Ernie Lepore and Matthew Stone
ECO, METAPHOR, AND INTERPRETATION: A CURE FOR THE COMMON CODE
As an ornament, the metaphor is of no interest to us, because, if it says more pleasantly that which can be said otherwise, then it could be explained wholly within the scope of a semantics of denotation. We are interested in the metaphor as an additive, not a substitutive, instrument of knowledge.
—Umberto Eco
Historians of ideas might disagree among themselves about whether the standard model of communication is already present in Plato, but no one would challenge its grip on the philosophical imagination ever since Locke. It is the model in which a “message is decoded on the basis of a Code shared by both the virtual poles of the chain.”1 More colloquially, it is the model that if someone has an idea he wants to convey to an audience, then he should encode it in a shared language. They can retrieve the message. Eco long ago argued that this model fails adequately to describe the actual functioning of communicative intercourses.2 His challenge has had little impact on the analytic philosophical tradition’s fixation on Locke’s model. This is unfortunate since he is obviously right and they are obviously wrong.
Here is a simple but definitive Eco-style argument against this standard communication model (SCM): it is easy to imagine, indeed, actual cases exist, where two speakers, though equally proficient with a shared language, still cannot communicate, not because of a performance defect, but, rather, because one only writes and reads the language while the other only speaks it. Many of us share this particular relationship with the ancient Romans. We can read and write Latin even if we never learned to speak it and, of course, in the past it was spoken by many who never wrote it.
This simple thought experiment reveals that communication requires its participants to share more than a common code; they must also share a system of articulation. It also puts into focus the need to keep separate the world from its conception, its conception from language used to express it, and linguistic expressions we use to characterize the world from the vehicles we exploit in presenting these linguistic items. We are unaware of a single acknowledgment of this four-part taxonomy in the analytic tradition. Diagrammatically it is as in Figure 1:
Figure 1
The reason for its neglect is pretty straightforward: most philosophers (and linguists) hold a rather naïve view about the nature of expressions, that is, about the metaphysics of linguistic items. The philosopher Donald Davidson, for example, held the view (one that we believe remains the status quo among linguists and analytic philosophers) that linguistic items should be identified with shapes. Eco has been transparently clear about the subtleties of coding and sign production for decades.3 So it is quite surprising, indeed scandalous, how long this tiresome view has stuck around; it is egregiously wrong.
One wonders which shape? Is the word “red” to be identified with its pronunciation (which one?) or its orthography? Or is it to be assumed that these are the same shape? That is a stretch, is it not? Among linguists, at least those in the generative tradition, the only relevant shape for language individuation is phonological. We find this view equally unpalatable. Should we conclude that there are as many words as there are phonological realizations? That is, do the British have a different word for “tissue” than Americans since they pronounce it differently? Moreover, this view has a hard time accommodating the intuitive temporal and modal profiles of words. Might not “red” have been pronounced differently than it actually is? Is it not intuitive that the word “red” has, as a matter of fact, changed its pronunciation across time? But given the identification of a sound and a word this ought not to be possible.4
What these considerations minimally point to is the incompleteness of SCM. It must be supplemented by a story that tells us how to separate language from its various vehicles of articulation. But the problems for SCM run deeper and we owe a debt of gratitude to Umberto Eco for his excavations here. For example, his sustained discussion of metaphor, and figurative speech in general, exposes serious flaws within SCM. Consider the metaphorical:
(1) The girl is a birch.
Were someone to assert it, what would its (or a) correct interpretation be? Aristotle told us it should be interpreted as a (disguised) simile, as in, “The girl is like a birch.” But this is too weak. As Nelson Goodman reminded us long ago, everything is like everything in some regard or other. To be told merely that the girl is like a birch is rather uninformative in a way that the metaphor is not. In the other direction, some have suggested that all there is to its meaning is its literal meaning, and so (1) read either literally or metaphorically always means or says no more than that the girl is a birch. However, were someone to use this sentence, her audience would surely infer that she must mean something other than its literal meaning since otherwise she would be shirking her responsibility as a participant in a conversation to speak only truths, or at least not obvious falsehoods. Given this constraint on rational agents participating in a conversational collaboration, an audience will ask itself what the speaker might be trying to convey to us beyond literal meaning.
The palpable problem with such an account is its inability to tell us anything of value about how we can come up with a correct interpretation. As Eco himself notes, much the same could be said of the failure of formal semantics to explain metaphor. Reacting to the compositional semantic formal treatments that seek a “transfer” of properties from the literal extension to the metaphorical one—whereby, for example, with (1), the girl acquires the property of being vegetal and the tree of being human—not only predicts the wrong results but does not tell us how we came up with them in the first place; much the same could be said about other competing treatments.
There is yet a deeper concern that occupies and consternates Eco, namely, that good metaphors are—to borrow his elegant expression—open, by which he means there is no obvious closed message or interpretation we must take away from a good metaphor. But, contrary to Eco, and this will be our focus, we worry that his being right about their openness undermines any project that seeks their interpretation in the first place. To see why will take a bit of stage setting.
I. THE PUBLICITY OF MEANING
We take it as a given that interpretations must be public. The whole point of uttering something with the intention of getting information across to an audience is that what the speaker intends to communicate should be available to his audience. This part of the SCM is correct. Proponents of publicity include a who’s who of twentieth-century theorists of language, including Frege, Wittgenstein, Quine, Davidson, Dummett, Grice, Lewis, Derrida, and, of course, Eco.5 It’s hard to imagine what its challenge could be short of mysticism. We presume that anyone pressing a nonpublic account of interpretation has lost his or her way.
In the standard Gricean framework this constraint takes the form of a requirement on what counts as speaker-meaning; it requires that the speaker intend for the audience to be able to retrieve his or her communicative intention. The details of the account matter little here, if for no other reason than that most commentators reject the original Gricean details (including Grice, who revised the account often and wound up with something so oblique and scholastic you would need a Ph.D. in psychology to unravel it); yet virtually everyone embraces his basic idea.6 But in order for the speaker to reasonably expect his audience to retrieve his communicative intention, he must render his intention transparent to them; there should be publicly available cues to guide them. If he means his words literally, it’s pretty easy to see how the story would go; the speaker presumes his audience shares his linguistic conventions, and so, with a literal utterance he presumes they’ll recognize that what he means is what his words (literally) mean (as determined by their shared conventions in the context in which they were used). Of course the process is subtle; presumably something about the context indicates which language is being spoken, that they were spoken literally, and so forth. And when this intent parts ways from the literal, the speaker thereby carries a duty to get across that he is so doing, as well as one to render transparent to his audience the ways in which he is parting ways from the literal. His utterance might be so absurd that his audience can’t take him to be speaking literally without assuming he’s gone mad; or it might be so false that they assume the speaker would have to take them for ignoramuses to presume that he means his words literally. And so it goes. Theories disagree with, and diverge over, the processes behind this sort of interpretation but they agree about the end game—to retrieve the speaker’s communicative intention.
There are linguistic exceptions; presumably no matter how much we try we can’t make our use of the first-person pronoun “I” pick out someone else even if that’s our speaker intention; even if a madman thinks he’s someone else, what he says when he utters “I’m Barack Obama” is false unless he’s Barack Obama. Other sorts of contextual constraints work in a like fashion. But, by and large, the model is clear enough. It, of course, allows for various infelicities.
In short, the effects of nonliteral uses of language cannot be characterized in terms of an audience’s retrieval and uptake of a single (or several) publically available specific proposition(s) that the speaker intends to communicate. This is because the interpretive practices that govern nonliteral uses of language are joint (or collaborative) activities; characteristically, with figurative speech, a speaker invites the hearer to share in whatever insights either one of them can discover in exploring some particular imaginative world in some particular novel way. Further, nothing in their joint activity requires that the resultant effects contribute propositional content; and further, nothing privileges just those effects (if any) that the speaker alone had in mind, that is, as determined by the speaker’s communicative intentions.
We further believe, though Professor Eco might not concur, that his significant contributions on the open text support our idea. Where we are unclear about his position, and where we hope that he will illuminate us in his reply, is why he continues to call the results of working with an open text an interpretation (or coming to “understand a metaphor”) since, as we assume is (or will be) transparent, we think those results issue in nothing like interpretation, if by that is meant ascribing an interpretation—semantic or pragmatic.
We will proceed, then, as follows: we will provide evidence for the claim that with respect to figurative speech, in particular metaphor, we exploit very different sorts of strategies than what normally underwrites interpretation. We will stress, à la Eco, the generality and open-endedness of these practices. They apply across extended discourse and even across speakers. Interlocutors can explore and amplify them, enriching what they extract from the exchange, but they do not clarify, disambiguate, or delimit what they extract in order to achieve a single precise meaning. We obviously will have to show how we think these critiques relate to Eco’s own discussion of the open text. We’ll end our discussion with our puzzle for Professor Eco about interpretation.
II. THE SCOPE OF METAPHORICAL INTERPRETATION
What is the point of a metaphor? What is a metaphorist trying to get across to her audience when she constructs and presents a metaphor? No one can deny that at a minimum metaphorical speech invites its audience to imagine something as something else entirely. Eco uses example (1) above to help clarify the creative engagement that an audience brings to this process.7 To appreciate the speaker’s point in using this particular metaphor, it’s hard to deny that the addressee must place some important properties and relationships in a source domain in correspondence with those of a target domain. This imaginative effort allows similarities and analogies across domains to deepen and enrich an understanding of the target. In the metaphorical interpretation of (1), for example, the task is to line up certain key features of a birch to features of the girl that are similar in some noteworthy respect. We might entertain the way she, like a birch, bends under social forces. This is a different sort of process from literally imagining the girl to be a tree. In grasping the point of the metaphor, we do not, and indeed should not, imagine the girl to actually lose her humanity. At best, we are supposed to find correspondences to aspects of a birch tree.
In short, when you reason about a presented metaphor, you go ahead and explore real-world analogies in a distinctive way peculiar to metaphor. Exploring a metaphor is a distinctive cognitive process. Interpreting irony, another case of figurative speech, is a very different sort of process where an interpreter is not reasoning about the actual speaker, but, rather, reconstructing an imaginary speaker; it should be obvious that the processes of reasoning about an actual interlocutor and creatively engaging with a fictional character are distinct, requiring different sorts of reasoning.
When we think of the creativity of metaphor, we like to cite a particular metaphor from the humorist Matt Groening, who wrote:
(2) Love is a snowmobile racing across the tundra; and then suddenly it flips over, pinning you underneath. At night, the ice weasels come.8
We, his audience, know that a snowmobile is a sport vehicle like a car on skis that you drive through the snow; it’s fun; it’s exhilarating, and it gives you a sense of adventure. We know the tundra is a frozen landscape without trees; it’s a place of relative safety. We know that weasels are small predatory animals known for their fierceness and trickery. Putting all this together, then, it is not unlikely that we will come to imagine a kind of prototypical course for a love affair, which starts with a sense of adventure and excitement and then goes horribly wrong, leaving us with gnawing feelings of torture and pain. But how did we come to this particular interpretation? After all, we all know that love is not really a snowmobile.
Our view is that when audiences reason about metaphor what they are doing is exploring real-world analogies in a distinctive way peculiar to metaphorical speech in particular. Exploring a metaphor is a distinctive cognitive process. Sarcasm and irony are different sorts of figurative speech, and require, we claim, invoking different sorts of cognitive processes.
We can understand (2) even though it is new and unfamiliar. But we wouldn’t want to posit a convention, a preestablished meaning, for any of its words or constructions. It really does seem that what’s doing the work here is an ability to understand Groening’s sentence as described, and then an ability to go on to make an analogy between the experience of being in love and a certain kind of history that could happen.
Since metaphors are creative and not conventional,9 we need to explain how it is possible that we could get anything like a special meaning from them. One Gricean tradition, as mentioned in passing above, provides a way: the idea is, very simply, that a metaphor is an utterance where what the speaker means in using the metaphor differs from what the metaphor literally means. Literal meaning might tell the audience something about snowmobiles, whereas what the speaker means tells the audience something about the trajectory of love affairs. The first theorist to flesh out this sort of theory of metaphor in any detail was John Searle.10
Searle’s idea, applied to the current case, is that when Groening utters, “Love is a snowmobile” he doesn’t mean (despite what this sentence literally means) that love is a snowmobile. Instead, what he means is that love is a snowmobile*, where snowmobile* introduces a brand new concept or property, related to being a snowmobile, which has something in common with real snowmobiles and what they are like to ride, or to work with.11 The speaker’s communicative intention in making this utterance is not therefore to talk about snowmobiles, but, rather, to talk about snowmobiles*. His audience recognizes as much because there are principles of metaphorical interpretation that are common ground between the speaker and the audience; and these principles provide hints about how to replace one property with an associated one. Only by tying one of these associated properties to what the speaker means by his utterance is the audience able to recover a message that makes sense, that is, one that is appropriate for the conversation.
Searle’s fundamental claim is that when confronting a metaphorical expression the audience must first decode its linguistic properties, mapping it onto a literal form of representation that preserves the literal denotations and compositional structure of the constituents of the expression. This stage is succeeded by a realization of its contextual inappropriateness (for example, its literal interpretation may be somehow defective), followed by the suppression of this literal representation and the ensuing search for nonliteral forms of interpretation. It is only by going through this three-stage process, according to Searle, that we are able to identify what the speaker’s communicative intention was in uttering a metaphorical expression.
So Groening, with his metaphorical (2), could not be committing himself to the truth of the propositions that his words literally express. It is common knowledge that he knows that love is not a snowmobile. He is therefore overtly violating a maxim of conversation that you not say what you believe to be false. According to Searle, such an overt violation or flouting of this maxim indicates what the speaker’s intention is.12 With metaphor, this is a simile, based on what is said. In short, literal meaning is primary; and figurative meaning is associated with literal meaning in simple and systematic ways, derived in the pragmatic process of utterance comprehension, and this derivation is triggered by the fact that the literal meaning is an overt departure from conversational maxims.
An analysis similar in spirit to Searle’s is provided by the theory of relevance.13 Relevance theory posits a special meaning for the metaphor, “Love is a snowmobile,” namely, that love is a snowmobile*; where, again, snowmobile* is intended to introduce a novel concept that replaces the ordinary meaning of the word “snowmobile” in a way that captures its metaphorical use. The speaker’s intention is to use this metaphorical meaning, and again, it is supposed to be recognizable by appeal to general principles. But the mechanisms that these two theories invoke are different.
According to the principle of optimal relevance, we know that not everything that follows from being a snowmobile could possibly be true of love; some important things, however, true of snowmobiles are also true of love; and these importantly related properties define for us the concept of a snowmobile*. And so snowmobile* is intended to capture the relevant snowmobile inferences that the hearer is prompted to by the speaker’s metaphorical utterance. For relevance theorists, snowmobile* is, then, an ad hoc concept: something that is not specifically given by any of the conventions of language, but something that is recoverable in context because of the constraints imposed by a communicative intention; or imposed by the pull of relevance.
In short, there’s nothing special about interpreting a metaphor; its interpretation is just the usual recognition of communicative intentions through the same general psychological principles that are involved in all of the cases of communication. Interestingly, however, in the vast literature on metaphor, most scholars, including Eco, argue that metaphorical interpretation is special; so it will be useful to review some of their insights in order to take stock of this idea that interpreting metaphors involves pragmatic psychological (nonlinguistic) inferences based on what a speaker means by a use of the metaphor.
It seems there is something distinctive, poetic, and untranslatable about a good metaphor. Because of these features theorists have resisted attempts to give easy paraphrases of metaphors. As noted in passing earlier, Aristotle proposed that a metaphor is basically a disguised simile: “The simile is also a metaphor. The difference is but slight. When the poet says of Achilles that he ‘Leapt on the foe as a lion,’ this is a simile; when he says of him ‘the lion leapt,’ it is a metaphor. . . . [Similes] are to be employed just as metaphors are employed, since they are really the same thing except for the difference mentioned.”14
Aristotle’s view provides a kind of paraphrase that’s simple and formulaically related to the literal meaning of what you said. His view became a target for theorists who thought there was something special about metaphor, that is, something poetic and untranslatable. Max Black was one of the theorists who put these objections into print most influentially.15 He argued that the simile theory cannot be right. For one, the interpretation of a metaphor is much more specific than a mere comparison. If you say, “the girl is a birch,” you don’t just mean the girl is like a birch in unspecified or indefinite ways. Usually you want to convey that the girl is flexible, bends in certain ways (though, as Eco emphasizes, not in the same way as the tree).16 That specific set of commonalities with a birch simply doesn’t come across with a simile analysis.
But how should we think of the interpretive effort in metaphor to discover important similarities? Also, as noted earlier, philosophers and psychologists have long observed that the concept of similarity is unconstrained, variable, and open ended—even more so than the concept of opposition involved in interpreting sarcasm.17 We need only agree that a search for similarities cannot be a search for differences to see that metaphor is distinct from sarcasm. In fact metaphor is much more flexible. It involves not just finding things that resemble one another, but, as we suggested above, finding analogies that map objects, properties, and relationships in a source domain systematically onto corresponding ones in a target domain.18 This mapping can lead us to have new insights into the target domain, as the analogies lead us to make inferences about the target that we draw readily in the source domain but that we find too obscure or difficult to draw directly about the target. Or the mapping might simply shape our understanding of the target by imparting new attention or salience to aspects of the target that may be familiar but insufficiently appreciated.19
A more serious objection that Black ran against the simile account is based on the fact that metaphorical thinking gives special insights that can’t be paraphrased. When you say (1), you’re not just saying she is flexible. You are also thinking of her as a birch tree, and this imagery, this perspective you are taking on her is as important to the effect and the point of the metaphor as any propositional information it may convey.
Black called his positive account of metaphor “the interaction theory of metaphor.” Briefly explaining his view will help you to understand what’s at stake in explicating the perspective-taking that most theorists think is an essential part of metaphorical interpretation. Metaphorical interpretation starts with a tension and incompatibility between what we know about the source domain (birch trees, let’s say) and what we know about the target domain (the girl, a person). In characterizing what we know, Black emphasizes the ramifying texture and cultural relevance of the facts that we must consider; he called these facts systems of associated commonplaces. For example, it’s salient about birch trees that they are flexible, that they easily bend. This distinctive pattern of ideas is an important part of what we have to reconcile with the girl in order to make sense of the metaphor. This tension is resolved by a basic operation; and this operation constitutes the distinctive perspective-taking metaphors allow. We take those aspects of the source that fit the target—a set of properties, let’s say, that conventionally are associated with the source domain, and then we filter out what we’re thinking about the target by emphasizing these aspects, which may be things we know about the target or not, but which prompt us to reassess and reinterpret what we know about the target in a way that highlights consequences of these important features of the source domain.
Eco pursues this idea of the operation of metaphorical thinking. He particularly wants to contrast it with general notions of comparison and general notions of exploring important details in imagery; one thing as another is quite different from merely comparing two things. Seeing one thing as another is different from realizing that particular features of something are important or distinctive or affectively laden.20 “The metaphor’s cleverness” lies “in making us see a certain resemblance between different things.”21
One way of eliciting this idea is to contrast metaphors with telling details. Telling details are facts or information that the speaker provides in setting the scene—or in informing the audience what he intends for them to think further about—that can develop open-ended understanding, where lots of information, lots of consequences are easily imagined through elaborating these telling details. What is more, it can be important to the audience’s appreciation of the information that they explore the consequences of those telling details because those telling details might tell them important emotional information about a scene: how they are meant to react to it, who they are supposed to empathize with, and what sort of feelings should be prompted.
We naturally expect that we can color our interpretation of a text by presenting telling details that prompt relevant inferences. The Chinese poet Li Po is someone who used telling details influentially in poetry; one line of his poem that Ezra Pound translated is:
I watch the moon through the clear autumn.
In the context of this poem we know this line to be uttered by a woman whose lover has failed to meet her at the appointed assignation. The moonlight conveys the stillness and loneliness of knowing she has been rejected, while the clear autumn night underscores the kind of arbitrariness of her lover’s failing to come to meet her.
There’s lots of information we can have in a text that colors our appreciation, information that we think more and more about, but interpreting metaphor goes beyond this, in particular, in perspective-taking, that is, in seeing one thing as another. This requires that in interpreting a metaphor we have to reconstruct a correspondence that allows us to reinterpret the girl’s properties as birch-like. This distinct aspect of metaphorical thinking comes from this perspective-taking, which is not something that automatically follows from any independent principles of pragmatic reasoning.
In effect we are opting for one side of Eco’s radical choices, namely, that “the mechanism of metaphor establishes linguistic activity, every rule or convention arising thereafter in order to discipline, to reduce (and impoverish) the metaphorizing potential that defines man as a symbolic animal.”22 And given that metaphor is special in this way and that it requires a distinctive kind of thinking, you might wonder why bother trying to assimilate metaphor to theories of speaker-meaning. This is exactly the conclusion that Davidson drew, namely, that there is no distinctive meaning in metaphor.23 In this regard metaphor does not constitute “a breakdown, a malfunction, an unaccountable outcome.”24
What is distinctive about metaphor is that we recognize this correspondence, take up a particular perspective on the topic and target of the metaphor, and think through this perspective to whatever insights we may gain. That’s all there is. There’s nothing more to say about meaning. A key question for Davidson, one which he pursues in a variety of ways (and which we have also explored in a variety of ways), is why is it that we can use metaphors in all the ways we do if no proposition is encoded by them?25
III. AGAINST METAPHORICAL MEANING
When Donne says “No man is an island; the belle tolls for thee,” he’s counting on his audience to draw certain conclusions from their engagement with the world as he just described it, but he’s leaving those conclusions open for his audience; he’s not packaging them into meanings. It might help in grasping this point to note how easy it is to give instructions in metaphor, a point that Eco himself emphasizes in noting that metaphor is not limited to language alone.26
Coaches in athletic events often tell students what to do by using metaphors. The butterfly is very difficult to swim. Your hips sink in the water; and at the same time, it’s still hard to get your arms out of the water and keep your mouth high enough to breathe. You have to relax and flow through it. And one way a coach can get swimmers to swim the butterfly the right way is by telling them: “Do this—with each stroke let a wave flow along the length of your body, from your head down your back and through your legs.” This wave is obviously just an imaginary one; it’s a metaphor for the orchestration of movements that a swimmer has to perform in order to swim this stroke successfully. How is it, then, that a coach can succeed in telling you what to do with a metaphor if this metaphor doesn’t contribute a meaning by which it informs you of what it is you are supposed to do? But why not take the language at face value? Suppose, for example, we say that the instruction just asks the swimmer to move and to use the image of a wave to guide the orchestration of her movement, that is, to think about what’s literally instructed of her as she tries to meet that instruction. Is there anything more we have to say—“no!”—because this is what we do when we follow this instruction? And this is what lets us swim the butterfly by following this instruction.
An approach to metaphor that does not posit metaphorical meanings has much to recommend it. Rather than saying a speaker derives an interpretation for a metaphor in the ordinary way, we say something radically different. We say that the metaphor has its own mechanism that distinctively applies to metaphors and only to metaphors, where the metaphorizer exploits his audience’s knowledge of one domain in order to put a perspective on something. Metaphors produce “something which, psychologically speaking, we could call ‘excitation.’”27
This is a distinctive open-ended inference that is not grounded in general principles of psychology, but, rather, in the particular power of a kind of imagination that’s part of our psychology—a very particular part of it. The information we acquire through a metaphor that comes from this process is not pragmatic in the sense of being part of speaker-meaning, that is, it is neither signaled by the speaker nor recognized by the hearer. Rather, it becomes an extension of the external world around us, a place where our perceptions and demonstrations can inform the thinking and interaction we do with one another, but it is not part of our communication, that is, it is not part of the communicative enterprise. This is the sense in which we suggest that a metaphor is just a different sort of thing than ascribing a speaker-meaning. We believe this is Eco’s point when he celebrates that “that metaphor is ‘good’ which does not allow the work of interpretation to grind to a halt . . . but which permits inspections that are diverse, complementary and contradictory.”28
This sort of flexibility and open-endedness means that metaphorizers do not intend their audiences to calculate any specific proposition that is distinct from the proposition their words literally express.29 Metaphorical interpretation, in fact, seems to draw so tightly on the psychology of experience, attention, and inference—those chains of association that lead a wandering mind from one idea to another—that it is a mistake to describe the process in terms of the calculation of propositions at all.30 In any case, even when there is information that an audience is supposed to take away from a confrontation with a metaphor, and even when we can be fairly precise about what this information is, the metaphorical imagination that leads us to these insights is a matter of exploring similarities and correspondences between our understanding of the source and target domains of the metaphor.
IV. DIFFERENCES WITH ECO
Metaphorical meaning, if it existed, would require an audience to recognize the specific content that a speaker wishes to get across with his words, and to use the signal of the metaphor as the basis for the uptake of that content. This does not happen in normal confrontations with metaphors; as Eco himself notes, “once the process of unlimited semiosis has started, it is difficult to say where and when the metaphorical interpretation stops.”31 We agree with him and so we reject metaphorical meaning. But then we are puzzled by Eco’s talk of coming to understand a metaphor,32 of assuming a code verified against a simile,33 and of interpreting metaphors as “rule-governed creativity” or “rule-changing creativity.”34 This may be just a loose way of talking about what happens to us when we are confronted by a metaphor; and if so we have no qualms. But if he means that a theory of meaning must address not only literal meaning but a metaphorical one as well, then we disagree with him and indeed we do so partly for the very reasons he advances.
We are not denying that metaphors can be used with the intention of drawing a hearer’s attention to similarities, but this is not the same as an intention to convey propositional content. We do agree with Eco that a metaphor succeeds, if it does, from the cognitive effort an audience puts into exploring the similarities suggested by the metaphorical imagery, but we disagree with him if he takes this to mean that the cognitive effort succeeds and derives from recognizing a speaker’s intention to convey propositional content. We believe that taking our perspective allows us to afford a specific place for metaphor in the architecture of cooperative interaction without ascribing any meaning to a metaphor or its uses beyond the literal.
V. CONCLUSION
We have distinguished metaphorical thinking—developing imagery, seeing one thing as another, noticing similarities—from merely grasping a proposition, namely, the one that is speaker-meant. But why should anyone care if the point of a metaphor is characterized as a meaning or not? What’s in a name? Our answer is straightforward: both meaning and metaphor sit at the center of complex webs of phenomena, principles, and puzzles. There is much to gain from keeping these issues separate from one another. In the realm of meaning the hardest problems come in regimenting our knowledge of language and thus coming to understand how that knowledge is acquired, structured, and used in speakers’ faculty of language. Our intuitions about literal meaning place heavy constraints on key semantic notions: truth, reference, context, and logical form, among others. If we can locate metaphor elsewhere, it is good news for meaning.
The study of metaphor likewise benefits from putting issues of meaning aside. Eco, for one, has expended a great deal of creative energy over the past half century in trying to figure out what we do when we appreciate or craft a metaphor. We believe that it prejudges such questions to classify metaphor as carrying meaning: it suggests that we produce and understand metaphors in more or less the ordinary way that we produce and understand all other language. We most certainly do not.
Since Aristotle’s invocation of the special genius of the metaphorist, critics have worked to pin down the ingredients that make metaphor special. Eco locates the power of metaphor in the real relationships among things that only metaphorical thinking can bring us to notice. Barfield does as well in quoting Bacon on metaphor: “Neither are these only similitudes, as men of narrow observation may conceive them to be, but the same footsteps of nature, treading or printing upon several subjects or matters.” Barfield goes on:
This is the answer. It is these “footsteps of nature” whose noise we hear alike in primitive language and in the finest metaphors of poets. Men do not invent those mysterious relations between separate external objects, and between objects and feelings or ideas, which it is the function of poetry to reveal. These relations exist independently, not indeed of Thought, but of any individual thinker. And according to whether the footsteps are echoed in primitive language or, later on, in the made metaphors of poets, we hear them after a different fashion and for different reasons.35
Meaning does not work this way even if metaphor might. Therefore, divorcing metaphor from meaning opens the door to honor, to refine, and to defend such insights, in the terminology of psychology, action, and information. Where it belongs; and where Eco would have us place it.
ERNIE LEPORE AND MATTHEW STONE
RUTGERS UNIVERSITY
NOVEMBER 2015
NOTES
1. Umberto Eco, The Role of the Reader (Bloomington: Indiana University Press, 1979), 5.
2. For a more recent discussion of these shortcomings, see Umberto Eco, Kant and the Platypus: Essays on Language and Cognition, trans. Alastair McEwen (New York: Harcourt Brace & Co., 2000), 270–79.
3. See, in particular, Umberto Eco, A Theory of Semiotics (Bloomington: Indiana University Press, 1976), chs. 2–3.
4. See J. Hawthorne and E. Lepore, “On Words,” Journal of Philosophy 108, no. 9 (2011): 447–85. Of course if you separate wordhood from its various articulations, you then need a story to tell which articulations are articulating which words; but at least the position is not incoherent.
5. Eco, A Theory of Semiotics, 66–68. See in particular his discussion of meaning as cultural unit.
6. See, for example, Dan Sperber and Deidre Wilson, Relevance: Communication and Cognition, 2nd ed. (Oxford: Blackwell Publishers, 1995); R. H. Thomason, “Accommodation, Meaning, and Implicature: Interdisciplinary Foundations for Pragmatics,” in P. R. Cohen, J. Morgan, and M. E. Pollack, eds., Intentions in Communication (Cambridge, MA: MIT Press, 1990), 325–63; Herman Cappelen and Ernest Lepore, Insensitive Semantics: A Defense of Semantic Minimalism and Speech Act Pluralism (Oxford: Blackwell, 2005); and David Lewis, Convention: A Philosophical Study (Cambridge, MA: Harvard University Press, 1969).
7. See Eco, The Role of the Reader and Umberto Eco, Semiotics and the Philosophy of Language (Bloomington: Indiana University Press, 1984).
8. Matt Groening, The Big Book of Hell (New York: Pantheon, 1990).
9. Eco, The Role of the Reader, 68.
10. John Searle, “Metaphor,” in Expression and Meaning: Studies in the Theory of Speech Acts (Cambridge: Cambridge University Press, 1979). See also Eco, Semiotics and the Philosophy of Language, 89.
11. We are implicitly exploiting developments within relevance theory in framing this account; in particular, the practice of referring to a metaphorical meaning with snowmobile* is due to Robyn Carston. She calls snowmobile* an ad hoc concept; that is, something not specifically given by any of the conventions of language, but something you can recover in context, because of the constraints imposed by a communicative intention; or imposed by the pull of relevance. See Robyn Carston, Thoughts and Utterances: The Pragmatics of Explicit Communication (Oxford: Blackwell, 2002).
12. See Eco, Semiotics and the Philosophy of Language, 88: He calls this violation or flouting “a scandal.”
13. Dan Sperber and Deirdre Wilson, Relevance: Communication and Cognition. See also Robyn Carston, “Implicature, Explicature, and Truth-Theoretic Semantics,” in Steven Davis, ed., Pragmatics: A Reader (Oxford: Oxford University Press, 1991).
14. Aristotle, Rhetoric, in The Works of Aristotle, vol. 11, trans. W. Rhys Roberts, ed. W. D. Ross (Oxford: Clarendon Press, 1924), 1406b.
15. See Max Black, “Metaphor,” Proceedings of the Aristotelian Society, New Series, vol. 55 (1954–1955): 273–94.
16. Eco, Semiotics and the Philosophy of Language, 113.
17. See, for example, Nelson Goodman, Problems and Projects (Indianapolis: Bobbs-Merrill, 1972); and Amos Tversky, “Features of Similarity,” Psychological Review 84, no. 4 (1977): 327–52.
18. Dedre Gentner and Brian F. Bowdle, “Metaphor as Structure-Mapping,” in Raymond W. Gibbs, ed., The Cambridge Handbook of Metaphor and Thought (New York: Cambridge University Press, 2008), 109–28.
19. Elisabeth Camp, “Saying and Seeing-As: The Linguistic Uses and Cognitive Effects of Metaphor,” dissertation, University of California, Berkeley, 2003.
20. Others speak of “picturing,” “imagistic,” “framing” effects or “aspect seeing,” “brainstorming,” “imaginative play.”
21. Eco, Semiotics and the Philosophy of Language, 95. See also Kant and the Platypus, 31–35.
22. Ibid., 88.
23. Donald Davidson, “What Metaphors Mean,” in Sheldon Sacks, ed., On Metaphor (Chicago: University of Chicago Press, 1979), 29–45.
24. Eco, Semiotics and the Philosophy of Language, 88.
25. Ernie Lepore and Matthew Stone, “Against Metaphorical Meaning,” Topoi 29, no. 2 (2010): 165–80.
26. Eco, Semiotics and the Philosophy of Language, 89.
27. Eco, The Role of the Reader, 86.
28. Eco, Semiotics and the Philosophy of Language, 120.
29. Ibid., 123. Eco sometimes refers to this open-endedness as “ambiguity,” whereas we prefer the term “indeterminacy” since the latter clarifies the sense in which we think metaphors lack special meanings and Eco’s expression suggests they have many meanings.
30. See Davidson, “What Metaphors Mean.” See also Richard Rorty, “Unfamiliar Noises I: Hesse and Davidson on Metaphor,” Proceedings of the Aristotelian Society 61 (1987): 283–96.
31. Eco, Semiotics and the Philosophy of Language, 124. See also ibid., 127.
32. Ibid., 96.
33. Ibid., 101, 104.
34. Eco, The Role of the Reader, 68, 78–79.
35. Owen Barfield, Poetic Diction: A Study in Meaning (Middletown: Wesleyan University Press, 1973), 86.