No CrossRef data available.
Published online by Cambridge University Press: 19 June 2023
A thought experiment requires neither instrumentation nor embodied actors. Nor does it appear to introduce new empirical information about the world in which it is performed (Kuhn 1962, p. 241). Nevertheless, it presents some previously un-recognized property of that world with a logical force that no real experiment can match. Thought experiments are easy to replicate—this confers an important advantage over real ones, though the real advantage is thought to be that there is no need to conduct a real experiment in place of a thought experiment From a rationalist position such as Jim Brown's (1991, 1993) thought experiments are either mysterious or eise explicable in terms of our ability to intuit natural kinds directly. Empiricists such as John Norton (1991) regard them as disguised deductive arguments whose performance-reconstituted as formal arguments—discloses in the form of prernises, information we must already have.
This is a shortened version of papers presented to the session on “Instrumentation and Experiment” at the joint BSHS-HSS-CSHPS meeting, Toronto (July 1992); to the BSPS annual conference at the University of Durham (September 1992) and to the session on ‘Thought Experimentation: the Theoretician's Laboratory” at the Thirteenth Biennial Meeting of the PSA, Chicago (October 1992). I have benefited from commentaries by Alfred Nordmann (at Toronto) and Ian Hacking (at Chicago), and from discussion with participants at these meetings. The support of a Special Project Grant from the Joint Research Councils Cognitive Science/HCI Initiative and ofthe British Academy (for travel to Chicago) is gratefully acknowledged.