The discovery of the urea cycle: Computer models of scientific discovery

Abstract

The formation of scientific theories is a paradigmatic case of creative problem solving. Any historical account seeking to go beyond the purely descriptive chronology of scientific development attempts to provide some sort of historical explanation. The explanation of the development of science is a challenge to any approach to science studies. Computer modelling such processes places additional demands both on the theoretical accounts and on the technical representation through computer models. The explanation is causal by nature — certain historical events that allegedly contribute causally to the specific course of events are identified. The relevance of events for a historical process can only be established by some sort of difference test: Without the presence of a contributing event, the historical process would have been different for otherwise identical relevant circumstances. Such a test goes beyond the level of historical narratives and the description of what actually happened. It must involve some sort of reasoning about counterfactual historical scenarios.