2—
Device:
Aspects of History and Structure

System and Structure of Rhyme

When a writer begins to perceive that his need mocks his gear he has already begun the change from versifier to mature poet.^[1] In an early letter (1909), Wallace Stevens defines this combined limitation and necessity of the poetic device:

In the "June Book" I made "breeze" rhyme with "trees," and have never forgiven myself. It is a correct rhyme, of course—but unpardonably "expected." Indeed, none of my rhymes are (most likely) true "instruments of music." The words to be rhymes should not only sound alike, but they should enrich and deepen and enlarge each other, like two harmonious notes.^[2]

When the correct device is also the expected one and by definition outworn, the act of composition will bristle with difficulties, with unforgivable wrong choices. The device itself will be parodied, distorted, or avoided in such a way as to make its absence very remarkable. After the "June Book," Stevens tends to avoid end-of-line rhyme because he cannot invariably strike with it the true, unprecedented note. (One result: he becomes arguably the most skillful technician of blank verse since Milton.) During this same, early modernist moment we find Mayakovsky joking:

Maybe,
             only a handful
                                       of unprecendented rhymes
             remain undiscovered
                                               in Buenos Aires.^[3]

Similarly Brecht, in "Bad Times for the Lyric," claims that in his song "A rhyme / Seems . . . presumptuous," and Marianne Moore calls for the allocation of only those rhymes "synonymous with gusto." In the most powerful way, with a new access of technical virtuosity, modernism renews the demand of the Romantic poets: not, or not often, for the abolition of the device but for its highest degree of transparency.

It is statements by the pseudo-avant-garde and the anti-avant-garde which raise the alarm that rhyme is being suppressed in the twentieth century. Acquaintance with the modernist and postmodernist demands for unprecedented rhyme, such as those just quoted, tends to quiet such fears and to confrim T. S. Eliot's more sensible reading of the relation of rhyme to modernity. There is "no campaign against rhyme," Eliot wrote in 1917:

but it is possible that excessive devotion to rhyme has thickened the modern ear. The rejection of rhyme is not a leap at facility; on the contrary it imposes a much severer strain upon the language. When the comforting echo of rhyme is removed, success or failure in the choice of words, in the sentence structure, in the order, is at once more apparent. . . . And this liberation from rhyme might be as well a liberation of rhyme. Freed from its exacting task of supporting lame verse, it could be applied with greater effect where most needed.^[4]

In the age Eliot helped to define by his practice, rhyme is obligatory only for those unimaginative, unhistorical poets who write advertising copy or squibs on newspaper opinion pages. More than in any previous period, poets know when now to use the device: how to make its absence significant, how to supply its want with other devices of equivalence, or how to turn the rhyme (when it is used) so to avoid the unpardonably

expected chime. These technical matters will be discussed, so I need here only reaffirm the level at which Eliot has addressed the question; for his splendid statement contains in germ a theory, to be unfolded here, of frustrated expectation as the motive power of the device in cognition and in history.

Is rhyme gaining or failing? Though the definitive answer would be of some interest, there is no way to provide this needed answer through a priori conceptual reasoning or empirical research. Rhyme, especially at line end, remains the most obvious form of equivalence. In combination with meter, this narrowly defined figure of sound has been the emblem of poetic language since Chaucer. So ingrained is this identity-in-difference of minimal pairs, rhyme has become another term for poetry, just as, in turn, poetry is a surrogate for the human studies more generally, the moral consciousness of the era. This doubtful chain of synechdoche has been the cause of alarm since the early modernist period, in gloomy books and articles which connect the survival of poetry and the humanistic spirit with the survival of rhyme and meter.^[5] Yet rhyme has neither gained nor failed, either in the statistical or in the qualitative sense, but simply changed. Preferring not to use the teleological terms progress and degeneration , I describe rather than judge each historical period's deployment of rhyme. But of course the description is not entirely innocent. Inevitably, because of the predilections imposed by my own history and my understanding of the gaps to be filled in modern poetics, the following historical-structural description of rhyme emphasizes the modern over the traditional, intellectualized over primitive rhyming, success over failure; and yet I try to give more than a little play to the alternative perspectives by bringing them in when I can.

In "The Chances of Rhyme," a poem as yet little known, though it is one of the superlative lyrics of its generation, Charles Tomlinson has written against "all those who confuse

the fortuitousness / Of art with something to be met with only / At extremity's brink, reducing thus / Rhyme to a kind of rope's end."^[6] Free play, randomness, risk taking in the use of the inherited devices of art, such is the artistic method of an avant-garde era. Tomlinson is right: "To take chances, as to make rhymes / Is human," and such chance taking defines the post-Romantic writer. My argument extends Tomlinson's theory of art and specifies further meanings for his title. I take the chances of rhyme to be in the first instance historical: the possibility of innovating the device in a period suspicious of all conventions. How did we come to a state of affairs at which a poet wants every rhyme to be his own invention, inscaped and selved? That meaning I call system , referring to the way poets choose to relate rhyming schemas to concepts of genre and decorum in a given period (e.g., satire seems to require clipped couplets). There is also an ahistorical meaning: the cognitive unpredictability of the device in its effect on the reader. How put a blade of logic between the rhyme-paired words to determine their degree of seamlessness and surprise? That meaning I call structure , referring to the sequences of acts of attention of writer and reader as these meet in a poem. These two notions of system and structure direct the inquiry as coordinative definitions.

A Short History of Rhyme in English

"And thou in this shalt find thy monument / When tyrants' crests and tombs of brass are spent " (Shakespeare). Rhyme in English requires that two or more positioned words have the same vowel sound in the last stressed syllable: with different sound features preceding, the same sound features following that vowel. That last sentence is as much as a prosodic manual might want to say about the device, and yet there is more to say, albeit speculative, about the semantic relationship between rhyming units, about the early desynonymiza-

tion of language (no doubt a fiction) which the device seems to imply. The discovery of a rhyme argues, that is, the human limitation of language, its dispersal into separated sounds and meanings, and its lucky economies when, as J. S. Schütze said of rhyme in 1802, two meanings find the same sound.^[7] Might the rhyming words be related, somehow, in convergent meaning as well as similar sound? Some have wondered: Wimsatt, Hugh Kenner, and John Hollander have written on the phenomenon of semantic convergence of rhyme words; a linguist describes families of rhyming words; in a recent poem with rhymes of a planned randomness, Charles Tomlinson says that the chances of rhyme are "like the chances of meeting— / In the finding fortuitous, but once found, binding."^[8] That difference-in-identity of minimal pairs, batched at measurable intervals, creates a principle of cognitive eventfulness—except in rhymes like "proved-loved," Shakespearean, automatic, thereby strictly unworkable. (This is not to deny that "proved" and "loved" had a quite exact pairing of vowel sounds in the pronunciation of Shakespeare's time.) These extinct rhymes evoke not expectation but perceptual automatism in the reader, and I have called them Shakespearean as a rough way of suggesting that a poem's defeat of automatism—its making strange of theme and rhythm—is successful in cognition only because of a prior act of historical imagination. Bad rhyming comes of following routes closed off by literary history: not knowing what is prosodically possible in a given period, thus mistaking archaism for innovation. Examples follow in this chapter.

My study of device and modernity centers on the rhyming practice of the past quarter century in England and America. It hopes to subsume the knowledge of the whole history of the device which recent usage inevitably implies. (For the effective tradition at this moment is what we know of the tradition.) My version of recent poets' knowledge of tradition follows in this section; my aim in undertaking this history of

rhyme is to understand the origins of the present state of affairs where rhyme is no longer an appropriate synechdoche for the poetic fact. Richardson, Saintsbury, Lanz, and Pendlebury have performed the history of the progress of the device through traditionally defined period styles, usually associated with centuries.^[9] Now that we may rely on their valuable work, it is time for another history of the device. My survey describes a process that, within my own terms of reference (learned from such as Tynjanov and Shklovsky), constitutes literary innovation: namely, the era-by-era exposure of the device as a device. I shall not often describe how the device is abandoned, though that of course happens on occasion; rather I hope to show how a literary group or generation, or a single great writer in his or her work, regains the ability to use the device at will. Writers, so I argue, do this by unmasking the device, showing it to be merely what it is, and thereby dispensable. There is an objection to doing history in this fashion: it can be said that we find in the period previous to the epistemological-prosodic break evidences that prefigure an avant-garde intelligence, seeing thus, in the past, only ourselves. One response is that only after the epistemological break, and by virtue of its emergent historicism, are we able—with interpretive categories of some strength—to segment processes of change and variation before Romanticism.

Rhyme, a device of bold patterning, is an instance of that "shape as superinduced" Coleridge opposed to "form as proceeding." Its use in a given period will be a test of reigning conventions of traditional shapeliness. This has not always been recognized. The presence of a valid unrhyming verse since Whitman makes necessary, in fact, a form of history of the device which refuses to take the device as a neutral object.^[10] There is no time in the history of rhyme when this device was not a matter of contention, whether in practice or in prosodic theory. There are conflicting poetics of rhyming and rhymelessness, or of one type of rhyming as against

another. Usually the debate is between different sectors of a period's writers, but occasionally—as with Campion, Milton, Auden—one finds it in the work of a single writer. Incipiently the debate is present in any writer's work, and if one cares to look for them, there are hints of a fit of rhyme against rhyme in nearly every poet. Such contention seems necessary to the historical continuance of the device.

In a rhyme-rich language like Italian, there is perhaps too much that is unpardonably expected, so with major writers like Montale and Quasimodo, rhyme comes into very great disrepute. English is not rhyme poor: the code of the language permits and encourages rhyme, and yet imposes strict and conventional limits on the range of likely homophony. For example, it seems rhymes of the "prove-love" variety come into currency only by tacit agreement of writers needing the resource and willing to permit "eye rhymes" into the canon.^[11] Around this device, as around all others, there has developed a structure of convenient fictions which define its expressive capacities in a given period. These fictions can to some extent be specified. To compile a nearly complete record of prominent devices and their use in the present century would require voluminous quotation but would satisfy our craving for an explicit poetics. In the absence of such a document, and lacking any explanation from linguists as to why some literatures like Spanish or Old French can employ terminal assonance instead of rhyme, it seems imprudent to keep calling rhyme the differentia of prose and verse. This way of thinking is very likely another venerable convenience twentieth century practice has brought into doubt; and yet any account of the origin and use of rhyme in English must show the essential descriptive-normative role of such a concept up to the generation of Whitman, and indeed beyond. There are reasons for believing that far from being merely the "stub" of the line, as Vladimir Nabokov maintains,^[12] end rhyme since its first use has constituted the line and its meters. If in traditional verse,

rhyme and meter have in fact interanimated each other, it follows that without rhyme, lines will be built up and recognized by other methods, metrical and grammatical.

Though study of the sound systems of texts would yield good returns in any account of total meaning, we still lack a persuasive notation for the effects of sound which, even in mental performance, characterize the most intimate and speaking elements of a poet's verbal art.^[13] No wonder some still maintain seriously that rhyme is merely ornamental, an effect unrelated to meter, and that in any text sound is subordinate to sense. Again, because rhyme is still, in Harold Whitehall's words, "the most mysterious of all the sound patterns,"^[14] it is considered by many the unanalyzable quintessential feature—the most representative feature. Rhyme is seen both as absolutely detachable from and absolutely the essence of poetry. Such contradictory attitudes derive partly from the traditional mistaken notion that rhyme is not intrinsic to the line, an altogether simpler item than meter and therefore a diversion from the essentials of prosodic study.^[15]

In part, too, the difficulties are inherent in the device itself and its origins. Rhyme is, says Whitehall, "non-indigenous to most European literatures, and has apparently been acquired by us from some South Mid-Asian Semitic language by the process of acculturation."^[16] Rhyme was not a boldly prominent organizing feature in the ancient classical poetry of the West, and this fact conditioned its emergence in the modern vernaculars. The purely theoretical ictus of Greek and Latin quantitative prosodies was apparently a disincentive to rhyme. "As long as stress was ignored in verse, modern rhyme, which depends on it, was obviously impossible; it is recurrence of stress, not verse-ictus, which is needed for rhyme."^[17] The real beginnings of rhyme as we understand it, "harmonizing with stress-accent and supporting the rhythm of the verse, are to be found in the tenth century"^[18] —in Latin hymns, then vernacular hymns and leonine hexameters. So

the deep harmony of rhyme enters Europe at the same time as organized Christianity.

Since stressing and rhyming emerge in history at the same time, very likely in symbiotic relation, there is no prima facie reason to believe that rhyme has any less part in the constitution of the line than meter. "Nations who unite prose accent and arsis," according to one student of Latin hymns, "need to mark off their verses plainly. They do it by rhyme, the rhythmical repetition of letters."^[19] Now Georges Lote, the historian of French versification, says that rhyme in French as in Latin was "merely a means of helping syllabism by marking the end of the line."^[20] But even in a syllable-count measure like that of French, if rhyme marks the line's end, it must not merely help, it must constitute such a measure; meter, a principle of infinite expansiveness, cannot (in the period after classical prosody and before blank verse) brake itself . Lote admits as much, I think, when he remarks that in medieval hymns, rhyme "is a means of marking the end of the line in a way which would strike the ears of the hearers indifferent to quantity and to an accent of which the music took no account." If the hymnological hypothesis is correct, rhyme moves from Latin to the vernaculars in such a way as to implicate its end-marking function with the rhythm and syntax of the rest of the line. Perhaps, then, the end of the poetic line properly begins at the line's beginning.

Rhyme was destined to be more emphatic in English than in French because of its adding, upon the privileged position of last stress, the added prominence of harmonious sound. Perhaps with the lack of a regular principle of stress, French has from the start been less able to justify, or at least to emphasize, its rhymes; poetry in French reacted far more decisively than poetry in English to the early modernist war on devices, and now the fifty-year ban on rhyming and on syllable-counting measures seems to have become a simple condition of creation. From the language of Racine, of Baudelaire, rhyme

has more or less faded away as a literary device—perhaps for cultural reasons, perhaps because of the lack of a supporting principle of metrical stress. The device has been more tenacious in the language that received it as the result of French conquest.

There is not, as the result of Norman invasion, of change in the culture and language of England, of church domination of Europe, of the rise of vernaculars, any contention over the device of rhyme: this is the time of its establishment. Yet rhyme is very much an issue between the existing alliterative poetries in the native tradition and the end-rhyming poetries on the French-Italian model, which are emulated in the dominant dialect by Chaucer, for of course rhyme had existed in fraternal forms before the tenth century. In Old Saxon, Icelandic, and Anglo-Saxon, alliteration was structural, other rhyme incidental. Possibly the frenchification of English meant a movement of accent toward the end of words, or at least a diminishing of initial accent, making words more suitable for terminal-syllable rhyme. Despite the use of alliteration till the end of the fifteenth century in the North and West, rhyme on the last syllable of the last word becomes the norm in English since Chaucer. Assuming a social and literary superiority, Chaucer at least once mocks gently the alliterative verse of the provinces, and yet he also supplies in "Sir Thopas" a whole tale that exhibits the full range of ineptness in tail-rhyme verse (rime couée ).^[21] Chaucer once explicitly complains about the "scarcity of rhyme" in English, and elsewhere admits that now and then his lines "fail of a syllable." A recently deflectionalized language must sometimes have been difficult to rhyme into, and not even Chaucer could give English a display of virtuosity like the Provençal canzon or an intricate carrying metric like Dante's terza rima .

By writing in dominant-dialect English, with the end rhymes of Jean de Meun and Dante, Chaucer set the pattern of urbane literature for centuries. And yet end rhyme was but

one of several possibilities when Chaucer had just begun writing. The more various rhyme—rhyme in other places than the line end—seems to have been lost from the homegrown European literatures after end rhyme became the norm. Henry Lanz, a historian of this period's rhymes, rightly says that rhymes "gravitate toward rhythmically important places,"^[22] but he overemphasizes the line end as the line's critical spot. Beginnings and middles can be "ends" too, when they are breaking points of rhythm and syntax. Before the consolidation of end rhyme in the vernaculars, rhyming effects seem to have been more various and inventive than at any time since. The variegated usage of the past fifty years in English raises some doubt, at least, whether end rhyme is really the evolved and predestined highest form.

I am describing a moment of paradigm crisis and uncertainty. A poetic device fostered by the church in its hymns, and brought from France, is gaining influence over a strictly indigenous related device. For a time the two overlap, until the insurgent device takes over. The coexistence is also a form of debate, though never aggressively argued. Indeed so cordial are the relations between the differing prosodies, that on occasion they are employed in the same poem—either in the same line, as in Pearl , or in different parts of the same alliterative stanza, as here in Sir Gawain and the Green Knight (ca. 1350–1390):

An oper^[*]  noyse full newe nezed^[*]  biliue,
pat^[*]  pe^[*]  lude myzt^[*]  haf leue liflode;
For vnepe^[*]  watz pe noyce not a whyle sesed,
And pe fyrst cource in pe court kyndely serued,
per^[*]  hales in at pe halle dor an aghlich mayster,
On pe most on pe molde on mesure hyghe;
Fro pe swyre to pe swange so sware and so pik^[*] ,
And his lyndes and his lymes so longe and so grete,
Half etayn in erde I hope pat he were,
Bot mon most I algate mynn hym to bene,
And pat pe myriest in his muckel pat myzt ride;

For of bak and of brest al were his bodi sturne,
Both his wombe and his wast were worthily smale,
And alle his fetures folzande^[*] , in forme pat^[*]  he hade,
    ful clene;
  For wonder of his hwe men hade,
  Set in his semblaunt sene;
  He ferde as freke were fade,
  And oueral enker-grene.^[23]

The long lines in the body of the stanza are alliterative, but these modulate into the clipped lines of the closural "bob and wheel," also alliterative but rhymed as well. This wonderful stanza makes special use of its rhymes: the huge knight is first introduced and described here, making his entry into the poem; his most startling attribute, his color ("hwe"), reserved for the last word of the last line, clinched in with a startling rhyme ("oueral enker-grene," "entirely bright-green "). These intermediate forms of great beauty, as reinvented in the alliterative-rhyming verse of Gerard Manley Hopkins, suggest that certain prosodic resources remain in our stock of forms, genuine but dormant.

After a time when Skelton's exuberant chain rhyming of short lines seems the most interesting usage, if not the most characteristic, the device of rhyme is exploited in many of its possibilities.^[24] Wyatt and the English Petrarchans of the early sixteenth century take special interest in rhyme as a tool for weaving together different lengths and quantities of lines. Now enters in prosodic theory, on the basis of an already remarkable achievement since Chaucer, the belief by Puttenham and Sidney and others that English "before any vulgar language . . . is fit for both sorts" of poetry, quantitative and accentual, unrhymed and rhymed (Sidney, Apology for Poetry ). Rhyme was, in fact, the one sweet new form of equivalence which the vernaculars had added to poetry (though Puttenham, searching precedents, recalls that some literatures before the Greek employed rhyme). So it was argued, except by those diehards who wished to expose the de-

vice because it had no classical sanctions and because, so they felt, it went against the very grain of the language. The antagonism against rhyme was in its origins a theoretical position, a nobly conservative and classical scholasticism in such writers as Theodore Drant, Roger Ascham, Spenser's correspondent Gabriel Harvey, and Richard Stanyhurst in his perverse translation of Virgil (1582). But the antagonism was also local and contemporary, in the sense that Elizabethan opponents of rhyme wanted to raise the dignity of English by shaping verse in elevated quantitative measures and were embarrassed at the pinchbeck chiming of rhymed verse. Yet after the 1570s antirhyme sentiment persisted into the era of the most magnificent flowering of rhymed poetry: deep dissatisfaction with rhyme was manifest in the development of dramatic blank verse and in the antirhyme essay and unrhymed lyrics of Thomas Campion. That Campion was, like Milton, a masterly rhymer, is not to the point. What interests is the need to expose the device and the polemical practice of the exposure, for these by an irony of history make a classicizing scholar-poet, Campion, one justification for prosodic experiment in the twentieth century. William Carlos Williams, especially, has admired Campion's pugnacious example, his search for rhymeless measures.^[25]

In the next phase of this history, the quarrel shifts from quantitative verse against rhyme to blank verse against rhyme. The curious violence of Milton's famous preface to Paradise Lost and his development of a flexible medium of blank verse for Christian drama and epic are the most notable features of the period after Campion. Yet others in different ways are plainly dissatisfied with the rhyming poetry of the period 1603–1660. Ben Jonson has "A Fit of Rime Against Rime":

He that first invented thee,
May his joynts tormented bee,
             Cramp'd forever;

Still may syllables jarre with time,
Still may reason warre with rime,
              Resting never.^[26]

The metaphysical poets, from Donne through Marvell, arrange unpredictable rhymes on taut, uneven-length lines; Abraham Cowley writes carefully asymmetrical stanzas in his influential Pindar translations of 1656; and we have in Richard Crashaw an amazing scheme of pet rhymes, where the poet exposes the device by working up his own system of rhyme-word associations, reusing rhymes as personal symbols.^[27]

The success of Shakespeare and Milton, both as rhymers and unrhymers, became for later generations a dilemma. The rhymed poetry was so various and rich as to make it strictly inimitable, while the great unrhymed achievement was in itself a sustained critique of rhyme. ("Milton, the unrhymer, / singing among / the rest . . . / like a Communist," shows William Carlos Williams identifying with yet another prosodic rebel.)^[28] Those conscious latecomers, the English Augustans, found a solution of authority and elegance. For forty years and more on either side of 1700 there was agreement that, thenceforth and forever, literary innovation meant repetition and correction of existing forms and genres. This was, of course, part of a project to fix the language of all communication: vocabulary, orthography, prosody. Exposure of literary devices, like all innovation, is now defined as correction; and the device of rhyme, used so wantonly by Elizabethans, will now be tamed and resolved into couplets. By that form of selective history which enables all poets to evoke their talent and protect a literary identity, the new style finds its precedents not in Donne or Shakespeare or Milton, but in Denham and Waller. Rhyme itself is not abolished, but merely renovated by regulation.^[29] In the first (and perhaps last) explicitly written poetics which has legislative force,

Edward Bysshe's Art of English Poetry (1702), permissible homophones are described, and, as an aid to writers, the appendix consists of one of the earliest rhyming dictionaries. Also the distances between rhymes, both of sound and position, are prescribed. Sound repetitions are to come at stated intervals, like perfect chimes: any unpredictability of sense or sound would be "harsh," or "rough," an uncivil prosody. Rhyme has of course its illogical, bardic or gothic element, perhaps most evident when it is used as the vehicle of a laughing irony. So for the Augustans there is all the more reason for controlling a device that—they perceived with restrained regret—was not present in the classical poetries they loved and translated. They had this dangerous device by precedent and it was not to be expunged, and so on the one hand they pretended rhyme was completely natural, assuming couplets were the mind's true order and discouraging other styles; on the other hand they controlled the device with the constraints of number and elision in the syllable-counted heroic line, avoiding similar parts of speech at rhyme position. Couplet rhyme, until the death of Dr. Johnson is, then, a peak of achievement gained by determined restriction. It is the emblem of vanquished difficulty and controlled violence.^[30]

The period style of English neoclassicism, from the start, was challenged in its uniformitarian prosody by Miltonic writers (Thomson, Cowper), Pindaric writers (Cowley, West, Gray), biblical writers (Smart), bardic writers (Gray, Blake), Spenserians (Shenstone), and balladeers (Percy, Chatterton). In the final quarter of the eighteenth century the heroic couplet seems a trick used by all, but used badly. No one comes to rescue the form or, after Johnson's death in 1784, to defend it in theory. The couplet seems increasingly inappropriate for the new meditative, topographical, or lyrical modes, and its decline, relative to the flourishing of the insurgent types just mentioned, helps to force a major prosodic break just before 1800. With the poems collected in Lyrical Ballads , the couplet

is devalued, and rhyme becomes one among several prominent devices, no longer the central device. So just when emphatic accent comes to modify strict syllable counting in the meters of poetry, rhyme loses its priority—and loses accordingly all blame for illogical, punlike disturbances of the semantic table of values. Very few explicit statements about rhyme are made by the Romantics, who accept it as part of a vast, complex inheritance. But they accept it on their own terms, since unlike their immediate predecessors they actively register the conventionality of the convention. Accepting the truth that rhyme is soaked in usage, their practice stipulates that, so far as possible, rhyme must express the very speaking tones of the writer. One thinks of the complexity of feeling in Coleridge's simple ballad stanzas; or one remembers the exuberant sound repetition, unthinkable a generation before, of Southey's poem imitating (through rhyme) "The Falls of Lodore"; or Shelley's decision to go for speed in terza rima, Keats's density in the sonnet or in the richly worked Spenserian and ode stanzas. Realizing that rhyme is a product of history, not of nature, the Romantics reinvent existing forms with a fine sense of how the device is limited and enabled by previous work: Keats goes to Dryden to get the sort of couplet he wants in "Lamia," Byron to Pope for the slangy tones of "English Bards and Scotch Reviewers." After the break at 1795, such a reinvention of the device, like its avoidance or its use in experiments, will of necessity express some degree of personal meaning. An unconscionable demand is made on the device, but one our major poets can tolerate: for over 180 years they have produced a highly conscious poetry in accordance with this demand.

In philosophy the historical break meant a shift from uniformitarian to diversitarian thinking, in critical theory a shift from mimetic to organic theories of the poem. In prosody the couplet gave way to a reshuffling of genres and to stylistic pluralism. After Wordsworth, and increasingly after Whitman, rhyme and rhymelessness coexist. In the work, for

example, of Wordsworth, Tennyson, Eliot, Auden, Lowell, Ted Hughes, rhymeless verse is abundant, yet these poets are best known as rhymers. Roy Fuller, a scrupulous rhymer for over thirty years, has published New Poems (1968) without a single rhyme, apparently seeking alternative constraints; while Edward Dorn, a poet very different from Fuller, whose early work is unrhymed in the school of Whitman and Charles Olson, has written wittily deformed rhyming quatrains as part of his long poem, Slinger (1975). Like the editorial decision to divide the poems of D. H. Lawrence by outer form into rhyming and nonrhyming sections, these are typical events in a period of pluralism. Poets now know that only when conventions are recognized as historically developed devices can they be used to the fullest extent as personal language. For these reasons, free verse, which displays modernity's exposure of the rhyme device in the fullest way, has by no means abandoned rhyme; free verse of merit has perhaps used rhyme as densely as many types of more traditional writing. Charles Olson has denounced rhyme in his essay "Projective Verse" (1950), but often in poems he relies on calculated deformations of the device. In fact, since Wordsworth, rhyme is not consciously a fighting issue on the part of those writers who have given it over as a legislative convention. Disputes will arise only when those who wish to prescribe rhyme for all poetry encounter the Romantic need to use poetic devices to throw a light on the conventionality of devices. We have seen how earlier eras have felt a similar need, but never so intensely, never as an organicist premise at or near the center of a whole literary culture.

Rhyme and Reason

More than in any previous age, post-Romantic rhyming is successful when it surprises. If earlier, in his Essay on Criticism , Alexander Pope could say that "True ease in writing comes from art, not chance, / As those move easiest who

have learned to dance," now the act and art of writing are pretended away, the skill of fitting together left unmentioned. So the poem appears a run of ordinary sentences, though with continual reminders from the structure of equivalences that it is art after all. The reader is led—only half-wittingly as it were—to the point of rest: the whole poem is preparation—not fully perceived till the end—for its own closure.^[31] The effect of the post-Romantic rhyming poem is at once casual and calculated, sensuous and logical.

However, just as the art sentences that employ rhyme avoid the marks of labor and resemble talk, so do we tend, with ordinary speech, to neglect consideration of the very sounds of the device which constitute it. Analysis of poetry has been strongly weighted, from the start, toward effects of logic, meaning, and sense: admittedly these effects follow the line of least resistance. The lack of a strong notation for sound repetitions indicates a disinclination to go beyond the signified to the actual verbal signifier and its special aural character; this in turn suggests that such "defections from the semantic system," as Roland Barthes calls rhyme, "cannot be directly explained by means of that system."^[32] Apparently the world of sound, with its elusive interior economy, is irreducible: interiority, says W. J. Ong in his essay proposing "A Dialectic of Aural and Objective Correlatives," cannot be resolved in terms of surfaces.^[33] For this reason post-Romantic poets and their readers do not share the confidence of most linguists that literariness inheres in prominent stylistic devices. Even the "impersonal" theories of poetry, so characteristic of the period from Matthew Arnold to Eliot and the New Critics, must finally assume the device will be reinvented from within in order to convey personal speech. The post-Romantic poem either resists or reinvents the rhetoric which has existed before the poem's creation. That is the poem's way of protesting its fate of being taken as an object, its way of enacting in advance the aural correlative.

It would appear that students have come at the device either by way of sound or by way of logic, but not with the combined perspective of form and sense together. Roman Jakobson does well to ask if there is a semantic propinquity to match the oral one: "Although rhyme by definition is based on regular recurrence of equivalent phonemes or phonetic groups, it would be an unsound oversimplification to treat rhyme merely from the standpoint of sound. Rhyme necessarily involves the semantic relationship between rhyming units."^[34] Below, demonstrating that one main meaning of rhyme involves the permutation of positioned sounds, I follow Jakobson by drawing out the implications of a statement he quotes from Gerard Manley Hopkins, who described the principle of verse as "speech wholly or partially repeating the same figure of sound." Before taking up permutation and position, I want to consider the limitations that follow any decision to treat rhyme mainly from the standpoint of sense.

An early description of the elements of sound may be found in Hegel.^[35] Rhyme, he noted, emerged in the European vernaculars just when Christianity became institutionalized on a large scale, a time when language became more intimate and soulful in the Latin hymn. Rhyme was instrument of a new meaning, a meaning of the personal part of the text. By a migration of forms the rhyme device supplanted, in the modern languages, that which was the poetic part of poetry in the classics. According to Hegel, the decay of the quantitative principle made it necessary for writers to find another sort of equivalence:

If the quantitative principle is renounced, and yet despite of it, but in accordance with the necessary demand of art, the sensuous medium is permitted to retain a certain force of resistance as against the exclusive assertion of ideal content, . . . there remains no other means left at our disposal save the express and artificially modulated sound of articulate speech as such.^[36]

The sound of rhyme is "speech as such": in art speech the sensuous element, the very breathing and pronouncing, is brought to the foreground of the reader's attention. With any poetic device there will inevitably be a "certain force of resistance" because the look and sound of patterned words emphasizes the material quality of language. Rhymes will be to some extent separable in analysis from the paraphrasable (or "ideal") content of the poem that employs them. If we resist any separation of rhyme from poem, of rhyme from line, we do so, I imagine, because at the final stage of ideal content, the relationship of signifier and signified will have been dissolved. There will exist at that thinkable but perhaps never realizable extreme, as Hegel plainly saw, not a text but a chunk of pure rhetoric.

Hegel's sentence is quoted by Henry Lanz, but in a context that changes its dialectical form into a polemical defense of a sound function "which is largely indifferent to meaning":

In ordinary speech, in prose, we entirely forget about the physical existence of words as signs or sounds. Meaning, ideas, is what we get for it. With their physical reality forgotten . . . the words become transparent . . . fully resolved into what they mean. Poetry is called upon to save the physical element of words and bring it to our attention in the name of art. For art and beauty require visible or audible forms through which the ideas may shine. Thus sound, the music of words, acquires an independent artistic value which is largely indifferent to the meaning or sense of it. Therefore Hegel writes [the sentence quoted above]. For rhyme is a means of increasing the sensuous intensity of words in contrast to their logical meaning.^[37]

Hegel, though, has not said specific sense does not matter, has not said the "signified" is in any way independent or irrelevant. Nor does Hegel offer warrant for figures of salvation ("save the physical element") or of translucence ("through which the ideas may shine"). Indeed, as I have shown, the whole period since Wordsworth and Hegel is unhappy that literary form may be in some degree institutional, indifferent to the speaker's meaning and personal voice. Literature, of

course, would not exist or be recognized as such without semiautonomous systems or signifiers, generic or formal shapes that can be transmitted from period to period, conventions that can be taught like artistic "languages." Nevertheless, while our literature would acknowledge that Ferdinand de Saussure's division of language into signifier and signified is itself a post-Romantic instrument of precision and discovery, our literature must also despise the artificiality of that separation. To note this reluctant dualism is to describe the essential modernity of the era beginning in 1795.

In a recent argument similar to Hegel's, Sigurd Burckhardt (1956) has noted the radical difference between words and the media of other arts: clay, marble, pigment, tones. Language, the poet's material cause, is already a medium before he starts to fashion it—though, says Burckhardt, a medium that "lacks all corporeality": the linguistic signs have only a "secondary, referential substance" since they are tokens of "real" things. Words

already have what the artist first wants to give them—meaning—and fatally lack what he needs in order to shape them—body. I propose that the nature and primary function of the most important poetic devices—especially rhyme, meter, and metaphor—is to release words in some measure from their bondage to meaning, their purely referential role, and to give or restore to them the corporeality which a true medium needs. To attain the position of creative sovereignty over matter, the poet must first of all reduce language to something resembling a material. He can never do so completely, only proximately. But he can—and this is his first task—drive a wedge between words and their meanings, lessen as much as possible their designatory force and thereby inhibit our all too ready flight from them to the things they point to.^[38]

Unquestionably. I only emphasize that different periods understand differently this "dissociative" or "divestive" function of poetic devices. Usually eighteenth-century poets preferred the least possible dissociation: they wished to regulate strictly the inherited devices of rhyme and meter, and to circumscribe metaphor. Puns and by extension rhymes, says

Burckhardt, deny "the meaningfulness of words and so call into question the genuineness of the linguistic currency on which the social order depends." In a time when the word as sign has an immense primacy over the word as entity, rhyme is harnessed and the pun falls into disrepute. Since 1795 there has been no comparable attempt to restrict the subversive, independent-of-things nature of the language medium; rather an exaggeration of it, almost a logomimesis.

Until 1795, rhetoric and poetic were (or seemed) coextensive—at no time more purely (and precariously) than in Augustan prosody and poetic diction. Since the shift at that time, when classical rhetoric and the mimetic theory of the poem were not abandoned but subsumed, devices have been understood not as deep structures of the mind but as historical, volatile, the perpetuum mobile of literature in the sense that they are reinvented again after lapses of time. It has become one of the tasks of modern poetics to explain how the inevitable device helps a writer achieve a voice, a style, literariness itself, while under different conditions the same world of signs, once saturated in personality, must deprive poetic identity insofar as it preexists the poem's making and saying.

In his study of rhyme, Wimsatt has set himself the task of explaining the icon quality of poetic speech, which would make language into a material. Unwilling to let the device escape control of conscious reason, Wimsatt argues the primacy of sense in poetry. His aim, he says, is to develop the idea that sound features are the effects and outer signs of reasonable-logical determinants, the idea "that verse in general, and more particularly rhyme, make their special contribution to poetic structure in virtue of a studiously and accurately semantic character. They impose upon the logical pattern of expressed argument a kind of fixative counter-pattern of alogical implication."^[39] For Wimsatt, sound patterns are not detachable ornaments; nonetheless, they are strictly secondary in logic: "The art of words is an intellectual art, and the

emotions of poetry are simultaneous with conceptions and largely induced through the medium of conceptions. In literary art only the wedding of the alogical with the logical gives the former an aesthetic value." Accordingly, this argument considers the semantic element foundational.

One willingly grants to Wimsatt that the cognitive elements in the poem can exist without the aesthetic elements, while the opposite cannot be true in verbal art. The cognitive elements are the sine qua non, the "fixative" of the poem's sensuous materials. It does not, however, follow from this that the poem's meaning has chronological or logical priority. If the sensuous element, the physical words tangled in their literary patterns of equivalence, were absent, the logical element of the poem would be neither aesthetic nor even thinkable. If the discourse is to be literary each is the precondition of the other.

Very apt in this context is Craig La Drière's statement that "it is hard to say whether meanings or sounds more often initiate the poetic process, but there is no theoretical reason to suppose the primacy of either. The great fact is that, once the process is begun, all the elements have a priori equality."^[40] Is the sensuous body of language, as it appears in poems, nonlogical? Does the logical function remain unmodified by the physical shape and sound of the words? Can we trust the testimony of poets who say that chanced-upon rhymes have led them to unforseen semantic routes? I do not know the answers, but would agree with La Drière that we can "learn a great deal more than we now know about structures of meaning from the study of structures of sound."

My point is that, quite possibly, Wimsatt's position obscures rather than defines a working concept of literary style. Entering the debate as an antagonist of Lanz's caricature of Hegel, Wimsatt argues that rhyme is a nonlogical device, a device that gives language an unavoidable solidity, a feature it may not deserve. He is not happy with the possible equality

of sense and sensuousness in poetry, and that is why his preferred form of rhyme is the Augustan couplet, a form that reduced to the minimum (a glorious minimum!) all that was nonlogical and unpredictable. Insofar as Wimsatt's position affords an answer to Hegel, it proposes a reversion to pre-Hegelian theory and practice.

Having noted that English prosodists have little to say about the semantics of rhyme words (the usual emphasis is not on sense but on degree of likeness of sound), Wimsatt praises the Popian wit of those couplets "where different parts of speech rhyme in parallel lines." Rhymes using the same part of speech in parallel functions, as in Chaucer's nounnoun rhyme of "bokeler" with "daggere," he calls, by contrast, "tame rhymes," minimum rhyme, "only one step from homoeoteleuton." Now homoeoteleuton, or the presence of like endings that are not set out into verse lines, is an ancient category for the description of effects in prose; by developing this unwieldy term's implications, Wimsatt here maintains that "the difference between prose and verse is the difference between homoeoteleuton and rhyme." In another essay, "Rhetoric and Poems," this opposition of homoeoteleuton and rhyme is called "one of the most profound of rhetorical differences."^[41] In this aesthetic system, the distinction between prose and poetry is not one of essence, for poetry merely superadds a certain repertoire of devices. Yet paradoxically the kind of poetry most like prose turns out to be norm for all poetic rhyming.

Wimsatt finds Pope's versification the highest type of rhyming in English, the most complexly logical, especially in "chiastic rhyme, the most brilliant and complex of all the forms of rhyme variation." Doubtless the rhyming of noun with verb, pronoun with adverb, and the like, makes the reader work hard. The optimum instance of the theory, however, may not be Pope but Ogden Nash. Whether homoeoteleuton is so rudimentary a device as Wimsatt says, whether it is a prose device strictly, and whether its difference from

rhyme is "profound," are also matters for debate and inquiry. He finds "mere homoeoteleuton" in Chaucer and Byron. Their poetic practice, leading up to and then away rom optimum rhyme in Pope, would be closer to prose. Pope's optimum rhyme is a specially poetic writing, for Wimsatt, because more reasonable, in a definition of reason as semantic unlikeness.

Since I want to show that one sort of reason resides in effects of simple repetition, I would question the inference Wimsatt draws with the use of such terms as "mere" and "complex," the inference that homoeoteleuton is neither complex nor literary. With a meagre definition of sensory resistance as semantic unlikeness, he writes: "The more primitive and forthrightly emotional the poetry, the less it may demand the sensory resistance of verse nonparallel to logic. The more sophisticated and intellectualized the poetry, the more it will demand such resistance." Yet a great deal of poetry does not seem to fit these statements about poetic patterning. In ballads there is the refrain; in sestinas, the repeated end word; in villanelles and triolets, the recurring whole line; in Whitman's free verse, anaphora, the repeated line beginning; in American Indian poetry, the hypnotic repetition of words and phrases, a "rhyming" of thoughts and not of words. Wimsatt says rhyme of the highest sort repeats just the final syllable or syllables of last words that are semantically unlike; there must not be complete identity of exact repetition. He has defined his subject of study narrowly as one relation of a rhyme to a reason, disallowing homoeoteleuton as valid poetic device for the same reason that some persons object to blank verse and free verse, the relative absence of bold traditional patterning, of pronounced norm and variation. Homoeoteleuton he thinks to be without a norm; that is the meaning of his adjective mere .

Wimsatt, by intention, leaves out consideration of a positional difference, yet difference of placement does offer a perfectly complex alternate norm in post-Romantic poetry. The

same word in a different place has not all the same poetic aura. Exact repetition is in fact a variety of equivalence, not phonetic (though sound is involved) but locational. Homoeoteleuton, in its modern varieties, is in this account a variety of rhyme broadly defined, neither more nor less sophisticated than any other sort of rhyme. It is one kind of poetic device, neither more nor less "reasonable" than any other, with a special effect that depends on the way the poet sets out the poem and the way the reader takes it in.

Like Wimsatt, I should reassert the equal importance of the sensory and logical functions of rhyme. Yet in his special emphasis on the logical function, making it foundational, he at some points in the argument underestimates every other variety of rhyme before and after the time of Pope. Consider the parallelism and antithesis in this passage from Randall Jarrell's "A Well-To-Do Invalid," where the figure of sound is indeed a figure of logic, yet we have no conventional end rhymes whatever:

When you first introduced me to your nurse
I thought, "She's like your wife." I mean, I thought:
"She's like your nurse—" it was your wife.

She gave this old friend of her husband's
A pale ingratiating smile; we talked
And she agreed with me about everything.
I thought: "She's quite agreeable."
You gave a pleased laugh—you were feeling good.

She laughed and agreed with you.
                                                            I said to her
—that is, I didn't say to her: "You liar!"^[42]

The whole poem (in which the nurse-wife dies first, leaving the invalid "well with grief") is a textbook instance of the relations of parallelism to reason. Repetition is its organizing principle, and the enabling rhetoric of its ironies, for every return of a significant word or phrase undercuts the first

utterance. "Her dishonesty is so transparent / It has about it a kind of honesty." Semantic likes, suffering perhaps here and there the slight permutation of an added prefix, withdraw everything given, in a sustained fit of cruel reasoning. The effect is sly. It is gained by homoeoteleuton.

In free verse and the prose poem, with their unspecified ends and middles, there is far less obvious pattern than in conventionally rhymed poetry. But equivalences are not absent. Neither Pope nor W. K. Wimsatt would, I suspect, approve the extreme measures taken in Mallarmé's "Un Coup de dés," whose unit is not the word or line or even the page, but the foldout. One must imagine the binding fold running down the center:

COMME SI
    Une insinuation              simple
               au silence             enroulée avec ironie
                                                                             ou
                                                                                 le mystère
                                                                                                    précipité
                                                                                                                     hurlé
dans quelque proche   tourbillon d'hilarité et d'horreur
                       voltige   autour du gouffre    sans le joncher
                                                                                                  ni fuir
                                                                      et en berce le vierge indice
                                                                                                      COMME SI^[43]

Top left and bottom right this "unit" is bracketed by the "COMME SI" phrases ("as if"); and viewed as a whole resembles a feather, or quill pen as image of the poet's labor. Temporary, fleeting rhymes ("en berce le vierge") are brushed by as the passage finds no completion but rather syntactical and metaphysical suspension of the second "COMME SI." The mind works over a large body of syntax to get down to that second "as if," which is at once a pause and a promise of more to come in the poem's movement, over the page. The

chances of rhyme, of repetition of any sort, are fragile in such a setting, and yet they are there, in the play of like-sounding syllables ("précipité-hurlé"), and in the upper case of the bracketing phrases. Tipping down and right, the page foldout resists visual balance. The two "COMME SI" phrases are transitional, leaving the vast stanzalike page open at either end in syntax and logic. Rhymelike this instance of exact repetition performs the musical and logical functions of rhyme, over a distance of syntax which may well be forbidding, but then Mallarmé is hoping to extend the mind's cognitive powers with new patterns of pagination, white spacing, sentencing, and sounding.

Dylan Thomas gives no such help in his (superficially more approachable) introductory poem in Collected Poems (1957). This is over 100 lines long, rhyming inward from the first and last lines to reach a perfect contiguous couplet in the exact center. Since the distance tolerance of the mind for rhymes permits at best an intervening passage of, say, ten lines, Thomas has deliberately violated traditional rhyme schemes and ordinary linguistic capabilities. Over that distance no reader is willing to see if rhyme pairs are semantically like: certainly not the reader with Augustan tastes, whose tolerance for the spread of prosodic pattern rarely extends further than two lines.

Expectation

Any account of poetry as a reiterative figure of sound must provide a sketch of how a reader's frustrated expectation may bring aesthetic pleasure. By what laws are the recurring sounds held apart? All verbal behavior uses two basic modes of arrangement, selection and combination. Selection of rhymes is based on lexical equivalences of similarity and dissimilarity (similar word endings, dissimilar word beginnings), and of synonymity and antonymity (same or opposite, or at

least different, meaning of the rhyme words). The combination in rhyming poetry, the build up of the sequence (in Roman Jakobson's words), is based on the syntax, the coming-one-after-the-other of parts of the sentence. Typically and traditionally, the rhyming words, selected because homophones, are placed at line ends and mark off lines for the ear. In this way lexis and syntax are bound and overdetermined each by the other.^[44]

As a device, then, rhyme has to do both with the selection and the placement of the words that chime. Equivalence of selection of the words themselves as entities is in fact taken further, into the construction of the discourse, by certain additional rules. These rules of meter usually create lines of equal lengths: syllables are "equalized" in meter, so that accented syllables receive more or less equal stress, unaccented receive more or less equal unstress. (Formal meter requires only that we register difference between stress and unstress, and remains a strategy of "more or less" despite the fairly recent invention of a four-level accent system for the description of the prosody of English speech.) Since the relation of rhyme to reason is of combination as well as selection—rhyme exerting an influence on the arrangement of the rest of the line—rhyme and meter interanimate each other whereever they are found together.

From the perspective of human cognition of poetic language, many types of poetic shape are valid as they delay and resolve the energies of attention. Though there is a range of complexity, the sophisticated rhymes are no more or less reasonable or literary than the others. No matter how palpably different they may be in degree of intricacy of pattern, the rudimentary writings will also be termed complex because the precondition of literary equivalence is complexity. Any writer of art sentences separates his language from ordinary language by a conscious patterning, literature's version of the frame surrounding the picture or the framing silence before

and after the music. The attempt of writers after 1795 to snatch away this frame, and with it the concept of the device, is necessary, noble, and finally futile—except in the sense that this urge is the true motor of stylistic dynamism and innovation in the aesthetics of modernity.

It is now time to deal with examples that show how the sequence is built up by the permutation of positioned sounds, examples from less sophisticated writing which have implications for theory. Here I explore permutations over different distances without much concern for what has been called prominent or permissible in our manuals of usage. In all cases adduced, the figure of sound is recognizably a pattern, though not one which might dazzle a traditional prosodist.

Literariness becomes a kind of structural surplus by adding constraints and equivalences to ordinary language. Sometimes this poem-making process is a matter of rudimentary arrangement, segmentation, or transposition as in puns, spoonerisms, anagrams, or in one form of concrete poetry:

     Fowl
Fred owl
     Fox
Fred ox
     Fred
Fred red
     Harp
Hans Arp^[45]

Watching the addition of three letters and a space to "fowl," seeing an implication of the noun drawn out of it, to attend Ernst Jandl's "Names" is to be present at the creation of rhyme. Then the name "Fred" itself undergoes permutation and breakdown; after the first set of terms is exhausted there is a lateral sliding to "Harp," and a new, smaller set is begun on the same principle. The achievement is minor yet nobody will say it is unrelated to the effect of rhyme. The interest, though, of traditional rhyme is permutation over a greater

distance. In that considerable vacancy between bouts rimés , the poet is obliged to stuff or dilate middles. As to the reader: just when one has come through a stretch of language, one is again reminded, by the framing rhyme, that this is poetry. So long as the middles themselves were very definitely specified into a certain number and configuration of stressed and unstressed syllables, the eighteenth century loved this reminder given by the overspecification of line ends. The post-Romantics, Mallarmé and Jandl, have afforded extraordinary instances where word position and wordplay shade into rhyme, but most poetry will not risk such extremes of deviation.

Certain folk poems elide or wither the "middles" between rhymes, but not quite so far as in the Jandl piece:

I saw the ghostesses,
Sitting on the postesses,
Eating of their toastesses,
And fighting with their fistesses.^[46]

The anecdotal element is extended to provide an absurd pretext for the comic rhymes that fill the second half of each line. We have here a poem—admittedly not a good poem—where the alogical pattern, in Wimsatt's terminology, is plainly constitutive. Apparently all poetic sorts get the proportion they require of sensory to logical resistance, and in this playful instance the sensory aspects dominate the other. Here are two verses from the Midlands:

Tell-tale-tit,
Your tongue shall be slit,
And every dog in the town shall have a bit.

He got out of the muxy,
And fell into the pucksy.^[47]

In the first, each line doubles the length of the preceding line, yet all are pent to the ear, equalized by the rhyme. No matter what the line length, the rhyme has to thump, for it is the rhyme that creates the line; rhyme that marks off lines for the

ear and thus, in the absence of a quantitative prosody, helps build the text from the smallest unit to the whole utterance. The second of these is more interesting from the point of view of selection, not of combination, suggesting as it does that "minimal rhyme" is the basis and staple of the device: fraternal rhyme words on the same piece of grammar, as here, are more likely to be semantically related along the lines of word families.^[48] The rhyme on "muxy" is related to old words for the disagreeable, "mixen," "midden," and "muck," while "pucksy" (similar in being either dialectal or invented) intensifies a progression of ideas from bad to worse by repeating the nasty joke of the sound with the slight change of m to plosive p . The logic is crude, a simple antithesis. Alexander Pope would never venture such low dialectal sounds; nevertheless, far more intricately worked, such logics are often the groundwork of the Popian couplet as well.

From the point of view of combination, that variable usually neglected, I shall now consider a common utterance that accompanies, sometimes actually effects, performances. Jump-rope rhymes are said in chorus by twirlers, watchers, and the skipper, one line to each swing of the rope, and the rhyme itself usually commands the performer to step out of the game. Here are two rhymes recorded a century ago in Warwickshire and Shropshire:

Hickery, hoary, hairy Ann,
Busybody, overspan,
Pare, pare, virgin mare,
Pit, pat, out, one.

Ink, pink, pen and ink,
I command you for to wink,
Rottom, bottom, dish clout,
O. U. T. spells out,
So out goes she.^[49]

The equivalences are within the line in rhyme or alliteration as well as from line to line (in "Pare, pare . . . / Pit, pat" suc-

cessive lines open with two alliterating monosyllables). Yet only rarely, as in this from Northamptonshire, does this sort of utterance produce crossed rhymes instead of couplets:

Hinks, spinks, and devil winks,
The fat's beginning to fry,
Nobody's home but Jumping Joan,
Father, Mother, and I.
O. U. T. out,
With a long black snout,
Out, pout, out.

Now, though "pout" orders (or describes) the feelings of the skipper who is out of the game, the content of these poems is usually nonsense. The poems may almost be said to exist for their rhymes, which in turn, in their emphatic position, direct the game and command exits from it.

The children who play are not aware of speaking in lines, and yet lines there are, measurable by the inflexible rule of the rope's swing. This rhyme numbers the lines out loud as it goes:

Hickory (1),
Dickory (2),
Dock (3),
The mouse ran up the clock (4),
The clock struck one (5),
The mouse was gone (6),
O (7)
U (8)
T (9)
Spells OUT!

Almost all folk rhymes and jumping rhymes are, as here, within limits eclectic in meter and line length. The lines are all equivalent in time length, as determined by the swinging rope; but sometimes the words are sung out slowly with pauses, sometimes they are hurried, words jammed. The letter "O" is here equal in time with "The mouse ran up the clock," the sort of equality-in-inequality which is most often

solidified by rhyme. The rhyme rescues the variant meters too: encountering a saying that commands, "Butcher, butcher, kill a calf," we expect a partner that is also perfectly trochaic, but we get, "Hang him up, and eat half." The rhyme naturalizes it.

There is a more meditative and literary kind of children's poetry, where the child speaking as an individual (not as part of a jump-rope chorus, with its communal invention and performance) comes to terms with what he or she knows of the adult versification. This, by a seven-year-old American girl, is called "My Private Life":

In my private life, I think about things,
Like I was a Queen with Diamond rings.
And sometimes I think about trees and grass and flowers,
And lots of princesses locked up in towers.
And the only thing I like
Is living in my private life.^[50]

The first two couplets rhyme on nouns, "tame rhymes" in Wimsatt's terms. But "like-life" in the final couplet is remarkable, not only for the obliqueness of the sounds, one to another, but especially because the rhyme uses "like" as a verb, where at the beginning of the second line it is an adverb—homoeoteleuton shading into rhyme, a logical shift of no little subtlety. The title, when repeated as refrain in the poem's first and last phrases, gives the effect of global wholeness and closure. The author knows the difference between a masculine and a feminine rhyme, and though the lines are not strictly equivalent in length, their differences—even the thirteen syllables of line three—do not seem obtrusive because of the rhymes. Line lengths are apparently close enough; indeed one might argue a certain flexibility for the line to extend or contract as necessary according to the demands both of sense and of allocating rhymes. The poem with the help of rhymes, but not for their sake, involves thinking through beginning, middle, end. One does not want to say that the sensuous elements overwhelm the logical ones. Thus the child, having

read little or no poetry, but having heard it read, struggles to find sensible and acceptable types of equivalence in her acts of both selection and combination. It seems the passion for the principle of equivalence, for the poetic function in language, begins early in the history of the race and of the individual. One need not make especially strong claims for the resulting poetry from oral tradition or from children; such writing is perfect in its imperfection. Enough, here, to prize its hand-to-mouth rhyming as one mark of genuineness, and to note how the major poetry of nearly every age has reformed itself by inspecting the more direct usage of the basic devices to be found in the genres usually considered subliterary.

Allocation

In most rhyming poetry the device serves to articulate the major structural units; the "figure of sound" enables the "build up of the sequence." In composition and in the reader's developing comprehension of the text, rhyme, sentence, and line are mutually determinative. Rhyme, in gestalt terms, is figure on the ground of the sentence and the line, but at the same time it is ground for other prosodic or grammatical devices. Looking backward from the rhyme to its chiming word or words, the device is figure. Looking ahead to the different rhyme that follows next, unspecified and yet in form predictable on the basis of established pattern, the very same end word is in the cognitive sense a ground. From its line-end position the device can be taken either forward or back.

Considered as figure on the ground of the sentence and line, rhyme draws attention to its own prominence. Geoffrey Hart-man has studied the way the traditional device makes "poles" regress "to a line-end position . . . while the rest of the verse is inserted between these rhymed ends."^[51] In poetry, he says, the "elision of middle terms and overspecification of end terms" collapses the chain or sequence of the sentence. However not all poetry has the usual pattern, and in the time before free

verse "Milton's rejection of rhyme is related to Hopkins' freeing rhyme from its fixed terminal position and making the last first (in sound-shape, not merely line-place)." Even in that large body of traditional end-rhyme poetry—the corollary is crucial for Hartman—the strength of the end terms depends on our seeing or hearing the elided middle members of the chain. The poet will adjust the units of his lines, so as to arrive at the right time, metrically, syntactically, upon his rhyme sound. It means trimming, extending, attending to limits, thinking back from the end over the whole sequence, considering the device both as figure and ground. It is a commonplace worth testing that such "enabling restraints of formal bondage"^[52] help evoke the rest of the poem, developing compressed statement and linguistic innovation. In this section I inquire whether overdetermination at line end produces other sorts of overdetermination throughout line, stanza, and poem.

To return rhyme to its place in sentence and line will mean assessing the shape and strength of the middle terms. Some mild protest is in order against Victor Shklovsky's extreme statement about "filling the distances between rhymes with free phonic spots."^[53] He doubtless refers to the sense of randomness engendered by the clutter of words between the overleaping rhymes. The intervening words are free in the sense that they are exempt from the requirement to rhyme, and yet they are very much semantic-syntactic thralls unless the rhymes themselves are to be seen as detachable stubs or tags at line end. In fact, Mayakovsky seems right to speak of rhyme as the tightening nail, or even more strongly as the dynamite that explodes at the end of the line's fuse. Rhyme is a subeffect of poetic closure, and as such acts to resolve or reinforce other effects, metrical and syntactical, as the poem proceeds.

Linguists have devised a "depth hypothesis" to measure the amount of syntactical delay, or frustration, the ordinary auditor can withstand without losing the thread of the sen-

tence.^[54] Students of poetic devices may find this useful as they trace the distance back into line and stanza one has to go to find the syntactical justification for the word that rhymes. Within limits, it would appear that the further one has to go, the more fully justified the rhyme will seem. Such a method may be extended by correlating it with a notion of poetic register: the more meditative lyrical poems will usually push back the syntactical justification to the early part, even the first line, of a segment like a quatrain; while certain types of poems, such as comic verses, require less depth of syntactical regression in the justification of the rhyme. Here, the term "justify" may be understood in the compositor's sense as well as the logician's.

Poetry's figure of sound is recursive and involves a leading forward and back of recurring harmonies and significances. As the reader gets further into the sequence, more and more elements of the discourse become determined. The effects of chance are themselves diminished in the course of the utterance, as the types and limits of recurrence are settled. That, no doubt, is what Jiri^[*] Levy^[*] means when he says that the cause of rhyme's aesthetic effect is "unexpectedness of the specific sequences of sounds, based on a deviation from stochastic distribution."^[55] The chances of rhyme, that is, create a jolt in the reader when recurrent sound and significance interfere with the linear buildup of the sequence. Levy defines the three elementary principles of form in the art sentence, of value because they apply both to sound and to sense in the poem and because they are phrased as binary oppositions. These three principles are continuity-discontinuity, regularity-irregularity, and "regularity or irregularity in the arrangement of unequal units, i.e. a higher or lower entropy of the series . . ."; so defined, the three principles permit us to correlate the physical level of the text with the semantic level.^[56] The effective rhyme, defined in these terms, will confirm grammatical expectations, and yet, since homophony is impermissible in ordinary language, indeed indicates in that

context a kind of derangement, will at the same time frustrate semantic expectations. Teasingly, the device suggests a semantic link between words. Beyond the type of "reasonable rhymes" Wimsatt and others have discussed, this link can very likely never be proved, nor need it be.^[57] What counts is the way the device in its place will both confirm and unsettle expectations. Pushkin and his translator Walter Arndt, for example, know just what the reader likes and have their fun with him:

At last a crackling frost enfolded
Fields silvered o'er with early snows
(Alright—who am I to withhold it,
The rhyme you knew was coming—ROSE!)^[58]

The expected rhyme is given, despite its nonsense in context, and the poet shocks the reader out of his habitual associations of contrast, depriving him of the snows-rose cliché forever.

In admirable rhymed poems, there is a very high degree of unexpectedness in the intermittent shock when we discover a discourse, which seems natural, is in fact overdetermined by a line of devices of equivalence. How does the professional poet deploy these equivalences, hiding the labor as he crafts the line up to the rhyme, back from the rhyme? Consider the syntactical regression back from the final line in Robert Graves:

Impossible men: idle, illiterate,
     Self-pitying, dirty, sly,
For whose appearance even in City parks
     Excuses must be made to casual passers-by.^[59]

This workmanlike stanza takes one risk, the placement of "sly" at line end, precarious after the comma. This is justified by the submergence of vocalic effects (i sounds in lines 1–2), the dropping of rhymes off the ends of lines 1 and 3, and by the way "passers-by" almost diffidently picks up the rhyme. Lines 3–4 are built of a long appositional phrase, equal to "sly" in adding one last piece of information to the list of ad-

jectives, yet unequal in length. The syntactical justification for the rhyme word goes back at least to "For"—but eventually, I believe, to the colon in the first line with the potentiality implied there for the list that follows. If the example is accepted as typical, we must substitute for Shklovsky's "free phonic spots" another description by Juri Tynjanov, who speaks of "the crowdedness of the poetic lines." The poet may achieve this compression of meanings by sheer stretch of naked grammar, but more often will employ a phrasal sentencing, with apposition, clots of phrase, stops and starts. Most poets, rhyming in fairly short bursts, partition the line and poem into smaller subsets by caesura, juncture, and various effects of breaking off.

I have been arguing that sound and sense in literary composition determine each other reciprocally, no place more evidently than in rhyming practice. I have tried to show this, up to now, by using examples from fairly simple verse forms. Before further examples, and in order to schematize the theory of allocation developed thus far, I would like to consider the case of Chinese poetry.^[60] My attempt here, too, is to try to slow down and space out the analysis of the composition process.

Chinese, because it is morphematic and not alphabetic, partitions the line by ideograms and not words, gaining thus a more clipped rhythm than Western languages. Its very difference from the English manner of proceeding makes it valuable as a theoretical model for allocation. The use of end rhyme is exactly the same as in English, but the system of constraints is otherwise revealingly different. One highly common form of the eight-line-regulated poem contains, in each line, five ideograms; each of these is either deflected or level in tone. The rule is that deflected-tone and level-tone ideograms must alternate along the line. If this first line is to rhyme (with lines 2, 4, 6, and 8), it must end on a level-tone ideogram. Then, working back from the rhyme position, ideograms 2, 3, and 4

in the line are strictly determined while the first remains optional and may be either deflected or level. The model of Chinese shows, in a clearer way than in English, that the line as an entity progresses from indeterminacy to determinacy. This holds true of the presumed act of composition and also in the reader's version of that act when the line is taken in perceptually. If the line is to rhyme, the whole thing is determined all the way back to the second ideogram; only the first ideogram may be taken with any freedom between the two choices, and even then we have a normal mode and a variant. The Chinese poetic line is a string of binary oppositions: once the decision to rhyme is made for the first line of the poem, a whole series of other constraint operations comes into play. To choose to rhyme shuts off other options, while the choice not to rhyme provides slightly more free option in line 1. As one gets further into line or into sentence or poem, constraints increase in quantity and determining power.

Apparently Chinese, with its rhythms and excitements different from ours, cannot achieve the special expectation of syntactical delay or the pleasurable frustration of the English periodic sentence. With ideograms as equal units, juncture and disjuncture are insistent, but Chinese will not display the specific track of feeling of the Western languages, which do not so strongly employ separation of the parts of the line. There are in English, that is, more units (words) in a given line; therefore more partitions; and therefore the line is more possessed of continuity. The more you partition down, the more continuity you achieve. What results from this process may, however, be a spurious or rhapsodic continuity, of the sort one finds sometimes in a writer like Swinburne. Thus when, as with Pound, a writer wanted the laconicism of the clumped phrase, he consciously imitated, in English, the Chinese mode.

Meaning and organization are processes that occur in a medium and depend on that medium's properties. The linear

character of writing obliges us to measure, if only subjectively in ourselves as readers, what information theorists call the channel capacity of the receiver. The spans of immediate and long-term memory impose severe limits on the amount of information we are able to receive and remember. As readers, we will have psycholinguistic limits, historically and personally determined, upon our matching responses to a given writer's strings of language. One writer on the psychology of communication has argued that human memory can span seven items, plus or minus two.^[61] Without declaring what such an "item" might be in prosody, but rather remembering length of lines or of distance between rhymes in traditional poetry, we can at least say that poetry in all languages will develop a line neither too long nor too short for the active memory to find its proper play. Except in poems of the extreme avant-garde, lines have enough "crowdedness" of equivalence to keep continuous interest, but they do not extend the pattern-response capacity very far beyond normal endurance. Rhymes will seem bolder in short-line poems, because, in Geoffrey Hartman's term, there is less middle to be "elided." But in lines of any length, the compositional value of rhyme will be its ability to introduce a cross pattern against the system of stresses: rhyme must, as Henry Lanz says, "be something apart from rhythm in order to assist us rhythmically."^[62] Since counting stressed syllables in not a habit of natural language reception, we require as line marker a type of equivalence which does not get heard because it is rhythmically prominent, but because it is melodically pleasing. Thus in traditional poems, the actual unit of construction becomes the rhyme-linked pair, broader in its sweep than the single line of verse. Free verse, deprived of such a sweep of pattern, can, if it wants, give the line and its parts a more distinct integrity than traditional verse. Then again, as Jan M. Meijer has affirmed, some structural principles are dynamic forces. Once set in motion, meter and rhyme can go on and

on: "It is a structural principle other than rhyme that puts an end to the process."^[63] Meter wants rhyme to demarcate line ends, while rhyme craves some principle of substantive plot to halt the poem. The aesthetic structure absolutely requires the cognitive structure in order to realize itself: medium depends on meaning.

A highly developed constraint system can be shown to operate in texts whose form is prescribed, and whose very vocabulary is determined by arbitrary rule before the poem begins. In New Stateman Weekend Competition No. 2,213, "competitors were asked to compose a Shakespearean sonnet containing the following words or phrases: gentian, charisma, cradle, burgeon, cri de coeur , investment, helpless, alopecia, run of the mill, allocate, devices, runcorn."^[64] So length and number of lines, as well as disposition of rhymes into one kind of sonnet, is fixed. Roughly one-quarter of the poem's language is already specified: seventeen words in all, none rhyming. The best two solutions were:

When Runcorn's gentians first refused to burgeon
And what divided it from fuming Widnes
Flowed thick with chemicals instead of sturgeon,
Helpless to filter as nephritic kidneys,
Only some dotty poets really got
The sense of England's coming alopecia—
Trees allocated to the villa's plot,
To little rooms of glass and vine and freesia.
Most others, swayed by the charisma of
Investment, simply raised their  cri de coeur
For the still freer run of the mill. Their love
Of cash and devices made whole skies a blur.
Cradle of ugliness and muck, the North:
Why do I cherish all that you put forth?

Pope mocked Belinda on whose helpless nape
Her lover bold, despite the sylph's devices,
Committed with his scissors such foul rape
As should be starred with Coma Berenicis.

I sing a Runcorn Lady of the Game
Her charms her sole investment. Gentian eyes
Lured lovers; her charisma called; some came,
Run of the mill types; she'd been no great prize

From cradle days. But, when hair ceased to burgeon,
She allocated cash from funds not big
To have her cri de coeur  heard by a surgeon.
Bald patches? Alopecia! Buy a wig.

So should Belinda, shorn of crowning glory
Have born a Whig, not suffered Pope, a Tory.

In the first sonnet the octet is a single sentence. This sentence is a series of right-branching, or dependent, clauses. Every sentence unit is a subdivision of the next-higher unit. The length and complexity of the sentence is gained by embedding of clauses. In turn, this embedding is related to, perhaps caused by, the constraint introduced by the rhymes.^[65] Embedding of clauses is one of the phenomena of the intruded middle; the clauses fall like continual parentheses of thought, a recoil of the sentence upon itself to catch a detail or nuance before progressing forward.

We are either, it seems, running with the eyes and mind up to the rhymes, or away from them along the next line. Rhymes exert attraction toward themselves, then once past, repulsion; or propulsion along the ensuing line toward the next rhyme. I am describing the reader's experience of that elision of middles which interests other disciplines than literature.^[66] The couplet is the most memorable rhyme form, because of its maximal closure and its more regular arrangement of words as units of unequal length. Eighteenth-century rhyming tried with the strict couplet to revise all devices by multiplying, so far as possible, the elements of predictability. One difference between Pope and writers since 1795 lies in post-Romanticism's enormously expanded tolerance for effects of chance and the reader's frustration.

One of the less convincing solutions to the New Statesman task has this ending:

How can this metre—tell me, if you will—
Accommodate the phrase "run of  the mill"?

A good trick to perform what one claims is impossible, and that, in the act of claiming: it is of genuine metrical interest that the "of" in the last line gets emphasis by position and italics, and achieves the curious effect of seeming stressed and scudded at once. And yet the rhyme on "if you will" seems stitched onto the end of the line. Apparently rhymes are seamless only when the syntax leading up to them does not include effects of apposition or of parataxis or of interpolated parenthesis. The further back in the line or stanza one can trace the syntactic justification for the rhyme word, the easier it will be to afford the necessary slight shock of the unfamiliar. Here, as in the example from Robert Graves, the chances of rhyme involve our continually encountering in discourse the improbable effects of equivalence at regular or fairly regular intervals. These effects signal the aesthetic function.

By rhyming French or Latin terms with English ones, the two sonnets quoted in full create even more dissimilarity than in the case of mere grammatical unlikes; so the effect is still wittier ("coeur-blur"; "devices-Berenicis"). Other types of equivalence reinforce meanings in these sonnets:

Only some dotty poets really got
(internal rhyme)

The still freer run of the mill
(internal rhyme; imitative phrasing in "run of the mill," which skips a stress and suggests by prosody that freer attitude)

Bald patches? Alopecia! Buy a wig.
(trisected line; representation of dialogue; exaggerated juncture suggesting symptom, diagnosis, and treatment compressed into one line)

Have born a Whig, not suffered Pope, a Tory. (parallel or antithetical syntax; pun on Whig-wig; internal rhyme with "shorn" in previous line)

It might be said that in "Pope mocked Belinda," the rhyme on "not big" is awkward, because the final phrase inverts normal order of adjective and noun; but I think this is, like "lover bold" in line 2, a conscious oddity, deliberately prominent. It is the kind of phrase which trains irony on the subject in order to diminish it. In addition to meeting all the requirements of the competition, both poems have a very high density of various sorts of equivalence. For the sake of incongruity, I compare lines from a more relaxed convention that gets its effects not from connection but from disjunction, not from syntax but from juxtaposed catalogs. The equivalences here are those of simple antithesis, and the lines, from Thomas Hood, perhaps provide a very rough English equivalent to the Chinese manner of proceeding:

Now double entry—now a flowery trope—
Mingling poetic honey with trade wax—
Blogg, brothers—Milton—Grote and Prescott—Pope—
Bristles—and Hogg—Glyn Mills and Halifax—
Rogers—and Towgood—Hemp—The Bard of Hope—
Barilla—Byron—Tallow—Burns—and Flax!^[67]

The superimposition of two different and conflicting systems of rhyme and meaning gives to these last lines the effect of a Petrarchan sonnet: the contrast of ideal and trivial. Rhymes seem blundered upon, though of course they are carefully planned. There is no nuance or quantity of information which must be comprised in these lines at all comparable to the density or flow of the other sonnets. The tone of the catalog is desired, and what usually seems a defect—rhyme's very factitiousness—is here turned to a stylistic preference.

Under the conditions of the competition, successful poems were bound to come up with a density of style. The proportion of aesthetic information to the available containers is

high, of an extreme near to the situation taken to hyperbole by Anthony Burgess in his commentary on last-minute touches to his lyrics for the Broadway musical, Cyrano :

Telephone calls come through demanding fiddling modifications to lyrics. To an unchangeable musical rhythm, in twenty-four syllables, in two lines rhyming—ound, the following information must be imparted to the audience: It is fourteen years since my lover was killed in battle, and since that time I have been living in this convent, seeing the beloved as both dead and living at the same time, somehow a god of the season, but not really dead in the winter and resurgent in the spring, instead possessed of the winter's constancy, not grimness, as well as the loving liveliness of nature in the spring and summer, and clothed in the magnificence of autumn.^[68]

Perhaps because the two sonnets quoted are very special instances of highly overspecified relationships of end to middle, more than other rhyming poems they raise the question whether external constraints in themselves produce aesthetic density. In both sonnets, intruded middles have been strengthened, either to offset or to justify four features or restraints—the overspecified ends, the prescribed diction, the sonnet form, and the iambic ground pattern. It is hard to deny that in these examples there is a causal connection between formal limitation and multiplicity of poetic equivalences. Yet not all equivalences are of this highly specified nature. More random conventions of weakened conventional shape do exist, and from these emerge equivalences, which, though hard to recognize as such because of the obscuring of overt pattern, are equally valid. Modern poetry takes as its project the systematic blurring or dilution of the inherited constraints, and its poetics requires a description of the shapes that result, shapes weakened in the cognitive but not necessarily in the aesthetic sense.

Failure

The seasoned poet's task of allocating all the parts of the line is not essentially different in kind from the failed poet's. The

linguistic and cognitive differentials between literary poetry and inferior or nonliterary poetry are rather small. Slight differentials in quantity and placement make major differences in aesthetic valuation. This makes it hard to define the distinctive nature of literariness.

Yet the poetic devices, and the labor they imply, do seem highly obtrusive in the failed poem. When the egregious rhyme leaps out at us from the awful poem, we suddenly see how the meaning of this and all devices

lies in their relation to a structural whole of interrelated meanings and intentions. In breakdown, for a brief moment the meaning of the objects is lighted up, emerging directly. . . . A broken hammer at once shows what a hammer is.^[69]

If we conceive the poem, any poem, as a fabric worked up from the mutual interference of an aesthetic and a cognitive principle, it will follow that we should see each of these two principles trying to reduce the other to matter. To the extent that the aesthetic and the cognitive have not achieved a balance of forces, the not-fully-achieved work is a lump of words.^[70] As it happens the bad poet who rhymes is usually bad in a certain way. In the examples that follow, perhaps in all examples of failed rhyming, the cognitive principle is overpowering and reduces the aesthetic to language matter. The flagrant rhyme draws attention to its overdetermination. In the most laughable instances, one can always tell which of two rhyming words the poet who fails hit upon first.

Robert Lowell, like all translators of poetry, admits to a final inadequacy. All translation is failure in the sense that, as Lowell has said, sound effects are not

transferrable from one language to another. I know what Baudelaire's sound effects are like and I try to get something else in English. . . . The Baudelaire was very hard for me, just to rhyme. I first did them in blank verse, then tried to rhyme them. I really did countless versions, shifting, changing lines.^[71]

Those blank-verse versions were matrices, draft stages. It seems a particularly delicate stage, if one has decided that the

rhymes must be preserved. Rhyme must be justified structurally, not only back into the syntax of the line and stanza but also by reference to the exact level of diction in the original. Similarly in Walter Arndt's Pushkin, the Eugene Onegin stanza has an "ordered interplay of masculine and feminine rhymes, and three differently patterned iambic tetrameter quatrains with a couplet at the end."^[72] Arndt speaks of his work of rearrangement and omission: "Predictably, most occur at the end of lines." We are reminded by this that rhyme is a prominent effect; perhaps more than the meter itself, rhyme influences the character of the entire utterance. Arndt believes Vladimir Nabokov's Onegin translation, which omits the rhymes, fails to do justice to the original. The literal line-for-a-line method, he argues, is a form of naive realism about the nature of literary style:

Poetic utterance is not produced from some underlying, neutral, merely cognitive statement by linguistic manipulation; and if it were the manipulations could not be the same in language A as in language B, or else they would be the same language. Hence sequential literalness becomes worse than irrelevant.

One would have to show, as I cannot, that Arndt gives more of Pushkin than Nabokov. Clearly Nabokov's version is premised upon a fine disdain, a belief that the sound effects are not to be captured, so not worth botching; whereas Arndt, like Lowell, believes the translator represents, not reproduces rhyme in the original.

Though all translators fail, nonetheless the adequate ones are always in the first instance adept readers of texts in their native languages. This permits them to see that, as Arndt says,

the proper formal frame of accuracy, i.e. the largest allowable unit of form within which maximum fidelity must be achieved, is a delicate matter of balancing the poetic pulse of the original against the stylistic sense of the reader in the target language, and against his syntactic comprehension span.

Despite the success, though, of Arndt's modest Pushkin in rhymes and of Lowell's splendid Baudelaire, many translators—having held all the factors in balance—have decided that for certain poems rhyme will not come across. They omit some or all of the sound effects, and concentrate on presenting a poem's line of images and emotional logic. Some such decision was made by the French translator of Robert Lowell and John Berryman, Alain Bosquet, who gives the last couplet of Lowell's "To Speak of Woe That is in Marriage,"

Gored by the climacteric of his want
He stalls above me like an elephant,

as

Aguillonné par le retour de ses besoins,
Il s'empetre au-dessus de moi comme un éléphant.^[73]

Lowell's rhyme is imperfect, ferocious. Missing it, the French gives the inherently weak simile of the last line far more force than it has in Lowell's English. What is aggressively ugly in the original is carried, to a large extent, by the deformed rhymes.

John Berryman's "Note to Wang Wei" finds rhymes for every line, but M. Bosquet has been able to pair only half the poem's lines:

How could you be so happy, now some thousand years
Disheveled, puffs of dust?
It leaves me uneasy at last,
your poems tease me to the verge of tears
and your fate. It makes me think.
It makes me long for mountains & blue waters.
(I'm reconfirming, God of bolts & bangs,
of fugues & bucks, whose rocket burns & sings.)
I wish we could meet for a drink
in a 'freedom from ten thousand matters.'
Be dust myself pretty soon; not now.

Comment peux-tu être si heureux, défait depuis
quelque mille ans, boufées de poussière?

J'en éprouve un malaise à la fin:
tes poèmes me taquinent jusqu'au bord des larmes
comme ton destin. Cela me fait songer.
Cela me fait désirer montaignes & eaux bleues.
Me fait désirer quot permettre.
(Je me rassure, Dieu de tonnerres & d'éclairs,
de fugues & de béliers, dont la fusé brûle & chante.)
J'aurais aimé que nous nous rencontrions pour un verre
dans un 'délivrance des dix mille matières'.
Serai poussière moi-même bientôt; pas maintenant.

In the French, lines 1 and 3–7 are unrhymed, and it takes till line 8 for "poussière" to find its mate. Bosquet rhymes the last line with the ninth, but in Berryman it rhymes further back (line 7), producing a closer interweaving of end sounds in the final two-thirds of the original. Berryman's "not now" might seem spliced on the end of the last line, unjustifiably; but he wants clipped, pitiless tones here: cutting out personal pronouns and working with sentence torsos, he manages without a hint of self-pity to deal with his own mortality and need for fame. "Now" is rhymed back into the body of the poem (twice, on "how" and "allow" in line 7), so as not to appear obtrusive; partly hidden by distance, the rhyme is implicative. The French goes some way toward managing this terseness, yet ends with the diffused register of the long word, "maintenant." Unpredictably but always firmly, Berryman's rhyme binds the whole and keeps the parts separated. Rhyme in the French version does not, to the same extent, convey a tone by sound, meaning, and placement.

Translations, though, are rarely perfect failures. No inadequate translator will ever, like the immortal William McGonagall, be admired and reprinted for transcendent badness:

'Twas in the year of 1869, and on the 19th of November
Which the people in Southern Germany will long remember,
The great rain-storm which for twenty hours did pour down,
That the rivers were overflowed and petty streams all around.^[74]

The lines are stuffed full of grammatical (and ungrammatical) odd lots and lengths in order to make up the approximate weight to balance them. Words are cut ("'Twas") or phrases spun out ("did pour down"), clauses are connected by an unnecessary "which" or "that," or by running "ands" of parataxis: all in order to make the middles of lines roughly equal. The same is true of the meter, which without transition is either mechanic doggerel or utterly irregular. Thus does sublime inferior verse hang its lines on hooks of rhyme. Like folk poets and children, many bad poets believe that to have poetry, one must have rhyme. Here are the first two of twenty-five stanzas that were found framed on a landlady's wall in Pacific Beach, California. The author has McGonagall's anxious need to pin down date and place at the start:

It was on October thirty-first, nineteen hundred and thirty-one
That, the family and friends at the N—— home had such great fun!
I feel moved to write about it, and let Mrs. N—— know
How glad I am, that I was invited to R——'s wedding to go.

When I first heard that R—— was to be married some weeks ago
I thought, oh, that's far distant. But how quickly time goes. So
Now it is in the past. The event was last Saturday night,
That Dr. M—— tied the not, and I'm sure he did it right.^[75]

The need to rhyme devalues the middles of the poem, all the parts inside the rhymes. The predominance of the cognitive over the aesthetic principle, in Meijer's terms, skews any right relation of rhyme to the rest of the line; so one lurches from rhyme to rhyme, foothold to desperate foothold over fifteen or sixteen syllables. With the drastic extension of middles, all must be spelled out, embellished. With its dead metaphors and inversions, its oddity of juncture ("goes. So / "), its grasping for clichés to make the rhyme, and the rhyme conceived as makeweight of any clutter so as to get out to the right-hand end, the poem is a disaster. It is thereby unintentionally charming. One feels in the presence of the thing itself.

Many of the received rules for making up lines or for rhyming are here violated with the impunity of ignorance. Yet rule violation, whether deliberate or unconscious, is never a mistake in principle. What happens here is but a glorious extension of effects present in all good rhymed and metered verse. Only when poems like these, which succeed as formalized communications to special audiences, are brought near professional poetry for comparison, will the question of judgment arise and the rhymes seem, as Stevens would say, unpardonably expected. The special audience of such a poem would have a sense of poetic form, and of the limits and possibilities of the occasion, so withered and restricted as to permit such writing to have meaning and great value for them. They see the trouble the writer took on their behalf, the labor to produce such an evidently rule-bound object. The marks of labor and of duty are all over the poem that is traditional—and bad!

Awful poems have a terrible force of seriousness and concern. The urgency of the cognitive principle, the need to communicate, apparently short-circuits aesthetic premeditation. Two final instances:

Other days and other names come to mind with ease:
Nathan Hale, Lincoln, Wilson of the League;
And now he too is gone as well as these,
All felled by irony and hate's intrigue.
Yet now his words and spirit are made free
The newest witness to our liberty.

You're the dream in my heart
That's dearer to me
Than the air I breathe—
Barry Blue.

You're the singer I love
My dear turtle dove . . .
Oh, if I could be true
To this star who's so new—

Barry Blue . . .
I'd be over the moon with you!^[76]

The between-rhyme middles are devalued metrically, syntactically, in register shifts, inability to count, failures of euphony; while the unsubtle ends of lines positively shout their rhymes. The content has here rendered as malleable material the structure of the line; rhyme never had a chance. And these passages have the advantage of a more or less traditional scaffolding; ineffectual free verse is perhaps in even greater danger of building up a surfeit of like sounds that do not justify their positions, thus freeing but not renewing rhyme.

Innovation

Some tasks of construction, rhyme cannot perform; others, the unpardonably expected, it has performed only too well. With few exceptions modern poets decide to keep the device of rhyme but warp and reposition it. The need to reinvent the device from within has produced original types and placements for the ancient method of sound linkage. At no other time since the Middle Ages, when rhyme in English verse was an exotic foreign import, has the device been pulled and squeezed so experimentally. Most poets after 1910 are represented by Marianne Moore's statement: "Concealed rhyme and the interiorized climax usually please me better than the open rhyme and the insisted-on climax."^[77] In the work of many major poets this preference has resulted in a drastic weakening of shape; not the abandonment of form, but an increased unpredictability of rhyming effects that must afford another, not necessarily a keener, pleasure to the reader.

One way of arranging the modern examples in this section would be to set them out according to degrees of distortion and avoidance of the traditional line-end position. To this account could be added a survey of types and intensities of modern deviations of rhyme sounds from "perfect" rhyme, from a construct that could be educed from the sound values

of traditional rhyming between Chaucer and Tennyson. Instead I have chosen a literary schema that moves from sound play in rhyme to sense play in rhyme. The sequence of the following passages will put conventional rhyming behavior near the midpoint of a spectrum between two extremes of language pathology, beginning with nonsense or echolalia and ending with multiple sense or punning. More than any previous period's work, modern rhyme exemplifies all points on the range, including the extremes. So these examples suggest that the same modernity that officially denies the device may also, if with some violence, find its renewal indispensable.

It is easy to pass over lines like:

Co co rico co co rico

or

Drip drop drip drop drop drop drop

in Eliot's Waste Land , yet the bits of rhyming nonsense disorient the reader, suggesting the flashing-up of a hallucinative state:

Weialala leia
Wallala leialala.

The ending of Hugh MacDiarmid's "In the Slums of Glasgow" also stretches internal rhyme to the limit, almost completely obliterating meaning in the technical sense of the term; but the effect of being no longer, or not yet, rhyme, is exactly chosen as the final movement of a long meditation on poverty:

Now the babel of Glasgow dies away in our ears,
The great heart of Glasgow is sinking to rest,
Na nonanunno nunno nana nananana nanu,
Nunno nunnonanunneno nanena nunnannunanut.
We lie cheek to cheek in a quiet trance, the moon itself no more still.
There is no movement but your eyelashes fluttering against me,
And the fading sound of the work-a-day world,
Dadadoduddadadadaddadi dadadodudadidadoh.
Duddadam dadade dudde dadadadaddododah.^[78]

Eliot and MacDiarmid are not the first to mute the signification of words and to exaggerate the seeming autonomy of the system of sounds. In this mode Velimir Khlebnikov and Edith Sitwell were, among others, early experimenters. And Wallace Stevens, after his embarrassment at the overt rhyming in his "June Book," went on to weave into a lengthy blank verse poem the multiple sound values of the letter C . "You have to think of this incidentally as you read ["The Comedian as the Letter C"]," Stevens said in a letter, "You cannot think of it directly. . . . You have to read the poem and hear all this whistling and mocking and stressing and, in a minor way, orchestrating, going on in the background or, to say it as a lawyer might say it, 'in, on or about the words.'"^[79] Stevens is overmodest; the effect is not so noticeable as rhyme; yet neither is it minor. The "stressing" in this poem was implicated enough with essential meaning for him to draw attention to it in the title. In Stevens, and notably in Khlebnikov and Apollinaire, such effects are sometimes wrongly praised as "sensitivity to pure sound."^[80] So far as an aesthetic intention is evident, the sound is not "pure" but has value in and by its slight semantic residue. However, to the extent that any poet piles up sound effects his poem will seem to evade semantic logic, and the whistling and mocking of his quasi rhymes may appear the quintessence of imaginative play. Rhyme is such cases is always conscious challenge to ordinary language: flamboyant surplusage of sound pattern which—even in the word games of concrete poetry—never quite eludes the control of a discoverable, commonsense logic.

Encountering Steven's "Commedian as the Letter C" for the first time, one may not realize that a local consonance was part of a vast intentional orchestration. Here as elsewhere in recent writing the relation between instance and design is masked and appears unsystematic. For some purposes, for some poets, the line-end position is untouchable now, and the

device migrates into the line. W. H. Auden in "Pleasure Island" uses internal rhymes on the penultimate vowel, thus:

What there is as a surround to our figures
     Is very old, very big,
Very formidible indeed; the ocean
     Stares right past us as though
No one here was worth drowning. . . .^[81]

Kenneth Rexroth loads rhymes into the second section of "Andromeda":

Anguish and form and prayer
No excuse no betrayal
No dimension in space or time
Without caution without consequence without motion
The many blades of the revolving razors
The many tears of the breaking sorrow
The fear of the bear the ghost of the bear
The gear of care that is always here
When the cross of words spells zero
There are threes in the sea
There are red columns on the horizon
And fear everywhere
And every year no word at all for her pain^[82]

That sort of thing can go on endlessly, and yet Rexroth knows how to terminate one sequence and begin another, alternating tones and contexts as well as rhyme sounds, rhyme places. In the final lines of Norman Nicholson's "The Black Guillemot," line-end rhyme is only part of the clustering of equivalences:

But, turn the page of the weather,
Let the moon hand up the tides and the pressure-hose of spray
Swill down the lighthouse lantern—Then,
When boats keep warm in harbour and bird-watchers in bed,
When the tumble-home of the North Head's rusty hull
Takes the full heave of the storm,
The hundred white and the one black flock
Back to the same rock.^[83]

In every way this sounds an ending, with four ck words out of the last seven. The resource is the more powerful for being spared for definite effects.

Another sort of writing, neither primitive nor traditional, uses and overrides rhymes within the line and at line end. These verses are chanted to a drum by "The Last Poets," a group of militant Blacks who have original priorities as to what must be carved away from the device to keep it vivid. I take these lines from the album notes of a record:

Selfish desires are burning like fires among those who horde the gold
As they continue to keep the people asleep and the truth from being told
Racism and greed keep the people in need from getting what's rightfully theirs
Cheating, Stealing and Double Dealing as they exploit the people's fears.
Now Dow Jones owns the people's homes and all the surrounding land,
buying and selling their humble dwelling in the name of the master plan
'Cause paper money is like a bee without honey with no stinger to back him up,
And those who stole the people's gold are definitely corrupt. . . .^[84]

We will not find some of these rhymes in any rhyming dictionary. These lines are also aggressively destructive of formal meter. Yet this voice has its own measures and insistent tones; it would take a hearing of the record itself to bear out a claim I should like to make, namely that in this instance poem-as-heard is superior to poem-as-read. What appears here on the page is a set of notations for a voice that can ignore the conventions of meter. In this voice there is not the kind of wit one finds in twentieth-century Popian verse, and yet there is wit,

logical progression, speed. The poem savors the rhymes but drives beyond, ticking them off:

Now the U.S. mints on paper prints, millions every day
and use the
Eagle for their symbol, 'cause its a bird of prey.

Since this is to be taken in by the ear, in print it seems thrown onto the page. Yet there is nothing loose or unjustified. There are three stresses in each long line, resounding and lingering stresses only on the rhymes. Read with these stresses, the lines sort themselves out and convey their own decorum.

One might illustrate other possibilities with a poem whose rhyming segments have not been surpassed in subtlety in this century, Eliot's "Ash Wednesday"; or by such a splendid piece of chance-seeming rhyme as Robert Lowell's "Man and Wife." But instead, for something yet more difficult to explain by the more usual analytical tools, here are two passages from Edward Dorn's Slinger :

From a point on the arc
2 days minus 2 corners
We sure know where that's at Boss, um
We can find it  in the Dark

O Poet, Hey Poet! the sun like a sword
Cuts below the Tanner's Yard
and we must hear the effect formed
of the code name Rupert another time
Another time we will witness
how this double hydrocarbon hustles the future
but for now, Later on
Fresh Distortions have swept the screen
And from the smell brought in by the wind
we have news of the Master Nark
who trailed us into your cycle
Yet there is another, an Unknown
who tracks us
Someone whose fame  is his Name
A summer storm advances Though it is autumn
You will conclude in another Town

Away from the shades
When under the cool strokes of  Muthos
we'll find out about
which way that Epactos goes

For some while we parallel the train
whose shining rails are closed at both horizons
and this group in which our brain
is contained, speakes in the excellent tones
of the beginning of an ascent, feel them rising
into the realm of the surprising
bent over what they say
along the river Rio Grande
'earing the low chordes of the foothills
spitting the seeds of the Sandias
out of the corners of their eyes
as they rise
towards the land of the crazy Utes
over and thru the mordants
of the bridges and the buttes^[85]

Rhymes fall at line ends but not always or predictably. There is a wish to rhyme, but not to take the whole thing too seriously as a formal scheme. A line might just as easily end with "um." In the first passage, the end of a major part of the long poem, there is a massing of peculiar consonant chimes as a signal of closure ("sword-yard," "formed-time," "onscreen," rhymes like Wilfred Owen's pararhymes). The second passage is more systematic, but here too line length and the closeness of rhyme words in sound move within highly variable limits, so one is never fully sure a rhyme has been encountered. We have, too, the curious though not isolated case of rhymed lines that are by intention unmetered. These lines parade their artifice: their carefully disarranged measures open to the reader's regard the allocational tasks of the rhyming poet. The lines do not claim to be anything other than slightly jerry-built, and the humor in this is related to the humor of a mixed diction where, in the space of twenty lines,

"Nark" and "Muthos" and "hydrocarbon" can extend each others' frame of reference. With such conventions this writing is a laughing allusion to the resources and deficiencies of a reigning prosody. The poet has learned enough of that prosody to send it up, though here as everywhere, he will keep more elements than he wants to throw away.

A major theorist of rhyme, Viktor Zhirmunski, speaks of it as a secondary factor in composition, "a means of uniting the rhythmical line into a structural unit of a higher order (the stanza)."^[86] Zhirmunski's theory takes account of Mayakovsky's assertion that in this century secondary elements will usurp the primary place of meter itself; but he denies that there can be rhyme or rhythm without explicit, traditional scansion. He reaffirms the primacy of meter, independently in agreement with T. S. Eliot that there is no escape from meter, "only mastery." To contradict this I have already quoted measured but unmetered rhyming passages and more could be cited from such as Mayakovsky, Olson, Dorn, Robert Duncan, or, as here, from John Logan:

The fog
             stammers everywhere
along the rock
                          break-
water pier
                  and in the twilight air
the Peace
Bridge has its Buffalo steel feet
nearly all cut off.
                               Therefore it walks
like a lame centipede quite
impotent at first
then on to Canada.
Fishermen
                  are chattering in
the fog.
              It's not
just the sign of them on the shore
I hear

but also the live laughter
of those boys with their poles
moving toward us,
breaking through the holes in the mist
which again will fill.^[87]

Intended or not, in speech or in poetry, meter is always present. Metrical categories are always precise, or at any rate aim at precision. It would be possible to attempt scansion of this or the Dorn passages, but would it be appropriate? Would metrical analysis help us to describe the key patterns of these poems where meter was so obviously not intended? It seems to me that the central pattern of Logan's lines is rather a figure of sound wherein words are keyed to one or more other words somewhere within the range of aural memory. Of many patterns, one: "pier" reaches above to "everywhere"; below to "air," "nearly," and both halves of "therefore." All these words are given emphasis by line break and juncture, and their placement gives energy to nearly every possible place in the line. This visual isolation of keyed sounds, combined with visual and aural continuities forward and back, makes the passage at once eventful and casual, reducing the prominence of the rhymes while at the same time catching each rhyme firmly. A pattern emerges and a tone is managed, though without the help of an abstract metrical scheme.

No other poet could employ these measures of Logan's, unless he repeated exactly the same poem. And yet this technique of unmetered rhyming and variable line breaking is clearly imitable, is indeed one of the major stylistic innovations of the twentieth century. The number of free verse poems on this model (a model not, of course, of Logan's devising) is already great, and in itself evidence that rhyme, far from having begun to perish, has undergone important (all but unrecognized) changes.

If we will listen for rhyme and rhyme-related parallelism in much free verse, we will be amazed at the richness and variety

of what we hear. The more so because, as Paul Kiparsky has noted, "where parallelism is used as a free feature it is always essential to the meaning . . . ; 'free verse' actually frees verse schemas for significant use; hence it can be a more difficult and more expressive form than regulated verse."^[88] Rhyme as instance, cut free from larger designs, would appear to have—"always," according to this linguist—full semantic justification, as indeed it does in W. S. Merwin's brief "Homeland":

The sky goes on living it goes
on living the sky
with all the barbed wire of the west
in its veins
and the sun goes down
driving a stake
through the black heart of Andrew Jackson^[89]

A detestation of Jackson is carried by the mocking of his name by harsh rhyme sounds twice in two lines. There is another use of rhyme as a free feature, much more common in free verse than in conventional, and far more likely to produce semantic relationships between rhyme words. A version of this occurs at the middle transition in Hopkins's choral song, when rhymes are conceived as replying to each other, like echoes:

                                      THE LEADEN ECHO

 . . . So be beginning, be beginning to despair.
O there's none; no no no there's none:
Be beginning to despair, to despair,
Despair, despair, despair, despair.

                                      THE GOLDEN ECHO

        Spare!
There is one, yet I have one (Hush there!). . . .^[90]

The second, echoing word is drawn from a nest of associations within the first, and either extends or, as in "despair-spare," denies the first word's logic. It is a strategy of addition, correction, or reversal, where the poet combs out a word's rhyming associations and lays them alongside one another. "There is no direction," writes W. C. Williams in Paterson :

            Whither, I
cannot say. I cannot say
more than how. The how (the howl) only
is at my disposal (proposal). . . .

Or again: "with the roar of the river / forever in our ears (arrears)."^[91] The rhyme follows its fellow, transforming and extending semantic context. A related method has been used to suggest how the mind races out of control, making metonymic connections along the lines of the rhymes, plunging into language pathology:

The sky is pitiless. I beg
your pardon? OK then
the sky is pitted . . .
            Yesterday the
air was squeaky clean today
it's dull and lifeless as an
addict's armpit. Surely you mean
leafless. . . .
             Some words
are briefly worse than others:
get the librium gun and point
it and Kodak at that Kodiak.
You see: No hope. So don't
hope. Hop, skip, jump, or
lie down. . . .^[92]

James Schuyler, author of these lines, clearly knows (with Freud) that double talk is a special form of communication.

From Eliot and MacDiarmid to Schuyler, these examples may remind us how Keats as a child would reply to people by making a rhyme on their sentences, and then laugh.^[93] As a language habit, rhyme seems a derangement, seems to say something only about language, but there is always the possibility that it is also telling us something about ourselves. Modern use of the device, on the showing of these passages, may be said to force its artificiality and irrationality into the light. For poet and critic, this can be instructive: when we perceive that language, like nature in the scientist's estimation, is only partially organized, we regain incentive for basic discoveries.

By boycotting puns and confining rhymes, English neoclassical poets wished to control this subversive likelihood of hidden perspectives within language and ourselves. The consciously unified period style of Pope, Swift, and Johnson is the historical opposite of the stylistic pluralism of the post-Romantic era, and hence its abiding attraction. Those twentieth-century poets and critics who promote the chastening lesson of neoclassical examples seem partially to contravene the terrorist imperative of Rimbaud, "One must be absolutely modern." To concentrate intensively on the relation of rhyme to reason is their way of modifying the Romantic frame of mind these writers share with Rimbaud; critics like W. K. Wimsatt argue their particular modernity by insisting that there is more to rhyme than the reverberations of harmonious unreason. Faced with these two modernities, that of Rimbaud and that of Wimsatt, I should like to suggest that each requires the other for its completion.