions of semantics. Agent-based simulations indicate that simple signaling systems emerge fairly naturally on the basis of individual information maximization in environments of wandering food sources and predators. Simple signaling emerges by means of any (...) of various forms of updating on the behavior of immediate neighbors: imitation, localized genetic algorithms, and partial training in neural nets. Here the goal is to apply similar techniques to questions of pragmatics. The motivating idea is the same: the idea that important aspects of pragmatics, like important aspects of semantics, may fall out as a natural results of information maximization in informational networks. The attempt below is to simulate fundamental elements of the Gricean picture: in particular, to show within networks of very simple agents the emergence of behavior in accord with the Gricean maxims. What these simulations suggest is that important features of pragmatics, like important aspects of semantics, don't have to be added in a theory of informational networks. They come for free. (shrink)