Reviewed by
Corinna Elsenbroich
University of Surrey
The jury is still out regarding the scientific status of thought experiments but this collection of papers greatly contributes to the clarification of the case. The papers in the book concentrate on historical and methodological issues of thought experiments providing a range of analyses and thus clarifying a concept that has been hard to pin down. The book is divided into three sections, each containing three papers. p>
The first section looks at thought experiments through the lens of history. Ierodiakonou provides an analysis of the first recorded ancient thought experiments, that of a man standing at the edge of the universe, extending a hand or stick to prove the infinite nature of the universe. She compares different versions of this thought experiment and uses them for a tentative classification of thought experiments. Lauter looks at a specific set of ancient thought experiments presented in the De anima Neoplatonic commentaries. He concludes that an analysis of thought experiments needs to be sensitive to the domain in which they are applied with for example psychology and ethics employing much weaker counterfactual claims than physics and metaphysics. Gellard's paper jetsets forward in time to the medieval period in which thought experimentation is rife in philosophy. Gellard focuses on the debate on atomism.The second section looks at the possibility of thought experimentation. After all, experimentation is usually connected to the physical world and not just thought, so the possibility of thought experimentation has always been rather contentious. Knuuttila and Kukkonen analyze the status of thought experiments with impossible premises. Palmerino looks at Galileo's thought experiments. Despite their origin in medieval and Renaissance works she argues that Galileo breaks with tradition in that he uses the traditional method of argument but uses it to come radically new conclusions. He particularly breaks with medieval tradition by rejecting thought experiments that relied in any way on God. Virvidakis provides an analysis of thought experiments from a Kantian transcendental perspective, thus providing a new perspective to the traditional platonic or empirical interpretations of thought experimentation that can be found in the debate between Brown (2004) and Norton (2004).
The final section is on how thought experiments work. Engel discusses the relationship between the modal claims in thought experiments and their reduction to ordinary counterfactuals. Goffi and Roux analyse why some thought experiments work and others do not. They provide three conditions that a successful thought experiment needs to adhere to and substantiate their analysis with historic examples.
The most interesting paper in this volume for the simulation community is the final paper. Zeimbekis compares thought experiments to mental simulations. The initial claim is that during a thought experiment the mind simulates processes to reach conclusions. Zeimbekis distinguishes between mental simulations of mental processes and of physical processes. His analysis of the mental simulation of mental processes focuses on moral thought experiments and concludes that mental simulation on morality leads to an undesirable bias towards individualist moral theories. For physical processes he concludes that thought experiments cannot be seen as mental simulations as the simulacrum is essentially different from the target system. There is only the representation of a perception of a physical state followed by an inductive application of background knowledge as if we were to following this development of this state in vivo.
Where does this leave the simulation of social phenomena? The differentiation between the source and target processes of simulations is one the agent-based community should keep in mind when analyzing models and is discussed for example in David et al. (2004). In addition, the conclusion that mental process/mental process simulations can lead to an individualist bias could open an interesting debate on assumptions of individualism in current simulations.
As stated above, whether thought experimentation is a valid method in science and/or philosophy will not be resolved by this collection, but the perspective offered will refine the debate considerably and will provide a useful prism through which to assess agent-based modelling.
ReferencesDAVID, N, Marietto, MB, Sichman, JS and Coelho, H (2004) The Structure and Logic of Interdisciplinary Research in Agent-Based Social Simulation. Journal of Artificial Societies and Social Simulation, 7(3): https://www.jasss.org/7/3/4.html
NORTON,J (2004) 'Why thought experiments do not transcend empiricism'. In Christopher Hitchcock (Ed), Contemporary Debates in the Philosophy of Science. Malden: Blackwell, pp. 44-66
Return to Contents
of this issue