Shall I compare thee? Simile understanding and semantic categories

David Fishelov*

*Corresponding author for this work

Research output: Contribution to journalReview articlepeer-review

13 Scopus citations


The purpose of this article is to present a few research findings on the process of understanding simile, where special attention was given to the semantic relationship between its three major components: Tenor (hereafter T), Vehicle (hereafter V) and Predicate (hereafter P). By manipulating different degrees of conventionality among these three components and the explicit presence of a P, eight kinds of simile emerged (e.g. conventional T and V but with unconventional P). Questionnaires were formulated consisting of lists of sixteen similes, representing the eight kinds. Subjects were asked to provide a short interpretation of the given similes, to estimate whether the simile conveyed negative, positive or neutral connotations, and finally to grade the degree of difficulty they encountered in understanding it. Thus, for example, most subjects reported that (1) "John is like a snake" says that John is a cunning and dangerous person, that it conveys negative connotations and that they had no difficulty in understanding it. Other, more "difficult" similes got diversified answers. Thus, for example, some subjects claimed that they had difficulty in understanding a phrase like (2) "John is like the state of Israel," and responses concerning its connotations and its specific meanings were more heterogeneous. One conclusion based on the results is that subjects tend to cling to existing semantic categories not only in understanding conventional similes (and, by implication, metaphors and symbols), but also when faced with highly novel ones.

Original languageAmerican English
Pages (from-to)71-87
Number of pages17
JournalJournal of Literary Semantics
Issue number1
StatePublished - Apr 2007


Dive into the research topics of 'Shall I compare thee? Simile understanding and semantic categories'. Together they form a unique fingerprint.

Cite this