Psicodebate (Feb 2016)
Analogy, an Alternative Model. Critics to the standard model of analogical problems solving and proposals for an alternative one
Abstract
The authors made an extension of Hofstadter‘s criticisms against the standard approach in analogical thinking represented by the structure-mapping theory of Gentner and the multiconstraint theory of Holyoak and Thagard. Based on this extension, they proposed a non-serial model of analogical problem solving. Against the standard approach, the model postulates that: (a) people detect and evaluate differences between mapped elements before the subprocess of inference generation and consider them in order to control it, and (b) properties of an element that explain why the element could fill a certain role in the base problem resolution (PERs) play a crucial role in these detection and evaluation operations, and also in post-inferences subprocesses. An experiment showed that: (a) people detect and evaluate the relevance of differences between mapped elements before inference generation, (b) that they inhibit the generation of literal inferences when they face relevant differences, and (c) that they stop the subprocess when they recognize insuperable ones. The results also showed that base PERs are reactivated at different moments of analogical transfer. The data obtained are incompatible with the standard theories of analogical thinking, which treat inference generation as a syntactic mechanism and exclude contextual semantic analysis from the study of analogy.
Keywords