JOURNAL OF MEMORY AND LANGUAGE 35, 544–565 (1996) ARTICLE NO. 0029 Conceptual Metaphors and Figurative Language Interpretation: Food for Thought? MATTHEW S. MCGLONE Lafayette College How do people interpret metaphors such as The lecture was a three-course meal? Lakoff (1993) has proposed that figurative expressions are interpreted as instantiations of deep conceptual metaphors, such as IDEAS ARE FOOD. In contrast, Glucksberg (1991) has proposed that metaphors are interpreted as assertions of the topic’s (e.g., lecture) membership in an attributive category exemplified by the vehicle (e.g., three-course meal). Four experiments that test the predictions of the two views are reported. The results suggest that reference to a conceptual metaphor is not the modal strategy that people use when paraphrasing metaphors (Experiments 1 and 2), rating the similarity between metaphors (Experiment 3), or retrieving metaphors from memory (Experiment 4). In each of these situations, participants relied primarily on the stereotypi- cal properties of the vehicle concept. The results from these experiments are consistent with Glucksberg’s (1991) attributive categorization proposal.  1996 Academic Press, Inc. People use metaphors such as Our marriage way as literal comparisons, such as Nectarines was a rollercoaster ride in everyday dis- are like oranges. course, and they are easily understood by their Although attractive in their simplicity, com- addressees. How is this understanding accom- parison models fail for the important case in plished? Some theorists have argued that met- which the addressee is not aware of the rele- aphors are interpreted as implicit comparison vant properties that the topic and vehicle con- statements, rather than categorical assertions. cepts share. Consider once more the claim Our For example, Ortony (1979) and Gentner marriage was a rollercoaster ride. For people (1983; Wolff & Gentner, 1992) have proposed who are not familiar with the marriage in that metaphors of the form X is a Y are inter- question, there can be no a priori representa- preted as comparisons of the form X is like a tion of the marriage that includes properties Y. Once the implicit comparison is recognized, such as ‘‘exciting,’’ ‘‘scary,’’ or ‘‘unstable.’’ these theorists argue, the addressee then con- Yet these are exactly the sorts of properties ducts a search for matching properties in the that come to mind upon an uninformed read- topic (e.g., our marriage) and vehicle (e.g., ing of the statement. rollercoaster ride) concepts. The implication Comparison models are ill-equipped to deal of these ‘‘comparison models’’ is that meta- with any metaphor that is used to make infor- phors are understood in essentially the same mative assertions about a topic — i.e., to intro- duce properties that are not part of the address- ee’s mental representation of the topic. This This research was supported by a Public Health Service argument applies with equal force to many grant (HD25826) to Sam Glucksberg of Princeton Univer- literal comparisons. For example, if a person sity. I thank Sam for his support during all stages of this knows nothing about kumquats, then telling research. Thanks also to Susan Fussell, Boaz Keysar, Phil Johnson-Laird, Marcia Johnson, Deanna Manfredi, Lisa her that A kumquat is like an orange will intro- Torreano, and two anonymous reviewers for their com- duce new properties into her mental represen- ments on earlier versions of this paper. Address corre- tation of the concept ‘‘kumquat,’’ rather than spondence and reprint requests to Matthew S. McGlone, produce a match between ‘‘kumquat’’ and Department of Psychology, Lafayette College, Easton, PA 18042. E-mail: mcglonem@lafayette.edu. ‘‘orange’’ properties. 544 0749-596X/96 $18.00 Copyright  1996 by Academic Press, Inc. All rights of reproduction in any form reserved.