Methods Experts Use to Generate Analogies

Abstract

The first step in using a spontaneous analogy is to generate the analogous case. By “generation” here I mean accessing or constructing an analogous case and raising the question of whether there is a valid analogy relation between it and the target problem. This is often considered to be the most creative part and by some even an unconscious part of using an analogy, and therefore it may be the one which is least well understood. The classical view of analogy generation is that a related case is accessed in permanent memory by association and brought into conscious attention in working memory (Gentner, 1983; Holyoak and Thagard, 1989). The purpose of this chapter is to examine think-aloud evidence on spontaneous analogies generated by experts. I will present evidence that some analogies are indeed accessed by association but that many are generated via a transformation or via a principle and that the classical view misses certain methods for generating analogies, some of which are quite powerful. Thus this chapter focuses on subprocesses used for process P1 in Table 2.1 The first part of this chapter elaborates the definitions used in referring to analogies, and the criteria for recognizing them in transcripts in more detail than was possible in Chapter 2. A variety of analogies from the study of ten expert subjects solving the spring problem are then described. The number of significant analogies and examples of the different types of analogies generated, including several creative invented cases, are presented. Finally different methods of analogy generation are identified and discussed.