Summary: In this paper, Binkley takes on a set of assumptions about the nature of metaphor, namely, that metaphors taken at face value are either false or meaningless, and that the linguistic norm is literal. More specifically, he argues "(a) that metaphors need not involve false or nonsensical uses of language, (b) that metaphors can be true, i.e., that they can be used to state true propositions, and (c) that the truth-value of a metaphorical claim can be discerned in fundamentally the same way we discern the truth-value of a literal one" (171).
Why would someone suggest that "Richard is a fox" is false? The reason is apparently the fact that Richard is not a fox, but a man. But as Binkley notes, "Groups of words which can be used as contradictories will be contradictory only if they are used with the same sense (but for the negation) and in the same context" (172). Given that "Richard is a fox" is metaphorical and "Richard is not a fox" is literal, they do not contradict. The view that the two contradict can only occur if we confuse "Richard is a fox" taken metaphorically with "Richard is a fox" taken literally; but although they use the same words, they do not express the same proposition. Although these literal and metaphorical uses of the word "fox" are related, they are different uses.
To claim that "Richard is a fox" taken metaphorically is false might be taken to contradict the fact that Richard is not literally a fox; but again, the falsity of "Richard is a fox" does not follow from that fact. Then Brinkley notes:
One of the most widespread myths about metaphors is the idea that at the literal level they harbor an impropriety of language....The notion that the literal reading of a metaphor is false or nonsensical is too familiar to need thorough documentation. Yet there appear to be simple counter-examples to this almost universally held belief. Consider the trite "He lives in a glass house." This statement may, but need not, involve a strain in language, even at the literal level....We know that the statement is metaphorical not because we know the person in question lives in a brick house (we may have no information about where he lives), but because the conversation has been about his behavior and not about his residence. (173)
There is, in fact, no good reason to think that metaphors involve any sort of linguistic impropriety.
So we haven't found a contradictory to "Richard is a fox" yet. How about "Richard is not a fox" taken metaphorically? In a conversation, if you said Richard is a fox, I might protest, saying Richard is not a fox (his apparent cunning might really just be a lucky blunder). In such a case, Richard either is or is not a fox; we are disputing over the truth of what is said, in much the same way we might dispute over whether Richard is a good husband or a buffoon. "In all these cases, a claim can be made which is amenable to argument, which has more or less determinate criteria of evaluation, which can be supported and weakened with evidence, and so on" (174). There's some vagueness in each of these cases, but in all of them we all know more or less what is meant. To this Binkley adds another argument:
Furthermore, whether a metaphor can be true does not depend upon its age. A dead metaphor will act almost as though it were literal, and will consequently raise no special problems about truth. But a fresh metaphor is no less capable of stating truths if its author wants to make claims with it. (174)
Someone might argue, however, that metaphorical claims are true in a less direct way than literal claims, i.e., that they have truth only insofar as they are connected with literal claims, i.e., insofar as they are just literal statements embellished. It is true that metaphor depends on the literal in that we need to know the literal use of the words to recognize the metaphor; but this in itself is a limited sort of dependence, and doesn't require us to say that metaphorical language is parasitic on literal language simply speaking. We need to understand the literal use of the words to grasp the metaphor, "But once metaphorical meaning is secured, the words and the meaning are not mediated by a third term, the literal translation of the metaphor" (175). The dependence of metaphorical claims on literal claims has nothing to do with their truth or falsehood. Likewise, there doesn't seem to be any reason to demand a mediation by a literal paraphrase of the metaphor (e.g., "Richard is sly"). He notes as well that sometimes to clarify literal claims we have to resort to metaphorical paraphrases.
It is commonly thought that literal language is more precise than metaphorical language. It is true that literal language is very helpful for clarification; "however this is no reason to presume that it is any more precise in expressing meaning or any closer to 'true' meaning than the metaphorical" (174). As he notes, sometimes we mean precisely something that is vague. The literal may be more precise with respect to certain endeavors, e.g., clarification; but this does not make it a more exact expression of meaning.
Someone might hold, however, that the literal is still a more exact way of expressing reality; i.e., that "Richard is cunning" gives a more accurate picture than "Richard is a fox." This, however, need not affect their truth or falsehood at all: "Descriptive power may be affected by the precision of an expression, but truth-value is not" (177). This is true even of literal statements, which may be imprecise but still true or false. There are, likewise, criteria for the metaphorical use of statements just as there are criteria for the literal use of statements.
Binkley then goes on to identify one of the reasons why philosophers of language have so easily misled themselves on the subject.
When someone wants to put forth an example of a use of language he exhibits an expression out of context....
Because of the nature of linguistic example-giving, it will appear as though those expressions whose meanings are (on the average) less context-dependent or less in need of explication will be the most perspicuous conveyors of meaning. The examples of language which are most readily exhibited and most easily understood in the context of example-giving will be those which are least context-dependent for their meanings. Accordingly, those expressions which will appear to embody their meanings most limpidly when they are exhibited in philosophical discourse will be explicit literal sentences whose meanings are not highly sensitive to changes in context and do not rest heavily upon circumstances of their use. (178)
(Another way to put this, to use a phrase Binkley does not use, is that philosophers of language have been led by the observer-selection bias of their choice of examples.) Sentences that convey their meanings more clearly and directly as examples will not necessarily do so in actual use (and vice versa).
Three distinctions, if used appropriately, clarify how metaphors can be true or false.
a)Establishing the truth of an expression and establishing the meaning of an expression are two different activities.
b) Metaphor as a resource of language should be distinguished from the various uses to which metaphor can be put. Metaphors can be used as poetic devices; this does not mean they are exclusively so, since they can be used to state facts. "A 'literal claim' is nothing other than a claim made with literal language. Literal truth is not a kind of truth, but a truth expressed in literal language" (179) (I consider this the most important statement in the paper.)
c) We should avoid confusing the meaning of a metaphor (or any other expression) with its explication.
Evaluation: What can I say? I agree with every point Binkley makes; and most of them I had come to on my own through similar arguments before I came across this paper. (I came across it just half an hour ago when I was taking a break; and was so excited on finding this generally unread, wonderfully correct paper in a (relatively) obscure journal that I had to put it up on my weblog as LFPA.) As I've told someone before, it's a scary thing when the people who claim to be able to do philosophy of language can't even figure out that the phrase "the literal truth" involves a figurative use of the word "literal."