Studying Scientific Discovery by Computer Simulation.
Abstract
Scientific discovery is a complex process, but despite its complexity, it can be analyzed using two basic ideas from the theory of human problem solving - the notion of search through a problem space, and the notion of heuristics for diredting that search down profitable paths. In this paper we examine Black's discovery of the specific heat law in the light of these concepts, and consider how BACON, an empirical discovery program, employs heuristics for finding numeric relations, positing intrinsic properties, and expecting symmertical forms to rediscover this law. Finally, they examine the role of theory in scientific discovery, and describe a different approach to inducing Black's law that relies on the distinction between extensive and intensive terms, and on the notion of conservation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 30, 1983
- Accession Number
- ADA126438
Entities
People
- Gary L. Bradshaw
- Herbert Simon
- Patrick W. Langley
Organizations
- Carnegie Mellon University