Reversibility, Aptness, and the Conventionality of Metaphors and Similes

Abstract

Abstract Metaphors and similes relate a topic (e.g., "crime") and a vehicle ("disease"). As sentences, metaphors (e.g., "crime is a disease") have the same form as literal claims about a category and similes (e.g., "crime is like a disease") the form of a comparison. However, a traditional argument is these differences between metaphors and similes are superficial, and the 2 assert the same relation between the topic and the vehicle. We test this argument using "reversibility" (e.g., "crime-disease" vs. "disease-crime"). Experiment 1 found reversing topic-vehicle pairs lowered the comprehensibility of metaphors more than similes (indeed metaphors were more often judged uninterpretable). Also, reversed metaphors were less likely to preserve their original interpretation and were more often converted back into the order in which they are typically expressed. Experiment 2 found preference for the metaphor form was related to the aptness of the topic-vehicle pairing (i.e., whether it captures the topic's important features), not to the conventionality of the vehicle. The results support the view that metaphors and similes are distinct statements, metaphors functioning more like categorization statements and similes more like comparison statements.