Elek Lane (UC Berkeley), "Metaphor and ambiguity"

Philosophical Studies

Metaphor: Simple Semantics, Complex Coordination

By Elek Lane

Suppose we are remarking that, no matter what we do, Bill just won’t go away and leave us alone—despite how we pointedly sigh, glare, even verbally abuse him, Bill keeps coming back. I utter (1), with which you wholeheartedly agree.

(1) Bill is a boomerang.

I’ve summed up Bill’s habit of returning to your side (no matter how hard you wing him), and noted that it’s just in his nature to do so. And I’ve probably made you contemplate a funny image. It’s hard to describe it exactly, but suffice it to say that it involves imagining Bill to be what he manifestly is not. In short, you see him as a boomerang.

This illustrates two of the many things that we do with metaphors. First, they often cause us to gain new beliefs. Consider how after uttering (1), I might also utter (2)

(2) Oh and so is Charlie.

helping you to thereby learn something about Charlie. Second, metaphor is associated with a certain kind of imagination, perspective-taking, or what-have-you—a psychological state that I’ll just call “seeing-as”, and in which theorists of metaphor have long been interested (Black 1955, Davidson 1978, Moran 1989).

Before I get around to explaining my own view—spelled out in “Metaphor and Ambiguity” (Lane 2024b, accessible here)—let’s take the lay of the land and discuss a few choice-points for the theory of metaphor. Famously (and infamously, for some) Donald Davidson (1978) articulated a view now called “non-cognitivism”. (Non-cognitivism has been more recently defended by Ernie Lepore and Matthew Stone 2015.) For the non-cognitivist, a metaphorical utterance achieves its effects simply by inducing a certain kind of imaginative state—the state of “seeing-as”—in audiences. The juxtaposition of a metaphor’s primary and secondary topics causes audiences to see the former as the latter. And this may lead an audience to notice many things about the primary topic. But being caused to notice something is not being told that thing, and the non-cognitivist maintains that metaphorical utterances have only their literal meanings—that there simply is no such thing as metaphorical meaning or content.

Cognitivism, on the other hand, is the view that metaphors do bear some kind of extra-literal meaning and/or content. While different versions vary widely on how that information is represented—and even on the kind of information that is—they are united by a shared commitment to thinking that semantics and/or pragmatics has something proprietary to say about how metaphorical utterances can be used to transfer information.

I defend a cognitivist view of metaphor. It can be stated very simply: words are ambiguous between their literal and metaphorical interpretations. (Black 1955, Goodman 1979, and Skulsky 1986 may be precursors.) So the occurrence of “boomerang” in (1) is just a different lexical item than its occurrence in (6).

(1) Bill is a boomerang.

(6) I whacked the kangaroo with a boomerang.

From a semantic perspective, then, there is no real difference between the ambiguity displayed by words such as “duck”, “cape”, “jam”, “bug”, and so forth, in English and the ambiguity displayed by “boomerang” in the context of (1).

I don’t think that these literal-metaphorical ambiguities exist in conventional English, of course. They exist in conversation-specific “microlanguages” (Ludlow 2014). Here’s one (non-metaphorical) way in which a conversation-specific microlanguage may differ from conventional English: Suppose we stipulate “sconch” to mean small quantity of whiskey. The microlanguage we speak will then include one extra lexical item—it will be an “expansion” of conventional English (Armstrong 2016)—and we will be able to make sense of, and to, each other when we utter sentences such as (7) and (8).

(7) I like a sconch in the evening.

(8) Pour me a sconch?

We do this kind of thing all the time. There is nothing particularly mysterious about our ability to expand our vocabularies beyond the standard conventions of English.

Metaphor, I think, is essentially a more subtle technique for introducing new lexical items to one’s vocabulary. When a metaphor is uttered, the speaker repurposes old words, outfitting them with new meanings, and challenges her audience to figure out the microlanguage that will unlock her hidden message. Naturally, speakers must leave clues for their audiences (otherwise the game of metaphor would be neither useful nor fun); and I take it that one of the main clues is just the standard conventions of English—in other words, that what a word means in conventional English is an important clue that interpreters use in figuring out the meaning of its metaphorical homophone in the new vernacular.

Why accept such a view? For one, it accounts for data that non-cognitivism struggles to explain—for instance, the fact that it is felicitous to embed metaphors in the complement-clauses of attitude reports (which creates the appearance that these clauses determine non-literal contents), as in (9) and (10).

(9) Anita thinks that Bill is a boomerang.

(10) Anita wishes Bill would stop being a boomerang.

And for two, it neatly handles a number of linguistic puzzles in a way that sets it apart from other cognitivist theories of metaphor. The details are in the paper, but I take an ambiguity theory to be distinct in its ability to explain “direct rebuke” (or the fact that you can disagree with the point made by a metaphorical utterance by simply negating the uttered sentence), why toggling between literal and metaphorical readings of one and the same token word produces infelicity, why semantic entailments that are valid under literal interpretation often fail under metaphorical interpretation, and why metaphorical speech is in general productive (or why metaphorical meanings appear to provide inputs to semantic composition).

An ambiguity theory of metaphor is really a theory of what I call the “linguistic side” of metaphor, i.e., about the semantic and pragmatic profiles of metaphorical utterances. It focuses on where to locate metaphorical meaning and content in the overall analysis of the speech act—as Elisabeth Camp (2006: 166) puts it, it’s a theory of the “taxonomic status” of metaphorical meaning. But a global theory of metaphor—one that completely explains the phenomenon—will also have to say something about the cognitive side of the story (that is, about the cognitive capacities that support the production and comprehension of metaphorical speech). While an ambiguity theory is compatible with a number of different hypotheses about the cognition of metaphor, I argue elsewhere (Lane 2024a, accessible here) that metaphor cognition draws not only on theory of mind, but also especially on an ability to make-believe, imagine, or see one thing as another (I support this argument by considering a selective deficit present in some speakers with autism spectrum condition). Thus while “seeing-as” doesn’t show up in the semantics or pragmatics of metaphor on my account, its connection with metaphorical speech is explained by the hypothesis that “seeing-as” is characteristically involved in the cognition of metaphor.

One contribution of this work is its distinction between the linguistic and cognitive sides of a theory of metaphor—and relatedly, what we can expect to be explained in terms of the contents represented by metaphorical utterances (on the one hand) and what is better explained by the cognitive processing of metaphor (on the other). Some theorists prefer to explain metaphor’s relationship to “seeing-as” in terms of the content represented by metaphorical utterances. One example is Megan Stotts. Stotts (2021: 113) argues that some metaphorical assertions are actually “indirect directives that command the hearer to engage in an open-ended comparison”. Metaphor’s relationship to “seeing-as” is thus supposed to be explained by the contents of metaphorical utterances—contents which determine a “seeing-as-ish” satisfaction condition. Another example can be found in Camp (2017) where it is argued that metaphors presuppose perspectival content (i.e., “seeing-as-ish” content). I agree with both Stotts and Camp that metaphors often involve open-ended comparing/perspective-taking/seeing-as—but we get a cleaner explanation of the linguistic data by supposing this to be due to how metaphors are often (and maybe characteristically) processed, rather than supposing it to be due to metaphors’ contents.

So that’s a program for explaining metaphorical speech: keep the semantics simple, and posit that words are simply ambiguous between their literal and metaphorical interpretations—but allow for complexity in speaker-hearer coordination on these novel ambiguities, and for “seeing-as” to be characteristically involved in the cognition of novel metaphors.

References

Armstrong, J. (2016). The problem of Lexical innovation. Linguistics and Philosophy, 39, 87–118.

Black, M. (1955). Metaphor. Proceedings of the Aristotelian Society, 55 (1), 273–294.

Camp, Elisabeth (2006). Metaphor in the Mind: The Cognition of Metaphor. Philosophy Compass 1 (2): 154–170.

Camp, E. (2017). Why metaphors make good insults: Perspectives, presupposition, and pragmatics. Philosophical Studies, 174 (1), 47–64.

Davidson, D. (1978). What metaphors mean. Critical Inquiry, 5 (1), 31–47.

Goodman, N. (1979). Metaphor as moonlighting. Critical Inquiry, 6 (1), 125–130.

Lane, Elek (2024a). Distinguishing two kinds of fictionalism: metaphor, autism, and the imagination. Synthese 204 (128): 1–23.

Lane, Elek (2024b). Metaphor and ambiguity. Philosophical Studies 181 (11): 3059–3087.

Ludlow, Peter (2014). Living Words: Meaning Underdetermination and the Dynamic Lexicon. Oxford, GB: Oxford University Press.

Lepore, E., & Stone, M. (2015). Imagination and convention: Distinguishing grammar & inference in language. New York: Oxford University Press.

Moran, R. (1989). Seeing and believing: Metaphor, image, and force. Critical Inquiry, 16 (1), 87–112.

Skulsky, H. (1986). Metaphorese. Noûs, 20 (3), 351–369.

Stotts, M. (2021). Demystifying metaphor: a strategy for literal paraphrase. Philosophical Studies, 178, 113–132.