SOURCE: "The Condition of Music," in Unsuspected Eloquence: A History of the Relations between Poetry and Music, Yale University Press, 1981, pp. 287-346.

[In the following essay, Winn explores similarities and differences between poetry and music within the context of nineteenth- and twentieth-century movements toward autonomy in both arts, with particular emphasis on the efforts of the Parnassian and Symbolist poets to make the language of poetry independent of everyday meaning by emulating the effects of music.]

Wagner's ideas, not to mention his success and his arrogance, were bound to produce opponents, whose works, whether in music or in words, were just as certain to be dismissed as reactionary by his convinced followers. So when Eduard Hanslick published his slender volume Vom Musikalisch-Schönen (The Beautiful in Music) in 1854, it was possible for Wagnerites to dismiss it as mere polemic, the product of Hanslick's known antipathy to Wagner and friendship for Brahms. But the importance of Hanslick's work transcends not only these personal relations but the immediate concerns of nineteenth-century music as well; as Morris Weitz argues, "it is to music what Hume's Inquiry Concerning Human Understanding is to speculative philosophy, a devastating critique of unsupportable views and an attempt to state clearly and precisely the territories and boundaries of the areas they discuss." More than that: The Beautiful in Music separates and dismisses the various imitative and emotive notions about the function of music which Romantic thought had impressionistically blurred together; it boldly insists on music's autonomy, its independence not only from words (a battle already won in the eighteenth century) but from verbally definable feelings as well. The aesthetic position it stakes out is not only fundamental to twentieth-century thought about music, but strikingly similar to the literary aesthetics of I. A. Richards and the American New Critics.

For musicians, the idea of autonomy would prove a means of liberating themselves from all kinds of imitative demands; as the coherence of classical Viennese instrumental music had given the quietus to the Renaissance claim that music's function was to serve or express words, Hanslick's doctrine would still the Romantic claim that its function was to express or imitate some specific emotion. Thus, in the course of his argument, Hanslick finds the recitative, the legacy of the Musical Humanists, a particularly convenient target:

In the recitative, music degenerates into a mere shadow and relinquishes its individual sphere of action altogether. Is not this proof that the representing of definite states of mind is contrary to the nature of music, and that in their ultimate bearings they are antagonistic to one another? Let anyone play a long recitative, leaving out the words, and inquire into its musical merit and subject.

Two birds with one stone: not only does music "degenerate" when it makes itself the slave of words, but "the beautiful tends to disappear in proportion as the expression of some specific feeling is aimed at."

Poets, too, were interested in autonomy, though the word necessarily had a different meaning for them. As Hanslick proclaimed the independence of music from the demands of words or specific emotions, Edgar Allan Poe, in an essay first published in 1850, "allude[d] to the heresy of The Didactic," the claim that "the ultimate object of all Poetry is Truth." The old doctrine that poetry had a duty to instruct as well as to please, like the doctrine that music had a duty to imitate something, took a variety of forms in the nineteenth century; W. K. Wimsatt distinguishes three—"the Shelleyan and Carlylean rhapsodic retort to scientism," a vigorous assertion that poetry also dealt in truth; "the Arnoldian neo-classic idealism," a prophecy that poetry would assume the functions of religion and philosophy; and "the sociorealistic propagandism" that would eventually harden into Soviet demands for "socialist realism." Concerned to make a quite different claim for poetry, Poe argues that it has nothing to do with truth:

The demands of Truth are severe. She has no sympathy with the myrtles. All that which is so indispensable in Song, is precisely all that with which she has nothing whatever to do. It is but making her a flaunting paradox, to wreathe her in gems and flowers. In enforcing a truth, we need severity rather than efflorescence of language. We must be simple, precise, terse. We must be cool, calm, unimpassioned. In a word, we must be in that mood which, as nearly as possible, is the exact converse of the poetical.

A familiar ancient distinction, if one more or less absent from Western aesthetics since the Renaissance; it might be Plato or Augustine on music and rhetoric, with the striking difference that Poe chooses Song rather than Truth. His Greek and patristic forbears, arguing that eloquence and melody were incompatible with truth, had sought in vain to control the claims of beauty, even in their own writings, but Poe "define[s] … the Poetry of words as The Rhythmical Creation of Beauty." And like many who would follow him, including (in their several ways) Pater, Wilde, Mallarmé, Verlaine, Pound, and Auden, Poe links poetry to music in the same gesture with which he separates it from didacticism:

Contenting myself with the certainty that Music, in its various modes of metre, rhythm, and rhyme, is of so vast a moment in Poetry as never to be wisely rejected—is so vitally important an adjunct, that he is simply silly who declines its assistance, I will not now pause to maintain its absolute essentiality.

Twenty-three years later, in an essay on the painter Giorgione, Walter Pater would maintain that "all art constantly aspires towards the condition of music."

There were powerful conceptual similarities between the musical and literary movements toward "autonomy." Central in both cases was the positing of some aspect of the mind between intellect and feeling, a three-part scheme where older accounts had only two. Thus Hanslick:

It is rather curious that musicians and the older writers on aesthetics take into account only the contrast of "feeling" and "intellect," quite oblivious of the fact that the main point at issue lies halfway between the horns of this supposed dilemma. A musical composition originates in the composer's imagination and is intended for the imagination of the listener.

And Pater:

Art… is always striving to be independent of the mere intelligence, to become a matter of pure perception, to get rid of its responsibilities to its subject or material; the ideal examples of poetry and painting being those in which the constituent elements of the composition are so welded together, that the material or subject no longer strikes the intellect only; nor the form, the eye or the ear only; but form and matter, in their union or identity, present one single effect to the "imaginative reason," that complex faculty for which every thought and feeling is twin-born with its sensible analogue or symbol.

Yet even in these obviously similar passages we see equally important differences: Hanslick is concerned to separate the imagination from feeling on the one side and intellect on the other, to establish an area of the mind from which music can be said to arise in the composer and to which it can be said to appeal in the listener, without either the claim that music is an intellectual language or the claim that it is a language of the feelings. Pater, by contrast, is concerned to fuse intelligence and sensuality, the mind and the ear, to produce his "imaginative reason." The movement toward musical autonomy was a movement toward abstraction, a declaration of independence from imitative demands; the movement toward poetic autonomy saw and admired in music not only its physical sensuality, but its abstraction as well. For Pater, music "most completely realises" his combining ideal, the "perfect identification of form and matter"; therefore "music,… and not poetry, as is so often supposed, is the true type or measure of perfected art." Hanslick, denying the capacity of musical form to express an external subject, makes a simpler claim: "the form (the musical structure) is the real substance (subject) of music—in fact is the music itself."

Hanslick's definition of autonomy can be more abstract than Pater's because the art he describes is finally more abstract; he can claim that structure simply is meaning, while Pater must necessarily talk about "welding" the two together. If, with Hanslick, we define the condition of music as the abstraction resulting from the essential neutrality and plasticity of its materials, we must recognize, as Pater's verb suggests, that poetry may aspire to that condition but cannot reach it. Words … resist attempts to empty them of their ordinary morphemic content. I may write a poem in which the word "swan" comes to function as a symbol for some larger abstraction, but even a reader who understands my method will find it hard to read that word without some notion, however fleeting, of a large, white bird with a curved neck. By contrast, I may change utterly the meaning of a note or even a chord in a composition; I may do so within a theoretical system or (more radically) by changing the entire system. The note G may be the tonic of one tonal piece, but in another piece in another key, it will take on another function; and since, in music, function is meaning, the fact that G has been the tonic of some other piece will be entirely irrelevant. As Hanslick compactly puts it,

The fundamental difference consists in this; while sound in speech is but a sign, that is, a means for the purpose of expressing something which is quite distinct from the medium, sound in music is the end, that is, the ultimate and absolute object in view.

W. H. Auden, in a poem entitled "The Composer," makes the same distinction from a poet's point of view:

All the others translate: the painter sketches
A visible world to love or reject;
Rummaging into his living, the poet fetches
The images out that hurt and connect.
From Life to Art by painstaking adaption,
Relying on us to cover the rift;
Only your notes are pure contraption,
Only your song is an absolute gift.

Just how absolute or pure music could be has been proved in our century by a development Hanslick only dimly foresaw: the waning of the tonal system. Hanslick does acknowledge the speed with which music "uses up" its forms:

Modulations, cadences, intervals, and harmonious progressions become so hackneyed within fifty, nay, thirty years, that a truly original composer cannot well employ them any longer, and is thus compelled to think of a new musical phraseology.

But he imagines, as anyone would have in 1854, that triads "will ever remain the indestructible foundation upon which all future development must rest." The first perception proved more accurate than the second: as a result of the harmonic innovations of Hanslick's arch-enemy Wagner, among others, triadic progressions, stable keys, and ordinary cadences began to seem "hackneyed." By the turn of the century it was possible to hear, in the music of composers all over Europe (Debussy and Scriabin, for example) such powerful motion away from tonality that triadic conclusions, when employed, seemed hollow gestures toward convention. And by 1910, Arnold Schönberg was announcing his abandonment of tonality in language which still echoes Pater's notion of a fusion of expression and form, and which sounds positively Romantic in its acknowledgement of an "inner compulsion":

With the George songs I have for the first time succeeded in approaching an ideal of expression and form which has been in my mind for years. Until now, I lacked the strength and confidence to make it a reality. But now that I am conscious of having broken through every restriction of a bygone aesthetic; and though the goal toward which I am striving appears to me a certain one, I am, nonetheless, already feeling the resistance I shall have to overcome; I … suspect that even those who have so far believed in me will not want to acknowledge the necessary nature of this development … I am obeying an inner compulsion which is stronger than any upbringing.

The revolutionary changes wrought by Schönberg and others exposed as never before the limitations of the metaphorical description of music as a language. As Charles Rosen explains in his little book about Schönberg,

The so-called "breakdown of tonality" at the end of the nineteenth century revealed to what extent this exterior stability [of the tonal system] was an illusion; more precisely, it was a construction that depended substantially on the individual works of music much more than a linguistic system depends on individual acts of speech. Music is only metaphorically a language; a single work of music may transform and even create an entire musical system, while no act of speech may do more than marginally alter language.

If an individual work of music may alter and even create "language," then the conditions for understanding it must—at least partially—be made evident in the work itself.

Thus, to adopt for a moment the famous distinction of the linguist Saussure, the relation between a parole, an individual speech act, and the langue, or larger system, is quite different in poetry and in music. A word in a particular parole, even if that parole is a poem, will seem to most readers to refer inevitably to its dictionary meaning in the langue. But in music, each parole, each piece, must establish within itself the "conditions for understanding it." This was actually as true for Mozart as for Schönberg, but the familiarity of the tonal langue implied by Mozart's parole made it possible to adopt the comforting but false assumption that the tonal system was a stable language in which individual pieces might be written. One could thus believe that even if music did not express words or feelings, it followed dependable syntactic rules, and thus had some kind of extrinsic referent. Atonal music, by depriving us of the comfort of triads and cadences, scalar and chordal relations, all that Rosen aptly terms "prefabricated material," forces us to confront and acknowledge the uniqueness of each piece. It fully validates Hanslick's claim that musical structure is musical meaning.

Hanslick's prophetic assertions, especially as validated by the ever more apparent autonomy of individual musical works, would thus seem to indicate a final unbridgeable gap between poetry and music. But in the very crisis brought on by the abandonment of "prefabricated material," Schönberg drew on poetic form to solve the urgent problems of length and organization. And when he finally solved those problems in a more purely musical way by devising the "twelve-tone" or "serial" method of composition, he employed some procedures closely related to older kinds of poetic construction: the retrograde, for example, one of the basic operations performed upon a twelve-tone row, is the "crab" or backwards ordering… in the music of Machaut, which Schönberg heard in an early revival and praised for its constructive craft, a craft precisely paralleled in the anagrammatic devices of Machaut's poetry.

In poetry itself, Pater's words have not lost their validity: many of the nameable movements in French and English poetry of the late nineteenth and twentieth centuries may be usefully described as attempts to attain various metaphorical versions of the condition of music, motions toward musical technique along various axes; the poets and theorists of these movements, though not all of them have understood music with real precision, have usually acknowledged their desire to make poetry more like it.

The Parnassian poet Théodore Banville, for example, believed that poets should recover intricate medieval forms—rondels, rondeaux, ballades, and villanelles—forms with such demanding technical requirements that content would become subservient to "the implacable richness of the rhyme." Since rhyme depends upon an accidental likeness between words, not a syntactic or morphological one, the French Parnassians and their English admirers were aspiring to the condition of music along the old axis of construction, the axis along which Schönberg would later proceed in borrowing formal hints from poetry. Writing a rondel necessarily entails choosing words for their rhymes and adjusting both syntax and content to accommodate words thus chosen; Swinburne's rondels provide some familiar examples of the possibilities and limitations of this kind of aspiration toward music, but it has been a method infrequently employed in the twentieth century, where the ideal of "sincerity" has made it difficult for poets to accept the constraints of intricate forms. But W. H. Auden, perhaps the preeminent virtuoso among modern English poets, not only mastered numerous older forms but invented equally difficult new ones, some of which show curious affinities with serial techniques; these poems demonstrate the continuing possibility for a technical influence between music and poetry, the usefulness of "contraption" even in a century whose characteristic poetic style is, as Auden admits, "Good Drab."

The much more important Symbolist movement placed even greater emphasis on sound; James Robinson argues that "Mallarmé went on from Banville to conclude that the value of poetry lay more in the sound of words than in their sense," but Mallarmé's position was much more revolutionary than Banville's: given a choice between suggestiveness and clarity, Mallarmé chose suggestiveness; he even attacked the Parnassians as "deficient in mystery," apparently because their poems sometimes named objects:

The Parnassians take something in its entirety and simply exhibit it; in so doing, they fall short of mystery; they fail to give our minds that exquisite joy which consists of believing that we are creating something. To name an object is largely to destroy poetic enjoyment, which comes from gradual divination. The ideal is to suggest the object. It is the perfect use of this mystery which constitutes symbol.

For the Symbolists, the poem becomes a closed system, whose elements derive their meaning as much as possible from their place in that single formal structure, as little as possible from their everyday functions as names of things. Small wonder that Paul Valéry described Symbolism as the "intention of several groups of poets (not always friendly to one another) to recover from music the heritage due to them." In its fascination with sound, its hostility to ordinary syntax, and most of all its attempt to make the poem a self-contained world, Symbolism was one of the most thorough and serious attempts in history to push poetry in the direction of music. The final, summarizing document of the movement is aptly entitled Four Quartets.

So argues Hugh Kenner, explaining how Eliot and Pound, both influenced by the Symbolist movement, developed in different directions:

One poet moved out of Symbolism, one deeper into it. Commencing from the post-Symbolist nineties, Pound worked his way clear of systematized suggestiveness until his chief point of contact with 19th-century French verse was Théophile Gautier of the direct statement ("Carmen is thin") and his most Symbolist procedure an isolating of single words, not necessarily English. Eliot… worked more and more deeply into the central Symbolist poetic.

But Pound's hygienic program for clearing away the Symbolist mists was the use of the "image" (later the "vortex"), which he defined as "that which presents an intellectual and emotional complex in an instant of time." In that definition we may still hear the voices of Hanslick and Pater, the idea of an instantaneous fusion of intellect and emotion in a single image. Simultaneity is the axis along which this kind of poetry—whether we call it Imagism, Vorticism, or (more simply) the procedure of Pound—aspires to the condition of music. Pound's fascination with Chinese ideograms, beyond their alleged visual expression, lay in the fact that one ideogram might be made be out of several others like a chord out of several notes. Joyce's Finnegans Wake, forging new words (new chords) from disparate linguistic elements, strains toward this kind of harmony, and Pound's Cantos, assembling the melodies of many times and tongues, strain toward a more contrapuntal simultaneity. Again Kenner has caught it: "A polyphony, not of simultaneous elements which are impossible in poetry, but of something chiming from something we remember from earlier, earlier in this poem and out of earlier poems, such is Arnaut's way—and such Pound's."

As Kenner's aside acknowledges, actual polyphonic simultaneity is impossible in poetry, but the results of Pound's search for an equivalent are often impressive, as are the results of Mallarmé's attempt to make the poem a closed system, or the results of Auden's formal virtuosity. Hanslick's claim that music is not finally a language is correct, but these poets (like many of their predecessors) have seen in the musical possibilities for simultaneity, density, and design attractive models for poetry.

Similarly, the greatest of twentieth-century composers, while convinced of the autonomy of music, not by any means cut themselves off from the expressive possibilities we associate with poetry. At the simplest level, we ought to remember that the revolutionary works of Schönberg's "Expressionist" period—the George Lieder, Erwartung, and Pierrot Lunaire—all employ texts, and that the Sprechstimme introduced in Pierrot Lunaire, while it dramatizes the difference between speech and song, acknowledges the expressiveness of spoken poetry by incorporating its shifting and uncertain pitch into a carefully designed musical setting. Schönberg was neither naive nor sentimental about the relations between music and poetry; many of his pronouncements on the subject sound the orthodox Hanslickian doctrine of autonomy. But his greatest scorn was reserved for music he took to be empty of expression; he was no cold-blooded mathematician.

The other giant among twentieth-century composers, Igor Stravinsky, chose a path often followed by poets: the renewal of past techniques. Reaching back past Romanticism, he regained some of the compression, wit, and irony of eighteenth-century art. But even when writing in a strictly diatonic idiom, Stravinsky manages to indicate that that idiom is a mask, a persona; he is arguably the most rhetorical composer who ever lived. Because older music was more frequently heard in the twentieth century than in the nineteenth, thanks to recording and greatly expanded publication, neo-classic procedures which had been available to poets for centuries—allusion, parody, burlesque—were newly available to twentieth-century composers, a situation brilliantly exploited by Stravinsky. But like Pound and Eliot, similarly eclectic poets, Stravinsky remains recognizably himself no matter what mask he assumes. The ultimate example of the power of his compositional personality to transcend convention is the marvelous twelve-tone music he wrote after the death of Schönberg, music in which he masters a new idiom just as deftly as he had earlier mastered the idioms of Gesualdo, Haydn, and American jazz. And if, like Schönberg, Stravinsky made severe claims for musical autonomy, he too continued to set texts. His Shakespeare Songs, for example, are at once explorations of serial technique and witty examples of imitative word-painting.

Indeed, the great danger for poets and composers in our time may be the tendency to respond to our recognition of the gap between music and poetry by exaggerating it, by withdrawing into isolation. The weakest twentieth-century music, the dead academic serialism of the 1950s, results from a rigorous working out of constructive principles and an ascetic disregard for expression; in aleatoric music, where some elements are determined by chance, this asceticism has become a kind of despair. The weakest twentieth-century poetry, the self-indulgent amoebic "free verse" of the little magazines, so values its supposed expressive "authenticity" that it abhors even minimal craft and construction. If there is one central lesson in the history traced in [Unsuspected Eloquence], it is that great music and great poetry invariably involve both construction and expression. The healthy and accurate recognition by modern theorists of the fundamental gap between the two arts—the fact that music has by nature greater constructive resources, and poetry greater expressive resources—need not mean that analogies between musical and poetic procedures are pointless. The pursuit of such analogies, whether true or false, has been a factor in the making of great works in both arts, and may in turn enrich our understanding of them.

Richard Wagner's enterprise, as Hanslick recognized, was not a pursuit of analogies between distinct arts but a thoroughgoing attempt to merge the arts in a Gesamtkunstwerk. In a pointed personal version of the myth of Beethoven as "tone poet," Wagner argued that Beethoven finally turned to texts, specifically to Schiller's "Ode to Joy" in the final movement of the Ninth Symphony, because he had come to recognize that only the word could give music the precise expressive meaning for which (according to Wagner) his earlier work had been searching. With reverent gestures in the direction of ancient Greek tragedy, Wagner declared that the music of the future must be combined with poetry and dramatic spectacle. His opinions … sometimes resembled those of the Renaissance musical humanists, though there is no evidence that Wagner knew their work; in any case, Wagner was a much more influential spokesman for these ideas than Bardi or Baïf: not only was he an able propagandist, but he was an important composer whose significant musical innovations suggested that his opinions should be taken seriously.

For Hanslick, however, Wagner's operas were a "violation of music by words," and the metaphor of rape seems intentional. In order to explain his opposition to Wagner without seeming merely personal, Hanslick constructs a careful and skeptical philosophy of music, a philosophy whose main points are negative: he denies the validity of a number of conventional ways of discussing music, insisting instead on an aesthetic perspective which acknowledges music's autonomy and discusses music in musical terms. Some of his main points reappear in the equally skeptical, negative, and liberating literary aesthetics of the so-called New Critics.

Romantic composers of program music, like the Musical Humanists of the Renaissance, couched much of their talk about music in the language of emotion; they spoke of the power of music to express a composer's emotions or to arouse a listener's emotions. Hanslick, while quite careful to acknowledge that "music operates on our emotional faculty with greater intensity and rapidity than the product of any other art", nonetheless insists that "definite feelings and emotions are unsusceptible of being embodied in music." He separates and rejects what the New Critics would call the "intentionalist" and "affectivist" arguments:

On the one hand it is said that the aim and object of music is to excite emotions, i.e., pleasurable emotions; on the other hand, the emotions are said to be the subject matter which musical works are intended to illustrate.

Both propositions are alike in this, that one is as false as the other.

Aesthetic investigations must above all consider the beautiful object, and not the perceiving subject.

Even though Hanslick claimed (perhaps disingenuously) that the principle of concentration on the object had already been established in the aesthetics of the other arts, it took many years for literary critics to achieve a similarly rigorous insistence on separating the aesthetic object from the responses of its perceiver or the intentions of its maker. And when Monroe Beardsley and W. K. Wimsatt gave those principles their most exacting formulation—in the essays on "The Intentional Fallacy" and "The Affective Fallacy"—they acknowledged their indebtedness to Hanslick by choosing as an epigraph for the latter essay a cutting reductio ad absurdum from Hanslick's attack on musical affectivism: "We might as well study the properties of wine by getting drunk." The essay on intentionalism contains no explicit quotations from Hanslick, but closely follows the opinions he expressed in passages like these:

The beautiful, strictly speaking, aims at nothing.

In music there is no "intention" that can make up for "invention."

The limits to which a musical composition can bear the impress of the author's own personal temperament are fixed by a preeminently objective and plastic process.

Neither Hanslick nor his New Critical heirs would want to deny that poets and composers have intentions, or that their works may affect us powerfully; they are united in recognizing that neither of these phenomena "possesses the attributes of inevitableness, exclusiveness, and uniformity that a phenomenon from which aesthetic principles are to be deduced ought to have."

Hanslick takes a similar position about pleasure; he acknowledges its occurrence but denies its relevance to aesthetic investigation:

If the contemplation of something beautiful arouses pleasurable feelings, this effect is distinct from the beautiful as such. I may, indeed, place a beautiful object before an observer with the avowed purpose of giving him pleasure, but this purpose in no way affects the beauty of the object. The beautiful is and remains beautiful though it arouse no emotion whatever, and though there be no one to look at it.

Here the most striking literary analogue comes in the Principles of Literary Criticism of I. A. Richards (1928):

It is no less absurd to suppose that a competent reader sits down to read for the sake of pleasure, than to suppose that a mathematician sets out to solve an equation with a view to the pleasure its solution will afford him. The pleasure in both cases may, of course, be very great. But the pleasure, however great it may be, is no more the aim of the activity in the course of which it arises, than, for example, the noise made by a motorcycle—useful though it is as an indication of the way the machine is running—is the reason in the normal case for its having been started.

For both Hanslick and Richards, pleasure is at least as unsatisfactory a basis for aesthetic principle as intention or affect; it is a by-product of musical or poetic objects they would advocate studying in a more purely analytical way.

But Hanslick's most important point was one which could have no literary analogue: the denial of past and present attempts to describe music as a language. In a passionate conclusion to a long discussion of this point, Hanslick points to the "mischievous practical consequences" of

… those theories which try to impose on music the laws of development and construction peculiar to speech, as in former days, Rameau and Rousseau, and in modern times the disciples of Richard Wagner, have endeavored to do. In this attempt the life of the music is destroyed, the innate beauty of form annihilated in pursuit of the phantom "meaning." One of the most important tasks of the aesthetics of music would, therefore, be that of demonstrating with inexorable logic the fundamental difference between music and language, and of never departing from the principle that, wherever the question is a specifically musical one, all parallelisms with language are wholly irrelevant.

This stringent denial of all metaphorical attempts to describe music as a language is central to Hanslick's stated purpose: the establishing of musical "autonomy." But advocates of poetic "autonomy," as we have seen, made much of metaphors comparing poetry to music; here we confront a semantic inconsistency at least as ironic and revealing as those conflicting definitions of "imitation" in the eighteenth century. For Hanslick and the twentieth-century musicians who came to accept his doctrine, musical autonomy meant an escape from the extraneous requirement that music "mean" something which might be verbalized; for Poe, Wilde, and Pater, poetic autonomy meant an escape from a criticism centered on subject matter, and one of the most common and powerful ways to express what that escape might mean for poetry was to talk of aspiring to the condition of music. Even Richards, whom we hardly think of as a pale aesthete, sounds very much like Poe in drawing a clear line between poetry and truth:

It is evident that the bulk of poetry consists of statements which only the very foolish would think of attempting to verify.… Even when they are, on examination, frankly false, this is no defect.

And in seeking to explain why T. S. Eliot's poems were misread by those seeking a logical or intellectual structure, he falls back on the analogy to music:

If it were desired to label in three words the most characteristic feature of Mr. Eliot's technique, this might be done by calling his poetry a "music of ideas." The ideas are of all kinds, abstract and concrete, general and particular, and, like the musician's phrases, they are arranged, not that they may tell us something, but that their effects in us may combine into a coherent whole of feeling and attitude and produce a peculiar liberation of the will.

Ironically, the analogy leads Richards into an affectivist account of music that Hanslick would surely have rejected: another example of the tendency for poets and critics marching under the banner of autonomy to encourage loose analogies to music, while musicians marching under a similarly labeled banner have discouraged descriptions of music as a language. This inconsistency is not merely amusing; it ultimately confirms a basic difference between the arts, as does the fact that Hanslick's position has become something like orthodoxy among twentieth-century composers, while the tough Hanslickian positions of Beardsley and Wimsatt have met a constant and ingenious opposition from poets and literary critics. To put it as simply as possible, most people cling more tightly to intentional and affective claims as ways of discussing poetry than as ways of discussing music, and they think of subject matter as a primary category for describing a poem. If you tell me that you have just read an interesting poem, I am likely to ask you what it was "about," and your answer may well include some remarks about what the poet was "trying to do" and how the poem moved you; if you tell me that you have just heard an interesting piece of music, I cannot so easily inquire about its subject matter, and your only satisfactory answer to such an inquiry would be to sing or whistle some remembered portion of the musical material. Nor would claims about the composer's intentions or the music's emotional impact on you tell me very much about the piece; its notes would remain "pure contraption." And even the coiner of that phrase, for all his respect for composers, recognized the limits on poetry's aspirations toward music, its ultimate duty to say something. In "The Cave of Making," he admits to the ghost of Louis MacNeice:

           I should like to become, if possible, a minor atlantic Goethe,
with his passion for weather and stones but without his silliness
  re the Cross: at times a bore, but,
while knowing Speech can at best, a shadow echoing
   the silent light, bear witness
to the Truth it is not, he wished it were, as the Francophile
  gaggle of pure songsters
are too vain to. We're not musicians: to stink of Poetry
  is unbecoming, and never
to be dull shows a lack of taste.

Poetry, by this account, is not truth, but its aspirations toward truth-telling are as much a part of its problematic identity as its aspirations toward musical purity. Auden is making at the level of the whole art the same point Hanslick made earlier at the level of fundamental materials: that notes are plastic and neutral, while words mean things, even if we wish they did not.…

When modern poets have sought to emulate music along constructive lines embracing various kinds of formal complexity, they have often wished to escape critical fixations on the politics, morality, truth, or "sincerity" of their subject matter. We have already seen how that kind of analogy to music helped Poe attack "the heresy of The Didactic," and in France, where Poe's influence was to prove most powerful, Gautier had declared as early as 1835 that "form is everything." A group of poets called the Parnassians, formed in the 1860s, built an entire aesthetic theory on that principle. Reaching back to Villon and Ronsard, as Schönberg later reached back to Machaut, Théodore Banville devoted himself to virtuoso forms ultimately derived from the troubadour tradition. The title page of his Améthystes (1862) describes the contents as "composed according to the rhythms of Ronsard," and since those rhythms were musical in origin, we may think of Banville as cheerfully writing new words to an old tune, insisting on the primacy of form in order to deflect critical attention from the contents of such lines as these:

Vois, sur les violettes
Brillent, perles des soirs,
De fraîches gouttelettes!
Entends dans les bois noirs,
Frémissants de son vol,
Chanter le rossignol.

Reste ainsi, demi-nue,
A la fenêtre; viens,
Mon amante ingénue;
Dis si tu te souviens
Des mots que tu m'as dits
Naguère, au paradis!

[Look, on the violets / Sparkle, pearls of the evening, / Fresh dewdrops! / Hear from the dark woods / Trembling in his flight, / The nightingale singing. / Stay that way, half-naked, / At the window; come, / My innocent lover; / Say if you remember / The words you said to me / Before, in paradise!]

In the troubadour tradition,… a similar concentration on form produced a similar freezing of content, but in Victorian England, where the moral criticism of Arnold and Ruskin was dominant, the issue was not the conventional emptiness of the first stanza but the "indecency" of the second, for which the perfection of the rhyme between "demi-nue" and "ingénue" was no justification. When Swinburne, among others, began to employ such old French forms, he did so (as Robinson puts it) "defiantly, with accusations of indecency and atheism exploding in his ears." Tennyson referred to the intricate stanzas as "poisonous honey stol'n from France," a phrase in which we may hear not only Victorian prudishness but a distant echo of eighteenth-century British attitudes toward seductive "foreign" music; this suspicion of continental "art for art's sake" was apparently still alive in 1964, when Auden glanced disparagingly at the "Francophile gaggle of pure songsters."

But the real aesthetic issue raised by Parnassian poetry is neither moral nor national: it is the poetic imbalance between form and content produced when a poet commits himself to a stanza requiring repeated rhymes on the same syllable. Under these circumstances, the poet must find not only a number of words for each of his rhyming sounds, but a structure of meaning in which he may logically use the rhymes he has found; as Banville's phrase about the "implacable richness" of rhyme suggests, formal considerations may alter or even control the poem's content. The metaphor of "organic form," plausible enough when applied to Wordsworth's blank verse, cannot apply to any rondel, though the successful rondel will minimize the strain between form and meaning. Consequently, such poems often work best when taking poetic form itself for a subject, as in this example of Swinburne at his best:

A roundel is wrought as a ring or a starbright sphere,
With craft of delight and with cunning of sound unsought,
That the heart of the hearer may smile if to pleasure his ear

                         A roundel is wrought.
Its jewel of music is carven of all or of aught—
Love, laughter, or mourning—remembrance of rapture or fear—
That fancy may fashion to hang in the ear of thought.

As a bird's quick song runs round, and the hearts in us hear
Pause answer to pause, and again the same strain caught,
So moves the device whence, round as a pearl or tear,
                       A roundel is wrought.

Here Swinburne's own craft and cunning make the rhymes successful; even "tear," perhaps the word most distant from the basic subject of the poem, is neatly worked in through the witty, almost "Metaphysical" play on its roundness. But the claim of the second stanza, that a rondel may accommodate any subject, even strong emotion, is not borne out in Swinburne's other poems in this form. The series on the death of a baby, in which the ingenuity of the form makes the expression of grief almost grotesque, exemplifies the problem. Here is the worst of those poems:

A baby's feet, like sea-shells pink,
 Might tempt, should heaven see meet,
An angel's lips to kiss, we think,
 A baby's feet.
Like rose-hued sea-flowers toward the heat
They stretch and spread and wink
Their ten soft buds that part and meet.

No flower-bells that expand and shrink
 Gleam half so heavenly sweet
As shine on life's untrodden brink
 A baby's feet.

Here the strain to rhyme is everywhere evident; the extended comparison to flowers is forced by the need to accommodate such implausible words as "heat" and "wink," which seem tenuously connected to a baby's feet despite the metaphor, and the superfluous phrase "we think" in the first stanza serves only to slip in a rhyme.

Our sense of a poor fit between the formal music of this poem and its sentimental matter is analogous to the discomfort we may feel in the church music of Haydn and Mozart, where a solemn text like the Kyrie is frequently set as a bright operatic quartet. But in that case, we can ignore the inappropriateness of the setting and enjoy the music on its own terms, while we cannot (alas) ignore the content of Swinburne's poem. The demands of meaning constrain the constructive ingenuity of poets, but also make possible their most dazzling feats of virtuosity; we are impressed with "The Roundel" because it succeeds where it might easily have failed. During the tonal period, harmonic demands provided similar constraints and opportunities for composers employing canonic constructions; the thrill of a Bach fugue lies in the way the strictly canonic lines nonetheless produce plausible voice-leading and clear harmony. An atonal canon, by contrast, free from the requirements of tonality, is somewhat one-dimensional. As Rosen puts it, "the ingenuity of design of Schönberg's canons may dazzle and charm, [but] the virtuosity … has vanished with the disappearance of tonal harmony."

If poets could escape from meaning as composers had escaped from tonality, they might gain a similar freedom of construction, while risking a similar loss of virtuosity. Not surprisingly, there were such experiments among the Dadaists: in 1913 the Russian Futurist poet Kruchonykh published a series of vowels entitled "Heights (Universal Language)," and a few years later Hugo Ball, leader of the Zurich Dada circle, published a number of poems containing virtually no words in any language. Here is one:

tressli bessli nebogen leila
flusch kata
ballubasch
zack hitti zopp

zack hitti zopp
hitti betzli betzli
prusch kata
ballubasch
fasch kitti bimm

zitti kitillabi billabi billabi
zikko di zakkobam
fisch kitti bisch

bumbalo bumbalo bumbalo bambo
zitti kitillabi
zack hitti zopp

tressli bessli nebogen grügrü
blaulala violabimini bisch
violabimini bimini bimini
fusch kata
ballubasch
zick hiti zopp

Much of the charm of this poem is purely musical. It has strong rhythm ("bumbalo bumbalo bumbalo bambo"), marked cadences ("zack hitti zopp"), variation ("flusch … prusch … fusch"); we might even describe it as an imprecisely notated, percussive, vaguely pitched piece of minimal music. But even this radical attempt to attain the condition of music cannot escape problems of meaning: it has a title; it includes at least one real word ("fisch"); its coinages (e.g., "violabimini") frequently suggest real words; and it even includes a poetic allusion ("leila" and "blaulala" sound like distortions of the song of Wagner's Rhine maidens, "Weialala leia"). It is at least as impure as a piece of program music; we are encouraged to seek for some kind of meaning, despite the obvious lack of syntax. And this "impurity" is inevitable, for meaning in poetry, unlike tonality in music, is not a system which can be replaced, but a central and inescapable element.

The contemporary inheritors of the Dadaists and the Parnassians, the OuLiPo group in Paris, seem for the most part to recognize this fact. Their leader, George Perec, claims to "reject the noble image of literature as a divine inspiration," preferring to believe that "language is a kind of putty that we can shape," but Perec's shapes, which include a 5,000-letter palindrome and a substantial novel (La Disparition) without a single occurrence of the letter e, have meaning; indeed, some reviewers of the novel failed to notice its technical trick. Another production from the same group, Cent Mille Milliards de Poèmes by Raymond Queneau, shows obvious affinities with serial methods in music. It is a book containing 10 sonnets, each cut into 14 one-line strips; by turning over these strips, the reader may produce 10 different poems, all making sense! Machaut and Compiègne would be proud, for this kind of production is more like their work than like that of Hugo Ball; its ultimate accomplishment is intelligibility.

Despite their ingenuity and charm, the constructive experiments we have been examining were not and are not part of the main stream of Western poetry, in which there has been, on balance, a turning away from virtuosity of all kinds. In 1968, W. H. Auden spoke at the opening of the Salzburg festival, acknowledging his great fondness for music and opera. He said in part:

… In the contemporary world, opera is the only art-form involving words which can still employ the High or Sublime Style. In days gone by, the poet could write in a High Style all by himself. This seems to be no longer possible. The characteristic style of modern poetry, or of the modern poetry I admire, is what Professor C. S. Lewis has termed Good Drab. It is a quiet intimate tone of voice, the speech of one person addressing one person, not a large audience: whenever a modern poet raises his voice, he seems phoney, like a man wearing elevator shoes.

The nostalgia for the High Style here is as real as the recognition of its difficulties. And if Auden seems to be thinking mainly of diction, we should remember that choice of words powerfully affects form and meter, as Edward Sapir argues in a seminal essay on "The Musical Foundations of Verse":

Whatever be our favorite theory of the nature of diction in poetry, it must be granted unreservedly that any lexical, grammatical, or stylistic peculiarity that is not current in prose helps to accentuate the rhythmic contour if only because the attention is more or less forcibly drawn to it.

Constrained by his perception of modern life to employ a vocabulary and grammar "current in prose," Auden was not unlike Schönberg, who was constrained by his perception of musical truth to abandon the tonality he loved.

But neither composer nor poet yielded to the formless.… [Schönberg] eventually won through to a music at once highly formal and uncompromisingly modern, in part through renewing old techniques, in part through inventing new ones. Auden's career involved parallel achievements. He could imitate Anglo-Saxon alliterative verse, Skelton, Swift, or Hardy; he could write short lines, long lines, or lines scanned syllabically; he made such forms as the ballade, the rondeau, the villanelle, the sestina, and the canzone his own without producing a sense of artificial High Style, using them to accommodate wryly modern comments on politics, theology, and sex. He was rarely guilty of an empty line, preferring the true virtuosity in which formal song enhances expression to the hollow chiming of the "Francophile gaggle of pure songsters." And in a century where, as Sapir complains, "only a very small number of possible forms have been at all frequently employed," Auden invented a number of stanza-forms and rhyme-schemes, some of them as intricately new as serial procedures in music.…

It is another striking irony that the most thoroughgoing modern attempt to push poetry in the direction of Hanslick's kind of autonomy was perpetrated by a group of poets devoted to his enemy Wagner. In 1887 a group of French Symbolists, normally a mild-mannered lot, actually engaged in a street riot when a performance of Lohengrin was forbidden by the police. As early as 1861, when Baudelaire published an admiring essay on Tannäduser, there had been interest in and enthusiasm for Wagner among French poets. Wagner's Lettre sur la Musique (also 1861), a greatly condensed and somewhat misleading statement of his aesthetic principles, seems to have been the main point of contact, though there were occasional performances of his works in France. Some poets traveled to Germany to hear the operas or meet the great man: Villiers de l'Isle-Adam was such an enthusiast, visiting Wagner in 1869 and 1870. And when Wagner died in 1883, his mythic stature was assured: in the next year Edouard Dujardin organized a new periodical to be called the Revue Wagnérienne and secured a promise of a contribution from Stephane Mallarmé.

Mallarmé's contribution, entitled Richard Wagner, Revery of a French Poet, appeared in August, 1885; five months later, Mallarmé distilled from its already suggestive imagery a dense and mysterious sonnet on Wagner. In these works, and in the poet's famous Oxford speech, Music and Literature (1894), we encounter in all its complexity the "simultaneous need for, and distrust of, music" which Bradford Cook has identified as an important part of the Symbolist movement.

There were real affinities between Wagner and Mallarmé: Mallarmé's interest in producing an all-embracing, dramatic work of art, thought of simply as "L'OEuvre," the work, suggests Wagner's grandiose concept of the Gesamtkunstwerk; Wagner's dissolution of tonality and embracing of a fluid and continuous chromaticism suggests Mallarmé's extraordinary treatment of French syntax. But Mallarmé seems not to have understood the drive toward mimesis and literary precision apparent in the device of the leitmotif, indeed, his praise of Wagner suggests a view of music somewhat closer to Hanslick's:

Now at last we have music which is obedient only to its own most complex laws, above all to the vague and the intuitive.

If the sentence ended at the word "laws," we would have a plausible statement of autonomist doctrine, but Mallarmé goes on to celebrate "the vague and the intuitive" as laws. This attitude, applied to language, lies at the heart of the Symbolist movement, as Hugh Kenner perceives:

Mallarmé and Valéry and Eliot felt words as part of that echoing intricacy, Language, which permeates our minds and obeys not the laws of things but its own laws, which has an organism's power to mutate and adapt and survive, and exacts obligations from us because no heritage is more precious. The things against which its words brush are virtually extraneous to its integrity.… By the end of the [nineteenth] century, in France, whole poems have been made "too subtle for the intellect," held together, as effects are, by the extrasemantic affinities of their words.

Ignoring the musical attempt to name or describe objects, the leitmotif, Mallarmé praised Wagner's music, and music in general, for its imprecise, evocative effects. He had attacked the Parnassians for naming objects in poetry, and he heard in music the suggestiveness he hoped to achieve in words. The speech on Music and Literature includes a typical statement of this theme:

It is not description which can unveil the efficacy and beauty of monuments, seas, or the human face in all their maturity and native state, but rather evocation, allusion, suggestion.

The Symbolist movement reenacts and extends the Romantic tendency to celebrate and emulate the supposed vagueness of music; Kenner contends that "Symbolism is scientific Romanticism," and the scientific thoroughness with which the Symbolists explored "extra-semantic" devices of sound, undermined syntax, and sealed off their poems from the exterior world of things bears out his claim. In all these activities, the Symbolists imagined that they were moving closer to music.

At the level of basic materials, Mallarmé laments the fact that the sounds of words are not always appropriate to their meaning:

I am disappointed when I consider how impossible it is for language to express things by means of certain keys which would reproduce their brilliance and aura—keys which do exist as a part of the instrument of the human voice, or among languages, or sometimes even in one language. When compared to the opacity of the word ombre [shadow], the word tenebres [darkness] does not seem very dark; and how frustrating the perverseness and contradiction which lend dark tones to jour [day], bright tones to nuit [night]! We dream of words brilliant at once in meaning and sound, or darkening in meaning and so in sound, luminously and elementally self-succeeding.

In poetic practice, this longing for a more musical language led Mallarmé, Valéry, and most of all Verlaine to an intense emphasis on sound. Like older poets, they employed imitative devices: in Mallarmé's sonnet on Wagner (of which more presently), the hero makes his entrance in a noisy fanfare of trumpets achieved by crowding the line with dentals and plosives:

Trompettes tout haut d'or pâmé sur les vélins …
[Trumpets high of gold fainted on velum …]

But Symbolist manipulation of sound moved well beyond mere onomatopoeia. In Mallarmé's mature work, subtle patterns of organization by sound are more important than syntax. These patterns are not complex and orderly constructions … but unique and particular creations in whose mysterious power to suggest and evoke Mallarmé believed intensely. It is difficult to provide examples because these effects, almost by definition, defy analysis. Perhaps the strongest evidence of their presence is the continued capacity of the best Symbolist poetry to move us, even when we cannot parse it. "Genuine poetry can communicate before it is understood," wrote T.S. Eliot in 1929, thinking of Dante, and in 1942 he applied that highly Symbolist principle explicitly to Mallarmé:

One of the most obscure of modern poets was the French writer Stephane Mallarmé, of whom the French sometimes say that his language is so peculiar that it can be understood only by foreigners.… If we are moved by a poem, it has meant something, perhaps something important, to us … If, as we are aware, only a part of the meaning can be conveyed by paraphrase, that is because the poet is occupied with frontiers of consciousness beyond which words fail, though meanings still exist.

In its resistance to paraphrase, Symbolist poetry attains one aspect of the condition of music. Its dependence on the specific sounds of deliberately selected words, its distaste for naming objects and making statements, its fragmentation of syntax.… its tendency to employ words and phrases with a gestural rather than a lexical significance—all the features that make it impossible to translate or paraphrase—produce a close poetic analogue to that feature of music which Hanslick perceived so clearly: that its sounds are not signs of exterior objects but objects in their own right. Thus Cleanth Brooks:

In the Mallarmean poem the words acquire something of the bulk and density of things; the poem is treated almost as if it were a plastic object with weight and solidity and with even a certain opacity. For the words are not signs, transparently redacting ideas. Instead they have acquired something like bulk and mass.

The kind of bulk and mass, we might add, possessed by notes and chords.

All these aspects of Symbolism appear in Mallarmé's sonnet on Wagner:

Le silence déjà funèbre d'une moire
Dispose plus qu'un pli seul sur le mobilier
Que doit un tassement du principal pilier
Précipiter avec le manque de mémoire.
Notre si vieil ébat triomphal du grimoire
Hiéroglyphes dont s'exalte le millier
A propager de l'aile un frisson familier!
Enfouissez-le moi plutôt dans une armoire.
Du souriant fracas originel haï
Entre elles de clartés maîtresses a jailli
Jusque vers un parvis né pour leur simulacre,
Trompettes tout haut d'or pâmé sur les vélins,
Le dieu Richard Wagner irradiant un sacre
Mal tu par l'encre même en sanglots sibyllins.

[The silence already funereal of a veil / Covers with more than a fold over the furniture / Which a collapse of the principal pillar / Will efface with the loss of memory. / Our so old triumphant exercise of the conjuror's book, / Hieroglyphs with which the crowd is excited / At propagating on its wing a familiar thril / Rather hide it for me in a closet. / Hated by the original smiling noise / There burst forth from master lights / To the stage born for their representation, / Trumpets high of gold fainted on velum, / The god Richard Wagner radiating a consecration / Ill-silenced by the very ink in sibylline sobs.]

The Symbolist reluctance to name objects is evident here. A dark theatre is evoked without being overtly described; Mallarmé's embarrassment about the state of poetry appears obliquely in the desire to conceal the hieroglyphic conjuror's book; and Wagner appears less as a sound piercing the silence than as a light piercing the darkness. The climax of the poem is unquestionably the word "Trompettes," but naming a musical instrument is the least important function of that word. The trumpets stand for Wagner; they delay still further the appearance of his name and stand in grammatical apposition to that name, which is the long-delayed subject of the poem's strongest verb ("a jailli," burst forth); they are golden in color, so as to participate fully in the poem's most basic opposition, that between darkness and light; their gold, in the poem's strangest construction, has "fainted" onto parchment, and the force of that fainting or swooning surely attaches to the speaker as well, for whom Wagner's appearance in a flourish of trumpets is an overwhelming experience. One might mechanically line up the imagery of the poem this way: Wagner's music illuminates poetry as the "master lights" illuminate a dark theatre as gold trumpets illuminate an old manuscript. But the first part of that formula remains entirely unstated, and the other images are not discrete, one-to-one allegory but interpenetrating, fluid symbolism. Our paraphrase does violence to the poem by attempting to restate what Mallarmé evokes or suggests.

But we paraphrase because we need to find meanings in poems, just as, at the level of the individual word, we cannot entirely ignore the ordinary significance of "Trompettes," even though that meaning is less important here than the unique symbolic meanings built into this poem, in which the trumpets represent Wagner, shine like gold, and burst forth from the stage lights. Symbolist poetry can effect this kind of shift or enrichment of a word's meaning because a symbolist poem, like a piece of music, is a closed system, in which context is a primary determinant of meaning. Thus T. S. Eliot, in Four Quartets, relentlessly repeats his central symbols ("fire," "rose") until we think of them primarily as compressed allusions to their previous contexts, not as names of elements or flowers. When the fire and the rose become one, in the last line of that poem, the effect is like a musical conclusion in which two previously separate themes are shown to be harmonically compatible. That musical effect, not some analogy by form to any particular piece, earns the poem its title. Once we get past the "file-card" approach to Wagner's operas, something quite similar can happen: we may hear a leitmotif less as a phrase arbitrarily designating some object or concept ("the sword," "fate") than as a phrase we have heard before in other contexts and are now hearing, aware of its past, in a new context.

Perhaps Mallarmé heard these effects in Wagner, despite his relative ignorance of musical technique. But even in the enthusiastic Revery there are uncertainties, jealousies, reservations. Mallarmé points to the "strange defiance hurled at poets by him who has usurped their duty," and to the way the music "penetrates, envelops, and joins [the drama] by virtue of the composer's dazzling will." And the conclusion, from which he drew some of the imagery of the sonnet, depicts the poet as climbing only part of the way up the mountain of abstraction, for a curious reason:

Oh Genius! that is why I, a humble slave to eternal logic, oh Wagner!—that is why I suffer and reproach myself, in moments branded with weariness, because I am not among those who leave the universal pain and find lasting salvation by going directly to the house of your Art, which is their journey's end.… May I at least have my share in this delight? And half-way up the saintly mountain may I take my rest in your Temple, Whose dome trumpets abroad the most extensive dawning of truths ever known and, as far as the eye can reach from the parvis, urges on the steps of your elected as they walk upon the lawns?

That Mallarmé, the chief celebrant of Symbolist aspiration toward music, should find himself, by contrast with Wagner, a slave to logic indicates that even he was finally aware of the limits of that aspiration. In the later speech, he describes literature as "our mind's ambition (in the form of language) to define things," and if the method of that defining, for Mallarmé and his inheritors, was a sonorous suggestiveness which owed much to music, the sonnet to Wagner, hazy as it is, achieves a degree of defining much more exact than that achieved by any piece of music—even a Wagnerian leitmotif. Eliot would later" … insist that a 'musical poem' is a poem which has a musical pattern of sound and a musical pattern of the secondary meanings of the words which compose it, and that these two patterns are indissoluble and one." This definition, if somewhat clearer than any of Mallarmé's utterances on the subject, remains fundamentally Symbolist in its concentration on sound and secondary meaning; by implication, it also recognizes that primary meaning, with its capacity to define things, is the area where analogies between poetry and music fail.

[A] composer gives meaning to an otherwise neutral or meaningless note by giving it a function in a system: in tonal music, he chooses a key; in twelve-tone music, he devises a row. Symbolist poets, seeking similar control over their words, loaded them with so many functions and meanings that their ordinary denotative functions were obscured, and became less important than particular functions taken on by words in a particular poetic context, a context held together by "extra-semantic affinities" of sound and suggestion. But when a tonal composer wants to alter the function of a note, he tends to isolate it, to state it without harmonic support…, while a Symbolist poet alters the meaning of a word by multiplying its secondary associations in order to drown out the dictionary definition, as Mallarmé turns "Trompettes" into an explosion of Wagnerian light. The Symbolist poet makes his words more flexible by turning them into simultaneities, polyphonic intersections of multiple meanings.

Twentieth-century advances in linguistics have made all of us more aware of the extent to which all words are simultaneities. A word has a history; it may once have meant something else; it may once have been more overtly or physically metaphorical; it has cousins in the great linked family of languages, brothers within its own. What the Symbolists were trying to do to words—to enrich them, to complicate them, to make them more flexible, to free them from one fixed significance—was an intuition of a feature now recognized as pervasive in language, a feature explored in all its complexity by a poet who treated all languages as a part of Language: Ezra Pound.

Symbolism was an important part of Pound's heritage, to be sure; it alerted him to these matters. Even when writing a brief for increased technical precision by poets, Pound makes a generously inclusive gesture toward Symbolist aesthetics:

… I do not by any means mean that poetry is to be stripped of any of its powers of vague suggestion. Our life is, in so far as it is worth living, made up in great part of things indefinite, impalpable.

But this genuine respect for the indefinite was affected, inevitably, by Pound's close study of troubadour poetry, which he understood far more fully than Swinburne did. In the same essay, "I Gather the Limbs of Osiris" (1911-12), he recognizes the "polyphonic" quality of Arnaut's rhyming technique, as well as the fundamental difference between collaborations in which a composer sets a poem and those in which a poet writes words for a tune:

It is my personal belief that the true economy lies in making the tune first. We all of us compose verse to some sort of tune, and if the 'song' is to be sung we may as well compose to a 'musician's' tune straight away.

By 1918, echoing Dante, he has become quite dogmatic:

Poetry is a composition of words set to music. Most other definitions of it are indefensible, or metaphysical.… Poets who will not study music are defective.

During the same period, he was also learning about Chinese poetry, a poetry in which each word was a visible combination of meanings, not adequately describable in the Western notion of "parts of speech." His mentor, Ernest Fenollosa, had written:

A true noun, an isolated thing, does not exist in nature. Things are only the terminal points, or rather the meeting-points of action, cross-sections cut through possible actions, snapshots. Neither can a pure verb, an abstract motion, be possible in nature. The eye sees noun and verb as one, things in motion, motion in things.

The word as chord in a fused, organic fashion: if Symbolism was scientific Romanticism, the Imagism with which Pound briefly flirted in 1912 was compressed Symbolism. Heinrich Schenker, penetrating to the essence of tonality as tonal music wanted, reduced sonata movements to statements of a triad; his analytical method distinguishes ruthlessly between "foreground" and "background" levels of composition, between decorative "passing tones" or "neighbor notes" and more essential or structural events. But Schenker's reductions appear as analytical charts, while Pound's similar reductions of Symbolism appear as poetry:

The apparition of these faces in the crowd;
Petals on a wet, black bough.

A poem distilled from a much longer draft. The Cantos would carry this process further in one way, using Chinese characters and brief phrases in Greek, Italian, and Provencal as allusive reductions of other poetic and historical worlds. Kenner's assertion that this "isolating of single words, not necessarily English" was Pound's "most Symbolist procedure" identifies the source of the notion of the word as chord, but also acknowledges that Pound was to move beyond that notion, though not away from techniques derived from music; indeed, it was by means of his serious study of music that he progressed from the chordal, separate sonorities of Imagism to the intersecting melodic and contrapuntal lines of the Cantos. His opinions about music were as definite and cranky as his opinions about literature, and as honestly earned. Learning by doing, just as he taught himself languages by translating their literature and carpentry by building his own furniture, he composed a sort of opera (Villon), attempted to play the bassoon (a perfect instrument for him, with its rich, often awkward assortment of overtones), and wrote in 1924 a "Treatise on Harmony" which Kenner has accurately identified as one of the important keys to the Cantos.

The most important assertion of the "Treatise" is its Bach-like insistence on the primacy of the horizontal line: chords are not thought of as units (as in Rameau) but as intersections of contrapuntal lines:

The early students of harmony were so accustomed to think of music as something with a strong lateral or horizontal motion that they never imagined any one, ANY ONE could be stupid enough to think of it as static; it never entered their heads that people would make music like steam ascending from a morass.

This opinion represents more than a prejudice against nineteenth-century chordal practice (though it is certainly that, as Pound's concurrent distaste for the piano and fondness for antique instruments suggests): properly understood, it explains why Pound eventually broke free from Symbolist practice to achieve a contrapuntal poetry. Chordal moments remain, as emphatic chords sometimes punctuate an otherwise linear fugue of Bach, but the Cantos are made from "strong lateral or horizontal motions," including the narrative line of the Odyssey, the sexual energy of Ovid and Arnaut, the correspondence of early American presidents, and the misguided economics of Mussolini. It is an index of Pound's achievement that no single sample can serve as a touchstone, any more than a few isolated bars can tell us much about a fugue; in both cases, the manipulation of the lines in the total construct is the central accomplishment.

Here as in other areas, Pound stands revealed as the quintessential modern, for twelve-tone music, relentlessly linear, recovers true polyphony from the chordal mists of Impressionism.… [Polyphony originated] in the combining of distinct tunes, and no music centered on chords as chords has been able to sustain real development: Debussy's harmonies, as Schönberg realized, were "without constructive meaning" and so served only a "coloristic purpose." Impressionist music, starved of linear motion, had to be outgrown, though even Schönberg profited from its lessons in the handling of color; Symbolist poetry, starved of linear statement, had to be outgrown as well, though Pound salvaged from it some of his most memorable techniques. James Joyce, in prose, moved in the opposite direction: Ulysses constructs a giant counterpoint of ancient myth and tawdry Dublin reality, but in Finnegans Wake those elements have been fused, at the level of the individual word, into chords. On the first page, "Sir Tristram, violer d'amores," arrives "from North Armorica … to wielderfight his penisolate war," and in one word we hear "penis," "penultimate," "peninsular," "isolate," and perhaps "desolate." Other Joycean coinages involve several languages, performing chordally the combinations the Cantos perform contrapuntally.

The contrapuntal method … is the more authentically musical, not only in terms of the past, but in terms of the present: according to Roger Sessions, in 1960, "composers seem by and large no longer interested in chords as such," since atonality allows any vertical combination and finds its grammar in the ordering of intervals, not in chordal inflections. It is also a superior literary method, as we may discover by comparing the rewards of rereading the Cantos with those of rereading the Wake: rereading Joyce, we will gain local explosions of understanding, many of them comic, as we hear more of the component notes in the chordal words from which the work is forged; rereading Pound, we will begin to hear the "long lines" of thought, argument, and design which cross each other in every canto. For Pound had no Symbolist distaste for statement, narrative, opinion; like the Goethe Auden admired, he wished his poetry were truth: "I have tried," he confessed at the end, "to write Paradise." …

Just as both arts need Romantics and rhetoricians, both need the renewal that comes from aspiration toward each other. Aesthetics is a notoriously imprecise and subjective area, and the relative consensus in the twentieth century about the distinctions between music and poetry is an example of progress; no composer I know of now advocates a return to program music, nor do poets aspire to a hidden medieval complexity of form. But the most characteristic modern problems in both arts may be described as withdrawals into their own conditions. Thus Ernst Krenek, an advocate of "total serialization," in which the composer abdicates his decision-making capacity and allows chance procedures to dictate pitches and rhythms, alleges that what composers have always called inspiration is really chance:

Generally and traditionally "inspiration" is held in great respect as the most distinguished source of the creative process in art. It should be remembered that inspiration by definition is closely related to chance, for it is the very thing that cannot be controlled, manufactured, or premeditated in any way. This obviously answers the [dictionary] definition of chance as "the absence of any known reason why an event should turn out one way rather than another." Actually the composer has come to distrust his inspiration because it is not really as innocent as it was supposed to be, but rather conditioned by a tremendous body of recollection, tradition, training, and experience. In order to avoid the dictations of such ghosts, he prefers to set up an impersonal mechanism which will furnish, according to premeditated patterns, unpredictable situations.

This is the counsel of despair, and its results, whether generated by humans or machines, are contemptible. Meaning in music is a profound mystery, not finally susceptible to verbal accounts, but surely the "body of recollection, tradition, training, and experience" that a composer brings to his work is a central part of the process that generates meaning. The success of composers like George Crumb in an eclectic, collage style that alludes explicitly to parts of that tradition is only one of many current attempts to leave behind the grimly academic music that resulted from the principles laid out by Krenek and his allies. Some have overreacted, alleging that only by a return to tonality can music regain emotional meaning; this is nonsense. Expressive power in music has survived other changes of the musical language … and in the works of Schönberg and Stravinsky, it survives the death of tonality. Nor is the twelve-tone system the only possibility: the stunning music of Krysztof Penderecki is neither tonal nor Schönbergian, but it achieves a dramatic expression of its composer's "recollection.… and experience."

On the poetic side, the equivalent problems result from an exaggeration of Auden's healthy perception that the High Style is no longer possible. Some poets have evidently concluded that no style is possible, that a group of words becomes a poem when its author declares it one. The notion that formlessness guarantees authenticity, the apparent principle behind much "little magazine" verse, is as serious an abdication of artistic responsibility as Krenek's reliance on chance. An art that began as one part of mousike cannot so abandon its past as to break entirely with construction. Again there is a possible overreaction, the allegation that only a return to rhyme and regular metrics can save poetry; but this notion, like the equivalent advocacy of a return to tonality for music, misses the point. If contemporary poetry has a serious need, it is the need for formal innovation, for new principles of construction which can help poets shape, order, and control the expression which is their high calling.

Finally, as I have tried to show here, analogies between poetry and music can help those of us who read and listen as well as those of us who create: by recasting poetic problems in musical terms, or musical problems in poetic terms, we may gain a fresh perspective. The naive student who demands of poetry a breathless enthusiasm, and so suspects any highly formal poetry of infidelity to feeling, may find it hard to cling to his reductive aesthetics if asked to apply them to music. And the advanced literary theorist may also profit from thinking about the relations we have been studying, as W. K. Wimsatt realized:

A theorist of poetry, being inevitably a person who subscribes to metaphoric and analogical ways of thinking, is in a good position to avoid both the fault of running the arts confusedly together by metaphors which become literal and the opposite fault of cutting the arts off from one another too completely by the denial of all such relations. There are poetic dimensions which can never be described except metaphorically, and to keep open, if only tentatively, the time-honored metaphoric avenues may do something to prevent theoretical discussion from declining into the literalism of a quasi-scientific semantics.

My purpose in this history has been to demonstrate the origins of those "time-honored metaphoric avenues" in the explicit intersections of the histories of music and poetry.