MADISON ON METAPHOR, BY DONALD STEWART

In his book Understanding: A Phenomenological-Pragmatic Analysis, Gary Madison, true to his title, offers us a phenomenological-pragmatic account of metaphor. “The ‘meaning’ of metaphorical discourse, ” he says, “is nothing other than the practical transformation it brings about in the listening and speaking subject, the orientation it communicates to understanding.”¹ According to him, “a metaphor does not say what it means, for what it says, its literal meaning, is precisely what it does not mean.”² Rather, its true “meaning” is said to lie “entirely in its ‘perlocutionary force,’ in the effect the words have on us.”³

This view of the matter, though I believe incorrect, has the salutary merit of taking the now generally accepted view of metaphor initiated by I. A. Richards and developed in one form or another by such authors as Wheelwright, Beardsley, Black, Khatchadourian, Goodman, and Ricoeur to its logical and extreme conclusion—that metaphor is, speaking strictly, meaningless. His analysis, which sets itself against what he calls the “rationalist” view that metaphor is a figure of speech based on simile or “essential similarities” (Aristotle), contends that since metaphor is meaningless literally, it creates an entirely new (type of) meaning which is original, unparaphraseable without cognitive loss and quite literally unspecifiable. What Madison’s analysis achieves is a way out of the paradox so succinctly captured by Alston in the claim that “it is an extremely important fact about language that it is possible to use a word intelligibly without using it in any of its senses.”⁴ For Madison, the intelligibility of metaphor does not lie in its meaning at all, for it has none, but in its power to effect a change of attitude, direction, and, ultimately, understanding on the part of the listener. The power of metaphor is performative, not semantic. It changes the subject, directs the subject to a new way of looking at something, effects a new opening into the world.

The problem with this phenomenological-pragmatic view of metaphor, like that of the tradition from which it stems, is that it has no measure whatever of the appropriateness of the metaphorical attribution. It is all very well to say with Wheelwright, Black, and Madison that, far from discovering preexistent similarities, “the likenesses that metaphor articulates are actually, in a very important sense, created by metaphor,”⁵ but it is quite another to show that they are then really likenesses rather than merely distorting or idiosyncratic images. Even if we want to maintain that the primary function of metaphor is to create insight, we still want to be able to say that not all metaphor reduces to the wild diaphor of dada, and on this theory we cannot.

Let me take one of Madison’s most important examples, that of Saint Thomas’s account of the analogical predication of properties to God. Madison rejects Thomas’s “rationalist” account of analogical predication and maintains that “there is more to Thomas’ practice than filters through his rationalist theories.”⁶ The important thing about (metaphorically) calling God good is, according to Madison, that we turn ourselves to God, not the question whether God is good. But unless there is meaning to the phrase, then it is hard to see how this statement, rather than any other, turns us to God. Perlocutionary force is dependent, in almost every case, upon semantic meaning. “God is good” may turn us toward him, but it can do so only where there is meaning, and we can see that the attribution of goodness to him is, at least initially, not simply absurd, even if not wholly literal. Unless Madison is willing to admit either that similarities do inform metaphor antecedently, or that metaphors gain meaning in some other way, he is unable to deal with the problem of just how the understanding is changed.

Fortunately, there is more to Madison’s practice than filters through his phenomenological-pragmatic theory. In an earlier chapter entitled “Analogy and Being,” he gave an account of analogy as midway between the rationalism of those who hold that the meanings of terms are essences and the relativism of those who hold that words mean whatever you want them to. “Only analogy of proper proportionality,” he says, “constitutes a true analogy; only it provides a genuine alternative to both univocity and equivocity.”⁷ It is this insight which I would like to follow up, for it can, as Madison says, mediate between the essentialist view of meaning, which relegates metaphor to a merely figurative way of saying something literal, and Madison’s phenomenological-pragmatic view, which claims that metaphor has meaning only in what it does rather than in what it says.

Madison, if I read him correctly, seems, in fact, to have accepted the dichotomy between essentialism and relativism in accepting the distinction between metaphorical meaning (relativism) and literal meaning (essentialism). It is hard to see why he did not follow up the lead of analogy in this connection, though he seems to have been sidetracked into thinking of analogy merely as a way of speaking rather than as a basis for both metaphorical and literal speech.

If, as he claims, “metaphorical ‘likeness’ is prior to and more basic to the life of understanding than logical ‘identity,’ ”⁸ then it would seem reasonable to claim that such likeness is prior in reality as well as in metaphor. This is to say that two or more instances denoted univocally by a term are qualitatively, as well as numerically, distinct. The relation between them is one of likeness rather than identity. On this view, no two things, properties or events univocally called by the same name need be qualitatively or substantively identical, though they may turn out to be so. They are similar with respect to the univocal use of a given term to denote or describe them. Essences, in other words, are nominal. Further, we may choose to classify objects and properties according to different similarities than those we have in fact chosen.

When we do this with the intent of permanent reclassification, we attempt to redefine the terms we use. When we do so for specific purposes in specific contexts, we speak metaphorically. Thus, when Sir Walter Raleigh writes, “The hardest steel eaten with softest rust / The firm and solid tree rent and rotten,” we immediately pick out the similarity between the action of rust on steel and animals eating. Moreover, the similarity is such that we could redefine the animal kingdom so that rust would be included. It would be puerile to do so since we would miss important differences between animate and inanimate objects, but this does not mean that it is theoretically impossible. Indeed, philosophical understanding is often achieved in just this way, as when Berkeley defined being in terms of perception, or Hegel, substance in terms of subjectivity.

That is not what is at issue here, however. The point I want to make here is that the term “eaten” is meaningful in the normal way—on the basis of a perceived similarity—and that this is true of literal language as well. Look at the second line of the Raleigh verse. Is “rent” used metaphorically or not? Cloth is normally rent but trees seldom, yet it is much less metaphorical than the use of the term “eaten” in the previous line. “Rent” is used in its normal sense and the whole sentence makes perfectly good, if not quite literal, sense. So too, “eaten” is used in its normal sense, though it is used metaphorically. Further, both sentences are meaningful because both are based on a clear similarity. It is in this similarity that the appropriateness, even the truth, of metaphor resides.

There is, in principle, no difference in the way in which metaphor carries its meaning and the way literal language carries its. Both are founded upon similarities which can seldom, if ever, be reduced to univocal reference to essences. Just as the action of rust eating steel is only similar to that of, say, a man eating flesh, so the action of the man is only similar to that of a horse, both of those to the action of a fish eating flies, and all of these to, let us say, moths eating holes in sweaters. Similarity is not reducible to identity except nominally, and this is crucially important for it shows that it is a matter of decision what we call literal and what metaphorical. We fix our language in certain, by and large, consistent configurations of univocal uses. But these configurations could be otherwise, and it is this fact that both allows for and encourages metaphor.

Metaphor can pick out the similarities which our literal language must, perforce, overlook or relegate to mere accidental features of our experience. In calling a man a wolf, for example, we gain the immediacy which literal language must pass over in, let us say, calling him fierce. To do the job we should have to call him “wolf-fierce,” and while this, too, is not impossible it remains metaphorical precisely because men are “man-fierce,” not “wolffierce,” though both presumably are part of what is intended by the literal and univocal uses of the term “fierce” over men, wolves, lions, and so on. Metaphor gains the specificity, which to a large degree literal language overlooks and must overlook. In principle, however, there is no difference between the two types of language since, were we to accept the metaphorical classification of men as wolves on the basis of the similarities between them, we should be faced with a new literal structure which would require a new metaphor to (re)establish the difference specific to each.⁹

This also explains why writers such as Black and Madison want to say that metaphor creates similarity rather than discovers it. We “create” similarities when we see that dissimilars are yet similar, when we see that, though God’s goodness is different from our goodness, both are comparable. Aristotle said that the greatest thing of all is to be the master of metaphor, for then one “saw” the similarity among dissimilars, and he was right. The act of comparison is an essential element in discovering similarities.

Madison seems perfectly aware of this when he writes:

The problem with this analysis is that it implicitly accepts the “rationalist” hypothesis that there really are natural kinds or essences which define both literal language and, derivatively, metaphorical language. It is to pass over the fact which he himself maintains, that the world is, in its relations between particulars, analogical and open to any reasonably sensible choice among all the possible language systems which can describe it. Literal language is just the one we have hit upon; it has no special theoretical status over any other possible system even though it has very considerable practical advantages over most such “metaphorical” systems. To put it in terms of Saussure’s distinction betwen la langue and la parole, la parole métaphorique invokes a different langue than that which we now accept.

Madison is absolutely right in saying that reality is analogical. When he says that “to say that something is like something else is to say that it both is and is not that other thing,” he is fundamentally right. What it seems to me he fails to see is that this is true whether or not we speak in analogies, speak literally or speak metaphorically. It is not the fact that we use an analogy which creates the difference in unity which he characterizes by saying something is and is not something else; all speech rests upon this basic truth, for univocity is nothing other than the cancellation of the differences in the similarities upon which it is based.

When we say, then, that the statement that God is good is metaphorical, we are not consigning it to the class of speech whose meaning is merely performative. Rather, we are recognizing the vast difference between God’s goodness and our own while maintaining that the term is univocal over the two cases. This is not Thomas’s rationalist view. Neither is it Madison’s phenomenological-pragmatic view. Rather, it is a view which I hope does some justice to the insights of both sides.

ON STEWART ON MADISON ON METAPHOR,
BY G. B. MADISON

In a letter to me Don Stewart said: “I think I have understood your position—I hope I haven’t mangled it.” Well, I don’t think he has understood what he calls the Madisonian theory, and I do think that he’s mangled it. On the other hand, he may have understood it all too well and simply not liked what he saw there. Many people, rationalists especially, haven’t—not that I’m accusing Stewart of being a member of that dreadful bunch of people. On yet another hand (there’s always another hand!), I may not have understood his critique of me. That’s quite possible. With these caveats, respondeo:

Stewart’s main point is that there is no real difference in kind between metaphor and literal speech. He does not like my saying that there is. But there is, like it or not. (Why should we have two different terms, “metaphorical” and “literal,” anyway, if there were not two different things to be referred to here?) Let me try to say, with sufficient clarity, I hope, just what it is that I meant.

First of all, as regards generalities, the difference between metaphorical and literal discourse is a free-floating and ever-changing one. That is, new, living metaphors are ever degenerating into stale, dead metaphors, clichés, literalities (if I may be allowed to coin a word). Moreover, like newly minted coins, they rapidly lose their luster as they circulate throughout a speaking and writing community. Because of linguistic inflation also, they lose whatever intrinsic value they might have had, becoming the near-worthless (in that they no longer give rise to new, insightful understanding) medium for the exchange of hackneyed, trite, and trivial ways of understanding and talking about things. Then, if I may be allowed to play with the metaphor (talking about metaphor metaphorically is no doubt the best way of saying what it in fact is; let us not forget that the word metaphor is itself a metaphor for what it is, albeit a dead one), they, like paper currency, have absolutely no intrinsic value. They are rapidly worn out as they are passed from hand to hand, thoughtlessly. Fortunately, however, human inventiveness being what it is, crisp, new metaphors are constantly being printed and put into circulation.

It is a basic tenet of phenomenological hermeneutics (of the variety I represent, in any event) that it is precisely this wearing-out process which is the source of literal meanings, of “concepts” which are taken to be the straightforward description of what things as a matter of prosaic fact are. All literal, pedestrianly descriptive meanings are nothing more than the fossilized remains of living, creative, playful metaphors. Univocal concepts, expressing the unadorned, literal truth of things, are nothing more than dead metaphors. The end, banal result of this degenerative process is the formal, frozen, mummified language of symbolic logic. Here words are reduced to mere signs, mere signifiers devoid of transcendent signifieds, meaningless signs, therefore, in that they point to nothing beyond themselves and, in their pure diacriticality, are expressive of nothing other than themselves.

Thus, although the difference between the metaphorical and the literal is a shifting and sliding one, it is nevertheless as deep and momentous a difference as the difference between what in my book I referred to as active (creative, insightful) understanding and passive (habitual, routinized) understanding. There is as great and significant a difference between the two as there is between (using these words in a purely metaphorical sense) wakefulness and sleep, poetry and prose.

With regard to more linguistically specific matters, what I want to maintain is that the logic (or illogic) of metaphorical predication is not at all the same as that of nonmetaphorical, “literal” predication. This is what I was trying to get at in my book by talking about the is/is-not character of metaphorical statements. Unlike straightforward constative utterances, a metaphor says, at one and the same time, that something both is and is not (like) something else. If we read or understand metaphors as straightforward, literal statements (“This is that”), we totally misread and misunderstand them. Thus, although metaphors look like constative utterances, they most definitely are not constative utterances. Looks, as everyone knows, can be deceiving. Products of a playful, ironic turn of mind, metaphors are deliberately deceptive and misleading. They do not really mean what they merely seem to mean, so that if we take them literally, we are quite simply taken in by them. Or as I said in my book (and it is important to note that I was there speaking from a semiological or grammatical point of view, most definitely not from a phenomenological point of view, i.e., the way in which we actually experience metaphor):

One of the most striking differences between metaphors and literal statements lies precisely in the metaphor’s deceitfulness. Linguistic analysts such as Searle and others have pointed out that an essential feature of normal, constative utterances is that they carry with them an implicit claim that the utterers are sincere and mean what they say. This, from a grammatical point of view, is obviously not the case in regard to metaphorical utterances, and so it should be equally obvious that, linguistically speaking, metaphors are in a different class altogether from literal statements.

I said above, in quoting myself, that the real “meaning” of a metaphor lies not in what it “says” but in what it “shows.” The real meaning of a metaphor is something extralinguistic. I do not mean by that that its meaning is its “reference,” an extralinguistic state of affairs in the “real” world to which it is commonly said it is the business of language to “refer.” Such a state of affairs is, as I have argued in my book, nothing more than a rationalist’s fiction.¹ I mean, rather, that the meaning of a metaphor is not what it says but what it does, the perlocutionary effect it has on us, this being the change it brings about in our customary way of viewing things, of seeing them (all seeing is a seeing-as, and metaphor enables us to see things in new ways). Metaphors are really not constatives at all but performatives. All language says something, but when metaphors say something they turn around and cancel out what they say. They are (to use a metaphor that Wittgenstein plagiarized from Sextus Empiricus) like ladders that one kicks away after one has, by their means, ascended to a higher place, a heightened understanding.

To assert, as Stewart does, that this amounts to saying that metaphors are meaningless is to assert something that one has no right to assert—unless it be the case that meaning is a purely theoretical, linguistic, semiotic affair and that our lived experience and the realm of praxis is itself meaningless. Being an unrepentant and unabashed phenomenologist, I refuse to accept that, to believe that our lived-through, preverbalized experience is not meaningful in some meaningful sense of the term. Being that can be thought is language, as Gadamer says, but that does not mean that being is nothing but words, that, as Derrida has wildly and nihilisticaly proclaimed, there is no “hors-texte.” What that means is that “in the mirror of language everything that exists is reflected.”² But, of course, that does not mean that language is but a mirror whose function is merely to mirror a reality which is what it is outside of the mirror. After Rorty, it is perhaps best not to speak of mirrors at all.

I am not saying that metaphors have no meaning. I am saying that their meaning is their power to effect a change of attitudes, direction, and, ultimately, understanding of the part of the listener or reader. I am thus saying that the meaning of a metaphor is not merely linguistic, semantic, that it is indeed something much more than this. Metaphor is more than just an odd way of saying things with words, it is a superbly effective way of doing things with words, of altering our way of being-in-the-world. As Bachelard says, the poetic image, i.e., true metaphor, “becomes a new being in our language, expressing us by making us what it expressses . . . it is at once a becoming of expression and a becoming of our being. Here expression creates being.”³

Now for some specific rebuttals. Stewart feels that the big problem with my “phenomenological-pragmatic” theory of metaphor is that it has no way of dealing with the “appropriateness” of metaphorical attribution. Well, that’s no big problem for me. I feel quite happy with saying that metaphor does not discover preexistent similarities and that “the likenesses that metaphor articulates are actually, in a very important sense, created by metaphor.” Of course, Stewart is right: that means that I cannot show that these likenesses are “really likenesses rather than merely distorting or idiosyncratic images.” I could only show that if I believed that the characteristics things are said to have belong properly to them. Only if I subscribed to what I. A. Richards so nicely called The Proper Meaning Superstition. Only if, in a word, I believed in essences. But I most emphatically do not believe in essences, and my whole book was a sustained attempt to get rid of them once and for all. I view it as a kind of Essay on Liberation—liberation from the false consciousness inculcated in us by traditional metaphysics and its insidious idea-products, that of “essence” in particular. I simply do not believe in “essences,” those figments of the metaphysician’s metaphorical imagination that no man, woman, or child has ever actually seen. Plato never was able to respond unsophistically to Antisthenes’s question: Does the “essence” horse look anything like a real horse?

I am opposed to essences for the same reason that my forerunner in what I call the Counter-Tradition, William of Occam, was: They limit my freedom, and being a being who is essentially free, I don’t like that. I don’t want to be constrained by essences. I am opposed to them because they restrict the freeplay of my imagination. I don’t want to be limited in my talk about things to merely “factual” statements about what they supposedly are, or about what I am. Why do things have to be what they merely are? Why do I have to be what I merely am? Why can’t things be other than they are? Why can’t I be other and more than what I am? Essences represent the repression of desire, and a being of desire is again what I essentially am. I don’t want to be repressed. I don’t want to live in a world pinned down by essences any more than I would want to live in the ideal Republic of that person who first invented essences, Plato. Nietzsche said:

I for one hope that the day will never come when everything is defined, and we are stuck with things that are nothing more than what they are, which is to say, have been said to be. Being a hermeneuticist, I would like, as Rorty might say, to keep the conversation of mankind going. Besides, what is wrong with Dada or Surrealism? What is wrong with, as Max Ernst said, an umbrella making love to a sewing machine on a dissecting table? Or, in other words, “the pairing of two realities which apparently cannot be paired on a place apparently not suited to them”? Do I detect, in Stewart’s reservations, a trace of the metaphysician’s unspoken fear of that horrific prospect that things might not really be what they’re thought to be, that being and thinking might not overlap, his Angst in the face of the cosmic abyss, of the wild being (l’être sauvage) which transpires through the cracks in the cosmic egg produced by playful metaphors?

I do not deny that there are similarities that inform metaphor antecedently—indeed, I said as much in my book in talking about metaphor qua epiphor—but I maintain that these “similarities” are themselves simply the rigidified remains of previous metaphors which first created them.

While it is true that I reject essentialism, as Stewart recognizes, it is most definitely not true, as he nonetheless says, that I accept the dichotomy between essentialism and relativism and accordingly view metaphorical meaning in a merely relativist way. The essentialism-relativism dichotomy is the great either-or, as Richard Bernstein would say (Beyond Objectivism and Relativism), of modern objectivistic-subjectivistic philosophy which it is the goal of phenomenological hermeneutics, and the goal of my book, to overcome.

Stewart is adding insult to injury when he says that my analysis “implicitly accepts the ‘rationalist’ hypothesis that there really are natural kinds or essences which define both literal language and, derivatively, metaphorical language.” If dealing with a certain conceptuality in the attempt to explode or deconstruct it is to “accept” it, then I must plead guilty. Maybe, like Cratylus, I should not have said anything at all. It is not easy, as I said in my book, to avoid completely the double bind which menaces any critique of rationalism. So I am not surprised if I have been misunderstood or misrepresented. I did say, after all, that “the critique of rationalism is almost universally misunderstood, for it is taken in terms of precisely that which it rejects”(p. 39).

Stewart makes a number of points with which I fully concur, and for which I in fact congratulate him. For instance: “We fix our language in certain, by and large, consistent configurations of univocal uses. But these configurations could be otherwise, and it is this fact that both allows for and encourages metaphor.” The fact that Stewart can say this, and, when he does so, think that he is saying something different from what I said in my book just goes to show, however, how much he has mangled my position.

Stewart says that univocity “is nothing else than the cancellation of the differences in the similarities upon which it is based.” That is one thing I can agree with him on. But what does that mean? I am not sure Stewart would be prepared to accept what it implies: that purely univocal concepts are intrinsically meaningless, for where there are no differences there is nothing but pure Identity, which is as unthinkable and thus as unintelligible a concept as any dreamed up by metaphysicians, being on a par with other such unthinkables as Infinity and Eternity. All of these pure positivities are nothing but empty sounds. When we think them we are, like the yogi in trance, no longer thinking anything in particular at all.

Stewart says that when the term or “name” (as Saint Thomas would say) “good” is predicated both of God and man, the term “is univocal over the two cases.” He recognizes that this is not Thomas’s view, and that it most definitely is not mine. But it is a view which he hopes “does some justice to the insights of both sides.” Maybe someday he will tell me how. I am sure it must be in a very subtle way that it does so. As subtle as the way in which Duns Seotus, the doctor subtilis, argued for the univocity of “being”—so subtly in fact that hardly anyone but a Scotian has ever claimed to have understood exactly what it was he meant.