Ernst Cassirer, "The Power of Metaphor"

Chapter 6 of Ernst Cassirer, Language and Myth, translated by Susanne K. Langer, New York, Harper & Brothers, 1946 [Dover Publications, 1953], 83-99. The book was “[o]riginally published in German as Number VI of the ‘Studien der Bibliothek Warburg,’ under the editorship of Fritz Saxl” as Sprache und Mythos: Ein Beitrag zum Problem der Götternamen. Leipzig, B. G. Teubner, 1925, 87pp.

This little study, entitled Sprache und Mythos, reveals the genesis of some of those great conclusions for which he [Cassirer] is known to the world; it gives one a look into the mental laboratory where new ideas are generated and developed.

Susanne K. Langer, Translator’s Preface

Language and myth stand in an original and indissoluble correlation with one another, from which they both emerge but gradually as independent elements. They are two diverse shoots from the same parent stem, the same impulse of sym-bolic formulation, springing from the same basic mental activity, a concentration and heightening of simple sensory experience.

Ernst Cassirer

Posted August 29, 2008

The Power of Metaphor

The foregoing considerations have shown us how mythical and verbal thought are interwoven in every way; how the great structures of the mythic and linguistic realms, respectively, are determined and guided through long periods of their development by the same spiritual motives. Yet one fundamental motive has so far remained unnoticed, which not only illustrates their relationship, but offers an ultimate explanation of it.

That myth and language are subject to the same, or at least closely analogous, laws of evolution can really be seen and understood only in so far as we can uncover the common root from which both of them spring. The resemblances in their results, in the forms which they produce, point to a final community of function, of the principles whereby they operate.

In order to recognize this function and represent it in its abstract nakedness, we have to pursue the ways of myth and language not in their progress, but in regress—back to the point from which those two divergent lines emanate.

And this common center really seems to be demonstrable; for, no matter how widely the contents of myth and language may differ, yet the same form of mental conception is operative in both. It is the form which one may denote as metaphorical thinking; the nature and meaning of metaphor is what we must start with if we want to find, on the one hand, the unity of the verbal and the mythical worlds and, on the other, their difference.

It has frequently been noted that the intellectual link between language and myth is metaphor; but in the precise definition of the process, and even in regard to the general direction it is supposed to take, theories are widely at variance. The real source of metaphor is sought now in the construction of language, now in mythic imagination; sometimes it is supposed to be speech, which by its originally metaphorical nature begets myth, and is its eternal source; sometimes, on the contrary, the metaphor-ical character of words is regarded as a legacy which language has received from myth and holds in fee. Herder, in his prize essay on the origin of speech, emphasized the mythic aspect of all verbal and propositional conceptions.

As all nature sounds; so to Man, creature of sense, nothing could seem more natural than that it lives, and speaks, and acts. A certain savage sees a tree, with its majestic crown; the crown rustles! That is stirring godhead! The savage falls prostrate and worships! Behold the history of sensuous Man, that dark web, in its becoming, out of verbis nomina—and the easiest transition to abstract thought! For the savages of North America, for instance, everything is still animate; everything has its genius, its spirit. That it was likewise among Greeks and orientals, may be seen from their oldest dictionary and grammar—they are, as was all nature to their inventor, a pantheon! A realm of living, acting creatures. . . . The driving storm, the gentle zephyr, the clear fountain and the mighty ocean—their whole mythology lies in those treasure troves, in verbis and nominibus of the ancient languages; and the earliest dictionary was thus a sounding pantheon.⁷⁹

The romantics followed the way indicated by Herder; Schelling, too, sees in language a “faded mythology,” which preserves in formal and abstract distinctions what mythology still treats as living, concrete differences.⁸⁰

Exactly the opposite course was taken by the “comparative mythology” that was attempted in the second half of the nineteenth century, especially by Adalbert Kuhn and Max Müller. Since this school adopted the methodological principle of basing mythological comparisons on linguistic comparisons, the factual primacy of verbal concepts over mythic ones seemed to them to be implied in their procedure. Thus mythology appeared as a result of language. The “root metaphor” underlying all mythic formulations was regarded as an essentially verbal phenomenon, the basic character of which was to be investigated and understood. The homonymity or assonance of denotative terms was supposed to break and direct the way for mythic fantasy.

Let us consider, then, that there was, necessarily and really, a period in the history of our race when all the thoughts that went beyond the narrow horizon of our everyday life had to be expressed by means of metaphors, and that these metaphors had not yet become what they are to us, mere conventional and traditional expressions, but were felt and understood half in their original and half in their modified character . . . . Whenever any word, that was at first used metaphorically, is used without a clear conception of the steps that led from its original to its metaphorical meaning, there is danger of mythology; whenever those steps are forgotten and artificial steps put in their places, we have mythology, or, if I may say so, we have diseased language, whether that language refers to religious or secular interests . . . . What is commonly called mythology is but a part of a much more general phase through which all language has at one time or other to pass.⁸¹

Before one can attempt any decision between these antagonistic theories, this battle for the priority of language over mythology or myth over language, the basic concept of metaphor requires scrutiny and definition.

One can take it in a narrow sense, in which it comprises only the conscious denotation of one thought content by the name of another which resembles the former in some respect, or is somehow analogous to it. In that case, metaphor is a genuine “translation”; the two concepts between which it obtains are fixed and independent meanings, and betwixt them, as the given terminus a quo and terminus ad quem, the conceptual process takes place, which causes the transition from one to the other, whereby one is semantically made to stand proxy for the other.

Any attempt to probe the generic causes of this conceptual and nominal substitution, and to explain the extraordinarily wide and variegated use of this sort of metaphor (i.e., the conscious identification of avowedly diverse objects), especially in primitive forms of thinking and speaking, leads one back to an essential attitude of mythic thought and feeling. Heinz Werner, in his study of the origins of metaphor, has presented a very plausible argument for the supposition that this particular kind of metaphor, the circumlocution of one idea in terms of another, rests on quite definite motives arising from the magical view of the world, and more especially from certain name and word taboos.⁸²

But such a use of metaphor clearly presupposes that both the ideas and their verbal correlates are already given as definite quantities; only if these elements, as such, are verbally fixed and defined can they be exchanged for one another. Such transposition and substitution, which operate with a previously known vocabulary as their material, must be clearly distinguished from that genuine “radical metaphor” which is a condition of the very formulation of mythic as well as verbal conceptions.

Indeed, even the most primitive verbal utterance requires a transmutation of a certain cognitive or emotive experience into sound, i.e., into a medium that is foreign to the experience, and even quite disparate; just as the simplest mythical form can arise only by virtue of a transformation which removes a certain impression from the realm of the ordinary, the everyday and profane, and lifts it to the level of the “holy,” the sphere of mythico-religious “significance.” This involves not merely a transfer-ence, but a real μετάβασις εις αλλα γενος; in fact, it is not only a transition to another category [the meaning of the Greek phrase.—A.F.], but actually the creation of the category itself.

If, now, one were to ask which of these two types of metaphor begets the other—whether the meta-phorical expressions in speech are produced by the mythic point of view, or whether, on the contrary, this point of view could arise and develop only on the basis of language—the foregoing considerations show that this question is really specious. For, in the first place, we are not dealing here with a temporal relation of “before” and “after,” but with the logical relation between the forms of language and of myth, respectively; with the way the one conditions and determines the other.

This determination, however, can be conceived only as reciprocal. Language and myth stand in an original and indissoluble correlation with one another, from which they both emerge but gradually as independent elements. They are two diverse shoots from the same parent stem, the same impulse of symbolic formulation, springing from the same basic mental activity, a concentration and heightening of simple sensory experience.

In the vocables of speech and in primitive mythic figurations, the same inner process finds its consummation: they are both resolutions of an inner tension, the representation of subjective impulses and excitations in definite objective forms and figures. As Usener emphatically said:

It is not by any volition that the name of a thing is determined. People do not invent some arbitrary sound-complex, in order to introduce it as the sign of a certain object, as one might do with a token. The spiritual excitement caused by some object which presents itself in the outer world furnishes both the occasion and the means of its denomination. Sense impressions are what the self receives from its encounter with the not-self, and the liveliest of these naturally strive for vocal expression; they are the bases of the separate appellations which the speaking populace attempts.⁸³

Now this genesis corresponds precisely, feature for feature, with that of the “momentary gods.” Similarly, the significance of linguistic and mythic metaphors, respectively, will reveal itself, so that the spiritual power embodied in them may be properly understood, only as we trace them back to their common origin; if one seeks this significance and power in that peculiar concentration, that “intensifi-cation” of sense experience which underlies all linguistic as well as all mythico-religious formula-tions.

If we take our departure once more from the contrast which theoretical or “discursive” conception presents, we shall find indeed that the different directions which the growth of logical (discursive) and mythic-linguistic conception, respectively, have followed, may be seen just as clearly in their several results. The former begins with some individual, single perception, which we expand, and carry beyond its original bounds, by viewing it in more and more relationships. The intellectual process here involved is one of synthetic supplementation, the combination of the single instance with the totality, and its completion in the totality.

But by this relationship with the whole, the separate fact does not lose its concrete identity and limitation. It fits into the sum total of phenomena, yet remains set off from them as something independent and singular. The ever-growing relationship which connects an individual perception with others does not cause it to become merged with the others. Each separate “specimen” of a species is “contained” in the species; the species itself is “subsumed” under a higher genus; but this means, also, that they remain distinct, they do not coincide.

This fundamental relation is most readily and clearly expressed in the scheme which logicians are wont to use for the representation of the hierarchy of concepts, the order of inclusion and subsumption obtaining among genera and species. Here the logical determinations are represented as geometric determinations; every concept has a certain “area” that belongs to it and whereby it is distinguished from other conceptual spheres. No matter how much these areas may overlap, cover each other or interpenetrate—each one maintains its definitely bounded location in conceptual space. A concept maintains its sphere despite all its synthetic supplementation and extension; the new relations into which it may enter do not cause its boundaries to become effaced, but lead rather to their more distinct recognition.

If, now, we contrast this form of logical concep-tion by species and genera with the primitive form of mythic and linguistic conception, we find immediately that the two represent entirely different tendencies of thought. Whereas in the former a concentric expansion over ever-widening spheres of perception and conception takes place, we find exactly the opposite movement of thought giving rise to mythic ideation. The mental view is not widened, but compressed; it is, so to speak, distilled into a single point. Only by this process of distillation is the particular essence found and extracted which is to bear the special accent of “significance.” All light is concentrated in one focal point of “meaning,” while everything that lies outside these focal points of verbal or mythic conception remains practically invisible. It remains “unremarked” because, and in so far as, it remains unsupplied with any linguistic or mythic “marker.” In the realm of discursive conception there reigns a sort of diffuse light—and the further logical analysis proceeds, the further does this even clarity and luminosity extend. But in the ideational realm of myth and language there are always, besides those locations from which the strongest light proceeds, others that appear wrapped in profoundest darkness. While certain contents of perception become verbal-mythical centers of force, centers of significance, there are others which remain, one might say, beneath the threshold of meaning.

This fact, namely, that primitive mythical and linguistic concepts constitute such punctiform units, accounts for the fact that they do not permit of any further quantitative distinctions. Logical contempla-tion always has to be carefully directed toward the extension of concepts; classical syllogistic logic is ultimately nothing but a system of rules for combining, subsuming and superimposing concepts. But the conceptions embodied in language and myth must be taken not in extension, but in intension; not quantitatively, but qualitatively. Quantity is reduced to a purely casual property, a relatively immaterial and unimportant aspect. Two logical concepts, subsumed under the next-higher category, as their genus proximum, retain their distinctive characters despite the relationship into which they have been brought.

In mythico-linguistic thought, however, exactly the opposite tendency prevails. Here we find in operation a law which might actually be called the law of the leveling and extinction of specific differences. Every part of a whole is the whole itself; every specimen is equivalent to the entire species. The part does not merely represent the whole, or the specimen its class; they are identical with the totality to which they belong; not merely as mediating aids to reflective thought, but as genuine presences which actually contain the power, significance and efficacy of the whole.

Here one is reminded forcefully of the principle which might be called the basic principle of verbal as well as mythic “metaphor”—the principle of pars pro toto. It is a familiar fact that all mythic thinking is governed and permeated by this principle. Whoever has brought any part of a whole into his power has thereby acquired power, in the magical sense, over the whole itself. What significance the part in question may have in the structure and coherence of the whole, what function it fulfills, is relatively unimportant—the mere fact that it is or has been a part, that it has been connected with the whole, no matter how casually, is enough to lend it the full significance and power of that greater unity. For instance, to hold magical dominion over another person’s body one need only attain possession of his pared nails or cut-off hair, his spittle or his excrement; even his shadow, his reflection or his footprints serve the same purpose. The Pythagoreans still observed the injunction to smooth the bed soon after arising so that the imprint of the body, left upon the mattress, could not be used to the owner’s detriment.⁸⁴

Most of what is known as “magic of analogy” springs from the same fundamental attitude; and the very nature of this magic shows that the concept in question is not one of mere analogy, but of a real identification. If, for instance, a rain-making ceremony consists of sprinkling water on the ground to attract the rain, or rain-stopping magic is made by pouring water on red hot stones where it is consumed amid hissing noise,⁸⁵both ceremonies owe their true magical sense to the fact that the rain is not just represented, but is felt to be really present in each drop of water. The rain as a mythic “power,” the “daemon” of the rain is actually there, whole and undivided, in the sprinkled or evaporated water, and is thus amenable to magical control.

This mystic relationship which obtains between a whole and its parts holds also between genus and species, and between the species and its several instances. Here, too, each form is entirely merged with the other; the genus or species is not only represented by an individual member of it, but lives and acts in it. If, under the totemistic conception of the world, a group or clan is organized by totems, and if its individual members take their names from the totem animal or plant, this is never a mere arbitrary division by means of conventional verbal or mythical “insignia,” but a matter of genuine community of essence.⁸⁶

In other respects, too, wherever a genus is involved at all, it always appears to be wholly present and wholly effective. The god or daemon of vegetation lives in each individual sheaf of the harvest. Therefore, an ancient but still popular rural custom demands that the last sheaf be left out in the field; in this remnant, the power of the fertility-god is concentrated, from which the harvest of the coming year is to grow.⁸⁷ In Mexico and among the Cora Indians the corn-god is supposed to be present, fully and unrestrictedly, in every stalk and even every grain of corn. The Mexican corn-goddess Chicomeco-atl in her maidenhood is the green stalk, in her old age the corn harvest; but she is also each separate kernel and each particular dish. Likewise, there are several deities among the Coras who represent certain kinds of flowers, but are addressed as individual flowers. The same is true of all the Coras’ demoniac creatures: the cicada, the cricket, the grasshopper, the armadillo are simply treated as so many individual wholes.⁸⁸

If, therefore, ancient rhetoric names as one of the principal types of metaphor the substitution of a part for the whole, or vice versa, it is easy enough to see how this sort of metaphor arises directly out of the essential attitude of the mythic mind. But it is equally clear that for mythic thinking there is much more in metaphor than a bare “substitution,” a mere rhetorical figure of speech; that what seems to our subsequent reflection as a sheer transcription is mythically conceived as a genuine and direct identification.⁸⁹

In the light of this basic principle of mythic metaphor we can grasp and understand, somewhat more clearly, what is commonly called the metaphorical function of language. Even Quintilian pointed out that this function does not constitute any part of speech, but that it governs and characterizes all human talk; paene quidquid loquimur figura est. But if this is indeed the case—if metaphor, taken in this general sense, is not just a certain development of speech, but must be regarded as one of its essential conditions—then any effort to understand its function leads us back, once more, to the fundamental form of verbal conceiving.

Such conceiving stems ultimately from that same process of concentration, the compression of given sense experiences, which originally initiates every single verbal concept. If we assume that this sort of concentration occurs by virtue of several experiences, and along several lines, so that two different perceptual complexes might yield the same sort of “essence” as their inner significance, which gives them their meaning, then at this very point we should expect that first and firmest of all the connections which language can establish; for, as the nameless simply has no existence in language, but tends to be completely obscured, so whatever things bear the same appellation appear absolutely similar.

The similarity of the aspect fixed by the word causes all other heterogeneity among the perceptions in question to become more and more obscured, and finally to vanish altogether. Here again, a part usurps the place of the whole—indeed, it becomes and is the whole. By virtue of the “equivalence” principle, entities which appear entirely diverse in direct sense perception or from the standpoint of logical classification may be treated as similars in language, so that every statement made about one of them may be transferred and applied to the other. Preuss, in a characterization of magic-complex thinking, says: “If the Cora Indian classes butterflies, quite absurdly, as birds, this means that all the properties which he notes in the object are quite differently classified and related for him than they are for us from our analytical, scientific point of view.”⁹⁰

But the apparent absurdity of this and other such classifications disappears as soon as we realize that the formation of these primary concepts was guided by language. If we suppose that the element emphasized in the name, and therefore in the verbal concept of “bird,” as an essential characteristic was the element of “flight,” then by virtue of this element and by its mediation the butterfly does belong to the class of birds. Our own languages are still constantly producing such classifications, which contradict our empirical and scientific concepts of species and genera, as for instance the denotation “butterfly” (Dutch botervlieg), in some Germanic tongues called a “butter-bird.”

And at the same time one can see how such lingual “metaphors” react in their turn on mythic metaphor and prove to be an ever-fertile source for the latter. Every characteristic property which once gave a point of departure to qualifying conceptions and qualifying appellations may now serve to merge and identify the objects denoted by these names. If the visible image of lightning, as it is fixed by language, is concentrated upon the impression of “serpentine,” this causes the lightning to become a snake; if the sun is called “the heavenly flier,” it appears henceforth as an arrow or a bird-the sun-god of the Egyptian pantheon, for instance, who is represented with a falcon’s head.

For in this realm of thought there are no abstract denotations; every word is immediately transformed into a concrete mythical figure, a god or a daemon. Any sense impression, no matter how vague, if it be fixed and held in language, may thus become a starting point for the conception and denotation of a god. Among the names of the Lithuanian gods which Usener has listed, the snow-god Bizgulis, the “Shimmerer,” appears beside the god of cattle, the “Roarer” Baubis; also in relation to these we find the god of bees, Birbullis the “Hummer,” and the god of earthquake, the “Thresher” Drebkulys.⁹¹ Once a “Roarer God” in this sense was conceived, he could not but be recognized in the most diverse guises; he was naturally and directly heard, in the voice of the lion as in the roaring of the storm and the thunder of the ocean.

Again and again, in this respect, myth receives new life and wealth from language, as language does from myth. And this constant interaction and interpenetration attests the unity of the mental principle from which both are sprung, and of which they are simply different expressions, different manifestations and grades.

Yet in the advance of human mentality even this conjunction, close and essential though it seems to be, begins to disintegrate and dissolve. For language does not belong exclusively to the realm of myth; it bears within itself, from its very beginning, another power, the power of logic. How this power gradually waxes great, and breaks its way by means of language, we cannot undertake to set forth here. But in the course of that evolution, words are reduced more and more to the status of mere conceptual signs. And this process of separation and liberation is paralleled by another: art, like language, is originally bound up entirely with myth.

Myth, language and art begin as a concrete, undivided unity, which is only gradually resolved into a triad of independent modes of spiritual creativity. Consequently, the same mythic animation and hypostatization which is bestowed upon the words of human speech is originally accorded to images, to every kind of artistic representation. Especially in the magical realm, word magic is everywhere accompanied by picture magic.⁹² The image, too, achieves its purely representative, specifically “aesthetic” function only as the magic circle with which mythical consciousness surrounds it is broken, and it is recognized not as a mythico-magical form, but as a particular sort of formulation.

But although language and art both become emancipated, in this fashion, from their native soil of mythical thinking, the ideal, spiritual unity of the two is reasserted upon a higher level. If language is to grow into a vehicle of thought, an expression of concepts and judgments, this evolution can be achieved only at the price of foregoing the wealth and fullness of immediate experience. In the end, what is left of the concrete sense and feeling content it once possessed is little more than a bare skeleton.

But there is one intellectual realm in which the word not only preserves its original creative power, but is ever renewing it; in which it undergoes a sort of constant palingenesis, at once a sensuous and a spiritual reincarnation. This regeneration is achieved as language becomes an avenue of artistic expression. Here it recovers the fullness of life; but it is no longer a life mythically bound and fettered, but an aesthetically liberated life.

Among all types and forms of poetry, the lyric is the one which most clearly mirrors this ideal development. For lyric poetry is not only rooted in mythic motives as its beginning, but keeps its connection with myth even in its highest and purest products. The greatest lyric poets, for instance Hölderlin or Keats, are men in whom the mythic power of insight breaks forth again in its full intensity and objectifying power.

But this objectivity has discarded all material constraints. The spirit lives in the word of language and in the mythical image without falling under the control of either. What poetry expresses is neither the mythic word—picture of gods and daemons, nor the logical truth of abstract determinations and relations. The world of poetry stands apart from both, as a world of illusion and fantasy—but it is just in this mode of illusion that the realm of pure feeling can find utterance, and can therewith attain its full and concrete actualization.

Word and mythic image, which once confronted the human mind as hard realistic powers, have now cast off all reality and effectuality; they have become a light, bright ether in which the spirit can move without let or hindrance. This liberation is achieved not because the mind throws aside the sensuous forms of word and image, but in that it uses them both as organs of its own, and thereby recognizes them for what they really are: forms of its own self-revelation.

⁷⁹ “Ueber den Ursprung der Sprache,” Werke (ed. Suphan), V, pp. 53 f.

⁸⁰ Schelling, “Einleitung in die Philosophie der Mythologie,” Sämtliche Werke, 2nd div., I, p. 52.

⁸¹ Max Müller, Lectures on the Science of Language, second series (New York: Scribner, Armstrong & Co., 1875), pp. 372-376.

⁸²Heinz Werner, Die Ursprünge der Metapher (Leipzig, 1919), esp. chap. 3, pp. 74 ff.

⁸⁴ Jamblichos, Protreptichos p. 108, 3, quoted after Deubner, Magie und Religion (Freiburg, 1922), p. 8.

⁸⁵ See Parkinson, Thirty Years in the South Seas, p. 7; quoted by Werner, Die Ursprünge der Metapher, p. 56.

⁸⁶ Cf. my study, Die Begriffsform im mythischen Denken (Leipzig, 1922), pp. 16ff.

⁸⁷ Cf. Mannhardt, Wald- und Feldkulte, 2nd ed. (Berlin, 1904-1905), I, 212ff.

⁸⁸ See Preuss, in Globus, Vol. 87, p. 381; cf. esp. Die Nayarit-Expedition, Vol. I, pp. 47 If.

⁸⁹ This is the more obviously valid if we consider that for mythic and magical thought there is no such thing as a mere picture, since every image embodies the “nature” of its object, i.e., its “soul” or “daemon.” Cf., for example, Budge, Egyptian Magic, p. 65:

It has been said above that the name or the emblem or the picture of a god or a demon could become an amulet with power to protect him that wore it and that such power lasted as long as the substance of which it was made lasted, if the name, or emblem, or picture was not erased from it. But the Egyptians went a step further than this and they believed that it was possible to transmit to the figure of any man, or woman, or animal or living creature the soul of the being which it represented, and its qualities and attributes. The statue of a god in a temple contained the spirit of the god which it represented, and from time immemorial the people of Egypt believed that every statue and figure possessed an indwelling spirit.

The same belief is held to this day among all “primitive” peoples. Cf., for instance, Hetherwick:

Some animistic beliefs among the Yaos of British Central Africa” (see footnote above, p. 70): “The photographic camera was at first an object of dread, and when it was turned upon a group of natives they scattered in all directions with shrieks of terror . . . In their minds the lisoka (soul) was allied to the chiwilili or picture and the removal of it to the photographic plate would mean the disease or death of the shadeless body (pp. 89 f.).

⁹⁰ Preuss, Die geistige Kultur der Naturvölker (Leipzig, 1914), p. 10.

⁹² For further details see the second volume of my Philosophie der symbolischen Formen, esp. pp. 54ff.