The psychological relationship between metaphor and analogy has been the focus of continuing controversy. Using both production and comprehension paradigms, we investigated the processing of metaphors versus formally equivalent analogies. To precisely match the metaphor and analogy tasks, we generated materials based on Aristotelian metaphors, derived by translating proportional verbal analogies in the form A:B::C:D into metaphors in the form "A is the C of B" (omitting the D term). For both analogies and metaphors, we had participants rate their goodness (Experiment 1), produce expressions (Experiments 2a, 2b, and 2c), and select the best completion (Experiment 3). People assigned higher goodness scores to analogies and metaphors that were relationally valid, and the two expression types showed similar sensitivity to individual differences in both executive functions and verbal ability. But although a valid analogical basis proved necessary to create a successful metaphor, this was not sufficient. We found that good metaphors honor additional pragmatic constraints predictable from a set of lexical variables related to word generalness concreteness and familiarity. Good metaphors (but not analogies) describe topic concepts (denoted by A terms) in terms of their more general and familiar domain concepts (denoted by B terms) and more general vehicle concepts (denoted by C terms). Our findings demonstrate both formal commonalities and pragmatic differences in the processing of metaphors and analogies. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
Ichien et al. (Mon,) studied this question.