In the early 1980s, Robert Axelrod published several articles on The Evolution of Cooperation, discussing and interpreting the results of his well-known computer tournaments and of a series of subsequent simulations. Both the tournaments and simulations were conducted in order to find a suitable, evolutionary stable strategy for the iterated prisoner's dilemma, which is generally considered an appropriate model of a
certain type of social dilemma that arises when "the pursuit of self-interest by each leads to a poor outcome for all."
The results of the tournaments and simulations led to a generalized theory of the evolution of cooperation, which claims to provide an explanation for various historical, social and biological phenomena. Axelrod`s work contributed extensively to popularizing computer simulation as a scientific method in the social sciences. Besides the fact that his approach had an unquestionably high impact on succeeding research and ushered in the "simulation era" in the social sciences, the use Axelrod made of computer simulations raises questions about their methodological and epistemological
status: If, as Axelrod states in his paper "Advancing the Art of
Simulation in the Social Sciences", simulation can serve the purposes of prediction, proof and even scientific discovery, what need is there for conducting experiments any longer? Can't we simulate science?
Admittedly, this suggestion sounds somewhat exaggerated, but why exactly do most of us share the intuition that there are fundamental differences persisting between simulations and experiments? What are the characteristic features distinguishing them? Do computer simulations in general - and
Axelrod's tournaments in particular - resemble experiments insofar as their
potential to provide us with surprising results that permit further theorizing
is concerned? Or are they nothing else than mere "number-crunching techniques", using brute-force computational means in
order to generate data from theoretical knowledge and assumptions already built into the underlying model?
The question where to draw the conceptual line between simulation and
experiment has turned out to be of great interest to philosophy of science,
not least since the categorization might be relevant to the way the results
are assessed and used. The objective of this paper is to elaborate on the
distinctive characteristics of simulations in contrast to experiments.
At a glance
- Grade: 1.0 (1.0 is the best grade)
- Examines Robert Axelrod’s repeated prisoner’s dilemma computer tournament
- Includes the payoff matrix used by Axelrod
- 17 sources in the bibliography
Contents
1 Introduction
2 Axelrod’s computer tournament – a brief overview of its setup, results and further conclusions
3 Simulating science? Simulations versus experiments
3.1 Materiality
3.2 External validity and its justification
3.3 Interim findings
4 Are Axelrod’s conclusions externally valid?
5 Conclusion
Objectives and Themes
This paper investigates how computer simulations differ from experiments and whether simulation should be classified as theorizing, experimentation, or a distinct approach to scientific inquiry. It uses Robert Axelrod’s work on the evolution of cooperation to examine the justification and reach of simulation-based claims.
- Axelrod’s tournaments and the repeated prisoner’s dilemma
- Materiality and intervention in simulations and experiments
- Surprise, confounding results, and scientific discovery
- External validity and the role of background knowledge
- The relationship between simulated strategies and real-world cooperation
Excerpt from the Book
Materiality
Attempting to characterize the distinctive features of simulations and experiments, it is tempting to appeal to the fact that they do, prima facie, obviously differ in ontological respects: In an experiment, scientists intervene in a concrete, material system, whereas a simulation merely investigates an abstract, nonmaterial entity – a mathematical model.
Morgan (2003, p. 216) seems to have in mind this substantial disparity when she refers to simulations as “experiments without material intervention.” But apart from laboratory experiments and mathematical model experiments, which serve as limiting cases located at the opposed ends of an imaginary scale, Morgan (2003, p. 217) identifies two types of “modern hybrids”, computer simulations that, to a certain extent, rely on “semimaterial” inputs (e.g. photographs of real hipbone slices) or produce “pseudo-material” outputs, mimicking observations of a real-world process (e.g. stock market prices). Moreover, she conjectures that the degree of materiality directly correlates to an experiment’s or, respectively, a simulation’s epistemic power. Although mathematical model experiments – like Axelrod’s exploration of the repeated prisoner’s dilemma – “may function both as tools for theory development and as tools for understanding the world” (Morgan, 2003, p. 218) and therefore may well surprise the simulationist, they lack the power to produce confounding results. Even though TIT FOR TAT’s success in both tournaments came as a surprise, it could be understood and reconstructed in the aftermath, considering its properties and the given set of competing strategies. Things seem to be the same in the field of economics: Mathematical models may yield unexpected results, but, “having been surprised, we can go back through the model experiment and understand why such surprising results occured. That possibility may not be open to us with material experiments.” (Morgan, 2003, p. 220)
Following Morgan, this remarkable difference in epistemic power is due to the extent of attainable controllability, which, in turn, is linked to the degree of materiality involved in an experiment. Even though she acknowledges the fact that laboratory experiments, unlike raw observation in the field, take place under rather strict conditions of control, Morgan emphasizes that “the material world can only be controlled and manipulated to an extent” (Morgan, 2003, p. 219), while “confounding causes [...] can so easily be assumed away in the mathematical model experiment.” (Ibid.) A simulation – or its underlying model, transformed into a computable algorithm – is fully controllable and adjustable according to the actual requirements, and can thus be tailored to fit its specific purpose. On the other hand, in a laboratory experiment one cannot be absolutely sure to have ruled out every potentially disturbing factor.
Chapter Summaries
1 Introduction: The paper introduces Axelrod’s simulations and the philosophical question of how simulation relates to experiment. It outlines the concepts used to assess simulation’s scientific status.
2 Axelrod’s computer tournament – a brief overview of its setup, results and further conclusions: This chapter describes the repeated prisoner’s dilemma tournaments, the strategies submitted, and Axelrod’s subsequent population simulations and theory of cooperation.
3 Simulating science? Simulations versus experiments: The author compares accounts by Morgan, Parker, and Winsberg, focusing on materiality, experimental intervention, external validity, and background knowledge.
4 Are Axelrod’s conclusions externally valid?: This chapter applies the preceding discussion to Axelrod’s simulations, examining the relationship between their assumptions, their conclusions, and the real-world phenomena they seek to explain.
5 Conclusion: The final chapter brings together the argument about simulation and experiment and assesses its implications for Axelrod’s work.
Keywords
computer simulation, scientific experiments, philosophy of science, external validity, materiality, epistemic power, background knowledge, Robert Axelrod, evolution of cooperation, repeated prisoner’s dilemma, game theory, TIT FOR TAT, model building
Frequently Asked Questions
What is this paper about?
It examines the philosophical and methodological relationship between computer simulations and experiments, using Axelrod’s work on cooperation as a central case.
What are the central themes?
The paper addresses materiality, intervention, scientific discovery, external validity, background knowledge, and the interpretation of simulated cooperation.
What is the main research question?
The author asks whether simulation is a form of theorizing, a form of experimentation, or a distinct approach between theory and experiment.
What research method does the author use?
The paper develops a philosophical comparison of published accounts of simulation and experiment and applies that discussion to Axelrod’s computer tournaments.
What does the main discussion cover?
It reviews Axelrod’s tournament setup and compares Morgan’s, Parker’s, and Winsberg’s arguments before examining the external validity of Axelrod’s conclusions.
Which keywords characterize the paper?
Key terms include computer simulation, experiment, external validity, materiality, epistemic power, game theory, and the repeated prisoner’s dilemma.
How does the repeated prisoner’s dilemma feature in the discussion?
It provides the game-theoretical model for Axelrod’s tournaments, in which submitted decision rules competed through repeated interactions.
Why does the paper discuss materiality?
Materiality is one proposed way to distinguish simulation from experiment and to assess how researchers justify inferences from a model or experimental system to its target.
How is the external validity of Axelrod’s work assessed?
The author considers the simulations’ background knowledge, their relationship to empirical situations, and the sensitivity of their outcomes to modelling choices.
- Quote paper
- Bachelor of Arts (B.A.) Inga Bones (Author), 2010, Simulating Science?, Munich, GRIN Verlag, https://www.grin.com/document/148996
