This essay rests on at least four assumptions, none of which will be defended. I begin by setting out these four as clearly as I can, along with some explanation of why I offer no defense of them.
(1) There is no infallible sign that any given expression, spoken or written, is a metaphor. Until about twenty years ago it was commonly supposed that a necessary condition of something’s being a metaphor is that if taken literally it be absurd or self-contradictory or at least blatantly false. This would not have been a sufficient condition, of course, but it was thought to be necessary. It was my privilege to note that a metaphor need display no logical or semantic anomaly, and that indeed a metaphor taken literally need not even be false. 1 I promptly went on to the mistaken assertion that there must be some way in which metaphors are recognized, that every metaphorical expression must exhibit some anomaly if taken literally, although the oddity might be only pragmatic. I concluded that every metaphor taken literally must be something that, in the circumstances, the speaker or writer could not mean. Thus if the metaphor were a literal truth, like “Sydney is a warm city” or “No man is an island,” the literal truth would be so obvious in the circumstances that one could not suppose the author to have intended to communicate that truth, and one would be induced to look for another, metaphorical content. 2 That was a mistake.
I made this mistake because I took this for granted: “In understanding a metaphor, one must put together two or more elements in some novel way. The need for some novel construction is signaled by the impossibility of assembling the elements in the usual (literal) way.” And that is a compound mistake. It seems a neat congruence between the [End Page 223] mechanism of a metaphor and the signal that it is a metaphor, but it’s a mistake because there need be no signal. Generally, of course, one does find an obstacle to taking a metaphor literally, either within the sentence itself, or within the more complex entity which is the utterance of the sentence within a certain context (that is, within what Austin called “the total speech act in the total speech situation”); but there need be no obstacle. A sentence might support both literal and metaphorical understandings, and if so, one might just miss the metaphor. One finds the literal construction significant, and then does not go on to attempt a metaphorical construction even though a metaphor is there for the taking. Some people are better than others at understanding metaphors, and some people are better than others at finding them.
(2) There is a difference between expressions that are metaphorical and expressions that are not, that are literal. This is true despite the fact that there is no general reliable sign to indicate a metaphor, and also despite the fact that there may be no serviceable formula for defining “metaphor.”
Understanding a metaphor requires something more, something other than understanding a literal expression. Generally speaking, all who know a language can understand the language’s literal expressions. This understanding sometimes fails because of the syntactic or semantic difficulty of a given literal expression—for instance, because of the presence of unfamiliar words; but knowledge of a language is the capacity, potentially, to understand the language’s literal expressions. Those who know a language usually are able to understand metaphorical expressions in the language as well, but this understanding seems different. It requires wit, imagination, flexibility—a kind of initiative. And this initiating understanding is built on one’s capacity for literal understanding. One must be able to deal with an expression literally, and then go on to do more with it. Thus literal expressions, and all expressions regarded literally have a settled fixity—at least relative to metaphorical expressions and all expressions (that can be) regarded metaphorically.
The idea that there is no difference between the literal and the metaphorical is a foolish idea, especially when advanced as a consequence of the fact that all language...