Observed methods for generating analogies in scientific problem solving

Abstract

Evidence from videotapes of experts thinking aloud is presented which documents the spontaneous use of analogies in scientific problem solving. Four processes appear to be important in using an analogy: (1) generating the analogy; (2) establishing confidence in the validity of the analogy relation; (3) understanding the analogous case; and (4) applying findings to the original problem. This study concentrates on process (1). Evidence was found for three different methods of analogy generation: generation via a principle (1 case), generation via an association (8 cases), and generation via a transformation (18 cases). Although the mechanism underlying analogy generation is usually described as an association process, transformation processes, where the subject modifies or transforms some aspect of the original problem, may be just as important if not more important. In contrast to the usual view of an analogous case as already residing in memory, several of the analogous cases were quite novel, indicating that they were newly invented Gedanken experiments. The usefulness of some analogies appears to lie in a “provocative” function of activating additional knowledge schemas that is different from the commonly cited “direct transfer” function where established knowledge is transferred fairly directly from the analogous to the original case.