I—
PRONOUNCED DEFECTS:
THE TRANSEGMENTAL DRIFT
1—
"To Hear with Eyes":
Shakespeare as Proof Text
O O O O that Shakespeherian Rag—
T. S. Eliot, The Waste Land[1]
The first phase of my subject is that authorial idiolect, or "style," which has come to be the measure of literary language in English. Acoustically textured to the point of distraction, it is a verbal medium full of a phonemic fury of sound signifying not nothing—but signifying more energetically than any signified requires. This is a sense of the commanding "Shakespeherian Rag" that haunts all our literary writing and reading. When Eliot in The Waste Land spells out, stretches out, the incantatory Bardic name, he does so without the punning "hear" and "ear" which are there, nonetheless, to be read. But how read? The answer is obvious enough to elude most literary study: we listen while we read. Who can doubt it? How, though, to prove it? This chapter will attempt in some measure to close the distance between credence and demonstration.
It remains a humanist truism that literature speaks to us. The work of deconstruction might be understood to have resurrected the dead metaphor of such a notion in order to lay its ghost for good. Literature has no voice. It is text, not talk. But what is left in the wake of this widely successful campaign against phonocentrism? Let's take an extreme case. Who would deny that the rest of us read to ourselves differently from the case of a sighted reader deaf from birth? What then is it that we think we hear, or hear in thinking? Not the author's voice, granted. Yet if literature cannot be fairly said to speak to us, perhaps it speaks through us. Our being there in front of it is the precondition not of its existence but of its function. Or another way of putting our relation to a text: we speak to it, silently, in the loose sense of moving to a beat. A less extreme case. Imagine a cultural context within the use and dissemination of the English language in which the written words of that language, while referring to the same known entities or concepts as they do now for us, are nonetheless quite consistently different in aural quality. Easily done, since this
is of course the case with Elizabethan pronunciation. Readers silently ventriloquize a text according to the linguistic conventions of their time. To borrow from Shakespeare's sonnet 23, the role of the reader is "to hear with eyes." Texts are in this sense not begotten but made, produced in reception. Such is the auditory leeway opened precisely by the lack of literature's inherent voice. This chapter will eventually close in on a very particular and particularly eccentric effect, the transegmental drift: product of a phonemic operation off-center from the graphic signifier that occasions it. To attend to such phonic drifting is to track the smallest node, in effect, of a deconstructed writing's so-called difference from itself.
A "Writerly" Bard?
In Shakespeare's sonnet 15, the speaker, after a sustained contemplation of transience in its organic aspects, summons a specular image, though entirely undetailed, of his beloved young man, promising him, or it (his image), the immutability otherwise stolen from the youth, from any and all youth, by the violence of time: an "engrafting" of new vitality through poetry. I choose this sonnet because it happens to have most of an entire chapter devoted to it, the last and summary chapter, in Stephen Booth's Essay on Shakespeare's Sonnets and to be thoroughly and expertly glossed in Booth's subsequent edition of the sonnets.[2] These commentaries constitute a pair of interpretive texts (by any other name) that together mark a kind of watershed in late New Critical readings of our most read poet—and of poetry in general. To begin with, the lyric in question:
When I consider everything that grows
Holds in perfection but a little moment,
That this huge stage presenteth nought but shows
Whereon the stars in secret influence comment;
When I perceive that men as plants increase,
Cheered and checked ev'n by the selfsame sky,
Vaunt in their youthful sap, at height decrease,
And wear their brave state out of memory;
Then the conceit of this inconstant stay
Sets you most rich in youth before my sight,
Where wasteful time debateth with decay
To change your day of youth to sullied night;
And all in war with time for love of you,
As he takes from you, I engraft you new.
Booth's compressed editorial gloss on the syntactic duplicity of the opening lines is more expansively investigated in his Essay . The note in his edition of the sonnets—to the effect that "under pressure of syntactical necessities
introduced by Holds, a simple subject-verb-object construction comes to be understood as if it had been 'When I consider that everything that grows'" (SS, 155)—is enlarged upon in the Essay in view of its reading effect, the dynamics of its processing. Shortly after Stanley Fish argues in Surprised by Sin for a cognizance of similar crossed signals in readerly responses to Milton,[3] Booth writes that this syntactic self-revision "requires an easy but total reconstitution of the reader's conception of the kind of sentence he is reading" (ESS, 181). There is, however, something more cannily to the point—or is it uncannily?—in Booth's own compression of this issue in his edition: again, "comes to be understood as if it had been." It wasn't, it only seems to have been; the first meaning cannot rest easy, but its necessary revision, in light of our already elapsed experience of the first line, is itself a phantasmal rewriting. If only implicitly, Booth here parts company from Empson. Though sensitized to "double grammar" in the sonnets, the favored Empsonian mode of analysis, Booth would apparently resist the notion that such vacillating syntax is ever finally "resolved" into stability and coherence. Instead of saying that what seemed to be the sense of the first line turns out not to have been after all, Booth intriguingly suggests that what seems in retrospect to be the reconstituted first line is still only an "as if." Retroaction in reading does not recast the original but only rereads it as it would have to have been, but wasn't, in order to permit continuity with the grammar that ensues upon and reroutes it. Rethinking is not revision; the "last word" on a set of lines does not supplant, only subdues, our first impression; what is is not necessarily what was.
To unload these implications from Booth's editorial gloss is to appreciate at how narrow and elusive a range he can detect what he has claimed in his earlier Essay about such snags in the "syntactic fabric" of the sonnet: the fact that they "do not merely describe inconstancy but evoke a real sense of inconstancy from a real experience of it" (ESS, 181). This too is an unusually suggestive phrasing, at least from the hindsight of contemporary discussion. Such double grammar in the sonnets does not mime inconstancy, or even thematize it, but performs it autonomously in the linguistic sphere. At the risk of "double reading" Booth himself—or paraphrasing, that is, from a vantage two decades further along in critical history, what it seems as if he meant—it must nevertheless be remarked that the "real sense" of mutability in the lines has little to do with grammar developing a mimetic analogue of transience based on floral and astral prototypes. Rather, the ephemeral in language is its primary experience, as "real," as material, as any other. Its evocation is at most homologous with the declared themes of the poem, not subordinate to them. In this latter sense, style does not "answer to" theme—except by enacting its counterpart in words.
Just this line of thought also serves to underwrite Booth's attention to Shakespeare's phonemic and syllabic as well as syntactic play in the sonnets. Pursuing this interest, Booth's approach demonstrates as much congeniality, if not direct allegiance, to the Jakobsonian as to the Empsonian tradition, with the former's more systematic commitment to (Booth's phrase) "phonetic and ideational interplay" (SS, 157).[4] Such a commitment is ultimately in the service of that iteration-within-difference which Jakobson has elevated to the very definition of the poetic "function," which thereby "projects the principle of equivalence from the axis of selection into the axis of combination."[5] Words are not just linked by syntax, by combination, but fastened, again, by that "phonic affinity" (among other features) which intersects and inflects the axis of contiguity. One word is selected over another from the paradigm of alternatives not just to fill a space in the grammar of progression but to secure an equivalence (or resemblance) with some other word previously set in place, as in the aurally looped sequence in sonnet 15 likening, as well as linking, "consider" to "conceit" through the intermediary syllabic matter of "perceive." The paradigmatic principle of comparability thus invades the syntagmatic chain of combination; equivalence dictates contiguity; sameness overtakes difference; language becomes poetic.[6] Style in this sense is always a particular intersection of the vertical paradigm of available alternatives with the horizontal sequence into which the decidedly chosen word will be dropped into position. Plotted by such choices in succession, the poetic is the crossing of the contiguous by the interchangeable.
Working back, like Jakobson, to the binary linguistics of Saussure, Michael Riffaterre's more recent investigations into the language of literature also concentrate on pertinent oppositions as they are spread out across the syntactic line.[7] For Riffaterre, however, equivalences that cannot simultaneously be pressed into the logic of sequence still address each other in an oppositional play that could never be fully accounted for by the scrutiny of those surface features regularly studied as "style." In essays collected under the title Text Production, Riffaterre demonstrates that it is in fact surface texture's structuring difference from its origin and alternatives, as perceived in the reader's own "production" of the text, that generates what another of his titles identifies not as the "semantics" but as the Semiotics of Poetry . In true formalist manner, style is sidelined by questions of literariness, itself in Riffaterre a kind of binary alternative to literalness: an avoidance of the explicit by figural displacement. True to semiotics, too, literary textuality is constituted by an array of associations that can be traced only through "signs" of meaning, clues, gestures, vestiges. Rather than searching out, as Saussure did,[8] the anagrammatical displacements of a Latin verse line—dismemberments of sacred or heroic names to be reassembled in the reading—Riffaterre
instead looks for more obvious systemic disturbances that foreground the unsaid as unsaid, avoided, suppressed. For Riffaterre, "Saussure's stroke of genius . . . was to understand that the text's true center is outside the text and not behind it, hidden away, as victims of the intentional fallacy are fond of thinking" (TP, 76). When Riffaterre himself looks beyond rather than behind the text, however, what he finds is a veiled revelation far less encrypted than Saussure's anagrams. He finds a semiosis mobilized without being literalized. Rather than radiating centrifugally from some occulted core, the text is generated from point to point by its variance from a paradoxically external (and unwritten) center.
The relevance of this notion to Booth's edition of the sonnets grows quickly clear. The editor's scrupulous logging of those contemporaneous proverbs and aphorisms that "underlie" Shakespeare's sonnets, for instance, witnesses to an intertextual pressure which Riffaterre would theorize as the semiotically definitive relation of a given sonnet to its decentered matrix. More specifically, this relation is understood by Riffaterre as a semiotic reaction formation variously characterized in terms of denial, avoidance, suppression, conversion, condensation and displacement, all with their openly courted Freudian associations, including especially the return of the repressed. Already generated by the deflection of some received or proverbial formula, then, the text is further "converted" or "expanded" into readable units by local "actualizations" of this clichéd "source."[9] We encounter here the phrased, rather than tacit, semiosis of the paragram . "The model I propose for the lexical paragram is thus the expansion of a matrix. Since it is lexical, this expansion occurs in the form of words linked together grammatically, and is not phonetic or graphemic, as in Saussure's paragram" (TP, 77). Evident here is Rifaterre's tacit holding action against exactly the kind of phonemic reading which this chapter finds invited by the Shakespearean text—as the clear (if pulsing) signal of its literariness. The phonotext "produced" in this way, as we are to see, occupies a middle ground between the lexical/grammatical sine qua non of Riffaterre's method and the graphemic free association of an anagram. Between full semantic security and syllabic mayhem, then, lies the domain of phonemic reading.
A Riffaterrean take on the opening quatrain of sonnet 15, for example, might read the mentioned "comment" of the stars (their legibility as a sign system) to reveal the whole figural sequence of astrological reference in the text as the expanded variant of the proverbial notion that we read our destiny in the stars . Where Booth and others have noted the etymological hint of sidereal or astral observation in the sonnet's first verb, "consider" (from con - + sider-, sidus, star), a Riffaterrean semiotics of the text would take this verb as a "lexical paragram" providing for the astrological matrix its first "model" (the
guiding prototype of subsequent actualizations) in the etymological sediment of a single syllabic cluster. Though Riffaterre himself does not very often pursue matters even to the level of the syllable, let alone its constituent phonemes, his method would seem to allow for this in principle. The end of sonnet 15 offers a test case ready to hand.
As the text moves into the monosyllabic tread of the closing couplet, one word stands out. The disyllabic verb "engraft" is a horticultural trope that at the same time marks the sonnet's deepest predication: the power of sonnetizing in and of itself. Booth annotates by paraphrase the likely pun on poetic engraphment: "'As time withers you, I give you new life (by writing about you)'" (SS, 158). He goes on to suggest, however, that to become operable the pun needs (intertextual) support from the opening of the next sonnet: "Despite a probable pun on engraft . . . and its Greek root graphein, 'to write,' . . . a reader presumably does not recognize this first of several traditional claims for the immortalizing power of verse (see sonnets 18 and 19) until the line is glossed by the first quatrain of sonnet 16, which is both logically and syntactically linked with this one." This ensuing quatrain insists that there is "a mightier way" to fight time than with "my barren rhyme," an image whose denial of fructification ("barren") seems to retract what sonnet 15 has proffered with its primary botanical sense of "engraft." It is only the punning double of this phrase, of course, that has instituted the "traditional claims for the immortalizing power of verse" in the first place, those claims which Booth finds suppressed beyond immediate recognition—except through the route of metaphor and its decoding. But the metaphor itself, even if we catch its general application, is not very clearly pointed. The "cutting" by which a graft takes place harkens back in its derivation to the Latin stylus on its way to the Greek root for carve . Booth is surely right in supposing the line to allude to the "practice of replacing the wasted limbs of old trees with slips that grow to be new boughs" (SS, 158). But if this allusion in turn alludes to the speaker's expressed hopes for the marriage and propagation of the young man, then it cannot be the speaker who is to engineer this new grafting of a scion. Not only won't the whole metaphor thereby come to rest where we expect it to, but the apparent transitive grammar of its vehicle is internally unsettled. If the speaker's meaning is that by writing about you I "engraft you new," then this can mean—following Jakobson's treatment of similar ambiguities of the predicate in sonnet 129—either that "I make (by engrafting) you new again" (an adjectivalized adverb) or "I graft you anew " (an adverbalized adjective). In neither case, however, does the figural logic, as governed by the transitive grammar, come quite clear. Does the speaker mean that I cross something with you, or implant something in you, or attach something to you, or otherwise extend you by means of something—or, if you are the sole direct
object of the phrase, that I simply ingraft you, however that would come about?
And even in this last case, what could possibly be disclosed as the tenor of that metaphoric vehicle—or of any of the preceding grammatical alternatives, for that matter—except the whole of you in renewal? We are thus thrown back on the verb's punning application, which makes, albeit in its less idiomatic format, more syntactic sense after all. It does so precisely by its near-homophonic collapse of vehicle upon a tenor spelled out at last: the "graft" that is always "graphic" at base. But phonemic too, aurally overdetermined. The closing verb phrase thus manifests climatically in this sonnet—and not, we shall later see, for the first time—that graphonic interplay which will eventually come to dominate this commentary: as a test case for a micro-poetics of style beyond the lexically coded semiotics of poetry. The phrase "I engraft you new" can, in other words, just faintly be heard, be "produced" upon the inner ear of the reader (in what might be termed a paraphonic variant actualized by liaison of the t and tacit elision of an unwritten o ) as the phrase "I engraph t' you new." Or even as "I engraph t(o) you (
In his generally admiring review of Riffaterre's semiotic methodology, Paul de Man objected mostly to what he saw as a consistent sheltering of description from the deconstructive fact of sheer inscription.[10] This shortcoming is occasioned, on de Man's account, by Riffaterre's failure to attend closely enough to the master tropes of rhetorical analysis: namely, the pervasive catechresis (the alogical strain, paradox, or anomaly) of poetic figuration and the prosopopoeia ("giving face") that attempts to mask or naturalize it. In view of Shakespeare's explicit thematizing in this sonnet of such a giving (and preserving) of face through a nondescriptive engraphment, through sheer apostrophe, it is further worth noting that de Man bases his tropologically oriented critique of Riffaterre indirectly on Saussure's interest in the Greek root hypographein —or "signature"—as an important source for the "hypotextual" variants of Saussure's inscribed anagrammatic ciphers (30). What de Man objects to in Riffaterre's retreat from such an admitted graphological emphasis is that the resulting method subserves without proper qualification "the determined, stable principle of meaning in its full phenomenological and cognitive sense" (30). But how stable is this principle? The burden of the present discussion is that, in the phenomenality of a text when read, cognition
subtly diverges from direct semantic processing; inked letters enunciated to oneself often take thinking by surprise. A stress on inscription may be one corrective to a semiotics of description, but then so, on the other hand, is an emphasis on reception. The path through a text taken even by a resolute tropological reading, deconstructing all stable reference as it goes, will be phenomenologically impeded at every "turn," or troping, by the phonic as well as graphic materiality of the language in process—impeded, in short, by "text production" at the cognitive (if often subliminal) level. The full implications of this in connection with sonnet 15 we must hold for a few pages longer in abeyance.
Beyond Formalisms
For now, we need to expand still further the critical and theoretical context within which the play of phrasal alternatives in poetry can be addressed. We turn from Riffaterre's paragram (including here the "paraphone" as well) to the frequent stylistic investigations of two very different critics. In their divergent vocabularies and allegiances—the one a leading promulgator of poststructuralist theory in America, the other an inveterate British humanist—Geoffrey Hartman and Christopher Ricks each nurture in their work a sense of potentially subversive lexical free play that resembles Riffaterre's "agrammaticality" as the very definition of literariness. In "The Voice of the Shuttle: Language from the Point of View of Literature," an important essay from Beyond Formalism , Hartman skirts very near to Riffaterre with his idea of an embedded "seed phrase" bursting in various contortions upon the surface of a poetic line.[11] Like Ricks as well, Hartman is interested in the layered reticulation of a text (in the root sense of a textured fabric) woven upon a "silence" that will not quite stay put. In Ricks, too, this silence is every bit as "volcanic" as Hartman dubs it (342), and from it may erupt at any turn the rejected or dismembered phrases that might, alternatively, have gone to make it up: the matrix as cauldron. In Ricks's collected essays, The Force of Poetry , the ground of poetic meaning seethes with pressures upon it that do not quite become presences in it, the shimmering surface of a poetic line alive with tremors from beneath.[12]
Hartman's essay, however, is both narrower and broader at once. It proposes a complex interchange between the "microstructures of literature, entities studied by linguists" (341) and the largest narratological issues of mythic plotting. He moves, for instance, from such "cross-eyed" phonetic structures as chiastic alliteration to the fateful double cross of the Oedipus story. At its most provocative, in fact, his essay describes an even larger trajectory, moving from the space between syllables and words to that breathing space of "indeterminate middles" between "overspecified ends"
(339) which develops as a kind of structuralist definition of human life as plot. One of Hartman's examples of alliterative chiasmus, among several sound figurations he takes up, is the "famous pun" in Paradise Lost, "O Eve in evil hour." What Hartman does not do with this phrasing, this voweling, is instructively continuous with what he does. "To go from 'Eve' to 'evil' is a metaphor on the level of sound" (343), he writes, as if he were directly alluding to Jakobson's notion of echoic substitution in the axis of combination, his "figures of sound." Hartman's gravitation toward a "transformational poetics" of the "zero value" (341) is illustrated here between the bonded phonemic halves of ev/il(l) . What, this particular sound figure of "etiological distancing" (338) asks us to ask, is the direction of causal descent? Which came first: the evil or its namesake? When Hartman suggests that the power of the line derives from the overtones of derivation itself, "the sense of a third term or matrix, a common root from which both might have sprung" (343), his (unmentioned) overlap with the work of Riffaterre (as well as that of Jakobson) could hardly be more direct. What Riffaterre would advance in answer to Hartman's question about the priority of Eve or evil might well be that the whole conundrum is materialized as the variant of a more overtly formulaic teasing with false etymology. The text is produced in this sense as the paragrammatic dodge of a pat aphorism like "The Eve in all evil," as in the comparable (if commonplace) punning of a bromide like "Woman means woe to man." In any case, it should be clear that Hartman is intrigued to read in such a case the projection of paradigmatic equivalences onto syntactically deployed paronomasia in a manner that certainly calls up the "conversions" and "expansions" of Riffaterre's semiotics.
Beyond this, there is in linguistics a more problematic sense of juncture or "zero value" ambiguously negotiated between words, which Hartman hints at but never pursues directly. Illustrating the "language jam in which we are stuck" (343), which is often enacted in poetry as a jamming at intersections, and drawing on Elizabeth Sewell's linguistically imprecise but suggestive notion of the "soundlook" of words (341), Hartman offers as example of junctural irony the title of Wallace Stevens's "Le Monocle de mon Oncle." Comments Hartman: "Technically defined the difference in sound is a slight distortion of quantity but primarily a matter of what linguists call 'boundary' or 'juncture'—here typographically indicated by 'monocle' dividing into two words, 'mon oncle'" (341). Once again, a poetic test case is generated from, in Jakobson's sense, a metonymic projection of sameness in a paronomastic domino play. Hartman's allusions there to the linguistics of "juncture," however, have an even more direct bearing on the present explorations than do his literary examples. Though none of his quotations is strictly ambiguous in its phonemic boundaries, he tangentially invokes a whole range of debate, as
summarized in Kooij's Ambiguity in Natural Language , about the functioning of so-called suprasegmental phonemes—namely, pitch, stress, and juncture—to enforce lexical decidability. More so than with a chiming phrasal balance like "monocle" and "mon oncle," the classic problem for phonological "disambiguation" (the problem of telling where one word ends and the next begins) arises with such "minimal pairs" as "an aim" and "a name," "night rate" and "nitrate," "light housekeeper" and "lighthouse keeper"—or, for that matter, our inaugural instance from Austin, "I stink" and "Iced ink."[13] Such a phonic crux has sometimes been addressed not only by the general notion of "suprasegmentals" but by the concept of a sliding phoneme of "zero" value, a so-called zero allophone of internal open juncture (19). It is this linguistic feature to which Hartman may be alluding in his mention of "zero values." For many linguists, however, as Kooij observes, this offers a suspect account, a "fiction" (11) designed to rescue segmental autonomy. In the study of adjusted and suspended junctures, one is advised to recall Jakobson's postulate: "From a strictly articulatory point of view . . . there is no succession of sounds. Instead of following one another the sounds overlap."[14] In view of this fact, a purer example of junctural value under poetic pressure, one not requiring a footnoted disclaimer like Hartman's ("I am aware that 'oc' and 'onc' contain a different phoneme"[341]), would have been provided by another cryptic Stevens title, "United Dames of America." In this brief poem about political oratory, the title, in its outrageous homophonic spoof of "United Aims of America," becomes a virtual textbook example of ambiguous juncture in a minimal "pairing"—one phrase written, the other (silently) sounded.
We should now be able to return more alertly yet to the linguistically gauged plight of Eve in the garden, at exactly the point where Hartman himself broke off with her. The "zero value" reinvested with force is not just the lexical cipher at the morphemic boundary between syllables, "Eve" elided with "il(l)." Further, there is the actual liaison in speech that overdetermines both ends of the punning phrase. Given that its second vowel in standard pronunciation is /
for instance, or between h and its silent phantom double, would themselves be included, however, in Hartman's hedging analogy drawn from conjectural physics: "You probably feel as impatient with me, and all this talk about zero values, as Bishop Berkeley with Newton's infinitesimals. He called these entities, calculable only by Newton's theory of fluxions, the 'ghosts of departed quantities'" (347). My very slight "impatience," and I assume that of others, comes simply from Hartman's own, his itch to dart away from a further speculation on the linguistic—or call them semiotic—implications of the stylistic intricacies he puts under such intensive scrutiny. More even than his approach is inclined to tally, zero values count.
Larger questions are certainly at play, and at stake, in the structural, psychoanalytic, and philosophical reverberations of Hartman's essay than in any of the twenty papers collected by Ricks in The Force of Poetry . Yet neither critic pursues his speculations quite far enough for close readings to open upon a principle of textual generation. Ricks, for once, is almost as near to a delimited but genuine theoretical breakthrough as is Hartman. Indeed, Hartman's use of Berkeley's phrase "ghosts of departed quantities" reads like a paraphrase of Ricks evoking the poetic vanishing act of the "anti-pun." Words, phrases, syntactic understandings are momentarily enrolled only to be annulled by context, "surmised and then ruled out" (101). The phenomenon turns up everywhere in Ricks's essays, on writers from John Gower to Geoffrey Hill, being more or less implicit in his previous stylistic considerations of Milton and Keats as well. With Wordsworth's phrase "fleet waters of a drowning world," for instance, Ricks calls attention to the "anti-pun" on "fleet" as noun. The adjectival form of this monosyllabic "variant" (to recall Riffaterre's term) would, according to Ricks, "be careless or perverse if it were not positively (rather than forgetfully or willfully) setting aside the other sense" (101). He garners his characteristic authorial validation for this by quoting Wordsworth's definition of the poetic temperament as the "'disposition to be affected more than other men by absent things as if they were present'" (100). In what might well be taken as a "semiotics of poetry" capacious enough to embrace such contradictory semantic signals, absent definitions in Ricks's readings hover like ghostly revenants over the text, often with a sense of violent irrelevance. They create at once a linguistic siege and its own resistance.
Though Ricks would never put his case in strict linguistic terms, it is as though binary oppositions from the axis of selection bedevil the syntagmatic axis as phrasal impertinences—and must be staved off. In Robert Lowell, for instance, one semantic sense "gains its territory by fighting off" the invasive double. The denotation that emerges "means intensely by meaning the whole thing including its exclusion," a "phantasmal violence" (271) of "potent
absences" and "vigorous spectral combats" (268). Poems "muscular with these ripples of solicited misconstruction," these "warring possibilities" (269), take their part in the "anti-sense" of a "violence-acknowledging non-violence"—a phrase which itself, read (aloud) without the hyphens, aptly enunciates (Ricks leaves us to discover) its own seditious converse. Without for a moment admitting the claims of any tradition of reading other than the line of Empson and Donald Davie, though with a few approving asides about some of Hartman's insights, Ricks is nevertheless very much in phase here with the shared perception in Saussure and Riffaterre that "the text's true center is outside the text." The intertextual variants are generated for him by a tension between legible surface and (Ricks's recurrent word) its indirect "force" of evocation.
Like Riffaterre, too, Ricks tends to minimize the acoustic in his lexical investigations. He is willing to hear a homophonic pun in Gower between "braieth" and "prayeth." And he will homophonically pun himself in assessing Gower's instance of pertinent opposition in the paradigmatic axis as an "air's-breath" difference (10) between plosives. But he does not generally seem alert to the possibility of what I would call the antiphone rather than anti-pun, that antiphonal or contrapuntal phantom latching upon lexical units and breaking down their integrity. If a critic were to grant that there is a phonic slippage of this kind inherent in language, an irresolute and disruptive drifting that destablizes any given constraint upon the bonding of phonemes into words, then that critic would also have to face a fact about (to reverse Hartman's subtitle) literature-from-the-point-of-view-of-language. Such language is in some ways too heterogeneous and dispersive for an author-centered aesthetic like Ricks's, too freakish, eccentric, and insubordinate.
Antiphonemics
Returning to certain remaining effects in Shakespeare's sonnet 15 that depend on phonemic rather than lexical displacements should draw further into the open an important point of convergence between Hartman's and Riffaterre's respective methods. There is, for instance, the acoustic undercurrent that pulls upon the climactic personification of mutability in the poem: "Where wasteful time debateth with decay / To change your day of youth to sullied night." Again we find a brand of double grammar organizing, or disarranging, the line. Booth notes that idiomatic expectations lead us to imagine at first some contest between time and decay ("debate with " in the sense of "against"), a "false start" that must undergo almost immediate revision when "one realizes that time and decay are obvious allies" (SS , 158). Once more we encounter Booth's implicit reader-response criticism, though tied always to manifest linguistic signals in the text. This time Booth leaves the full upheaval of our
response unexplored. The shifting ground of prepositional grammar pulls the foundation out from under the very possibility of opposition; in the blink of an eye we are yanked up short by the recognition, foolish to have forgotten even for an instant, that between time and its own process of decay there is no contest at all—that everything in life colludes with decay, which even as a sound pattern brackets and engorges the next line's "day." Moreover, within the vexed and self-corrective phrase "debateth with" is seeded not only a syntactic irony but an additional phonemic insinuation. For there is in fact a larger phonic frame in place here than that which encloses "day" within "decay," namely the encompassing bracket "DEbateth with deCAY." In Hartman's vocabulary this is a "decontraction" of the syllabic terminals that happens to install yet another "etiologic distancing" between the cause (those inevitably conjoined temporal forces) and the effect (doom). The figuration of decay-as-death is so overdetermined that it further generates a virtual anagrammatic code (graphically in "DEbAteTH with," or again in "DEbAteth wiTH," and phonetically in "DebatETH with" and "DebatEth wiTH"). Over and above this, the genuinely pronounced effect, as it were, the one with more "phenomenal" charge, is the antiphonal hint triggered by the segmental detonation of "debaTETH (debaT/DETH)." Attuned to the text in this way, we can also hear another accordion-like play with phonemes that decontracts a matrix phrase and arrays it upon the lexical surface of the line as a paragram, if not quite an anagram this time. Concerning the phrase "the conceit of this inconstant stay," Booth's edition notes that it "functions like an oxymoron." It also may—may in a given reading—operate as a kind of phonemic pleonasm, its last two words standing thus as a paragrammatic decontraction of "inconstancy." As such, this antiphonal overtone would only foreground by contrast an emphasis on the paradox of a fleeting "stay."
This kind of reading draws on a phonemic patterning evident elsewhere in the sonnet apart from any lexical transformation, though tampering nonetheless with junctural value. An instance: "Vaunt in their youthful sap, at height decrease." If these words abut in silent enunciation with full phonemic distinction, then their very contrast is thus dramatized. On the other hand, the sort of dental "ambiguity" that permits the hint of "death" in "debateth " operates in this case across the lexical segmentation effected by juncture. The chiastic arc of the whole line would in this sense be replayed, if quickly read, in the phonemic hinge between the noun of apogee and the verb of downgrade. No sooner said than done and gone, in this sense the phonemic matter of "height" might be felt to yield on the instant to the verb of its reversal. As with the season of youth it describes, so might the noun "height" itself provide the first impetus of its own undoing. Either way, articulated in full contrast or
nudged toward elision, the phrasing could well precipitate a phonemic reading thematically cued by the logic of antinomy within evanescence.
Since such an effect would scarcely be as forceful, even if registered, as "inconstant stay" or "debateth with decay"—in their service as phonemic paragrams for "inconstancy" and "death"—we might rather leave the sonnet with a more emphatic example of microlinguistic compression, a dissonant actualization of arguably the poem's chief and deepest matrix. To hear as much in this opening line will require an exemplary recognition of the interplay between graphic and phonic signifiers in the "Shakespeherian" text: a graphonic tension. This chapter's inductive survey has so far followed out certain common strands uniting the stylistically alert editorial glosses of Booth, the semantic dismemberments and phonic reassemblages of Saussure in his anagrammatic phase, the structuralist stylistics of Jakobson, the formalist literary semiotics of Riffaterre, the renovated New Criticism of Ricks, and the "microlinguistics" of Hartman in that transitional period self-denominated as "Beyond Formalism" but not yet dedicated to the deconstructive enterprise. If these pages have so far succeeded in suggesting how Booth's interest in "phonetic and ideational interplay" could be linked through Jakobson's "figures of sound" to the embedded phonetic patterning of Saussure's studies, how at the same time Booth's deciphering of multivalent syntax derives straight from Empson's explication of ambiguous double grammar, and if these pages have further tracked a palpable connection between the relationship in Riffaterre of paragram to matrix and Ricks's preoccupation with the anti-pun, as well as between Hartman's finessed zero values and the manifold conversions and expansions of Riffaterre's system—if so, then we should be ready to back out of sonnet 15 by way of its first diffusive phrasing: a test of phonemic differentials in their (in every sense) marginal wordplay.
Sonnet 15 may be heard to turn the corner of its first line, to complete its initial grammar, by citing its own theme of mortality under erasure . In this way the first line becomes virtually the sonnet's last word, the inaugural transformation of a primal matrix. Represented in shorthand, the Empsonian ambiguity of grammar has so far been read to go like this: "When I consider (that) everything that grows / holds-in perfection (or holds still in its perfection), . . . then is when I . . ." What might be further said to operate here, at the level of the phonotext, is a phonemic as well as linear "enjambment" which happens to coincide metrically with the onset of the sonnet's next line. The first line runs on, and at its turning, so also—by grammatical doubling and revision—turns the phrase ("grows/Holds") that comprises its enjambment; turns, too, the phonic matter that frames (both internally and externally) the second word of that phrase. Like poetic lines themselves, lexical units are not cleanly sliced off at their borders but spliced together, edging into each other
on the sly, the slide. If the far end of the word "Holds" tends, for instance, to erode toward a semantic ambiguity, it is the other, the near end of the same lexeme, that—more malleable still—leans toward surrendering its aspirated yet all but silent first phoneme. In so doing, it momentarily "sounds," rather than spells out, the most deeply secreted, because in every sense the most hopelessly clichéd, seed phrase or matrix from which the poem is generated: the brute truth about all life on earth, which always and inevitably "grows / (H)old(s)." That the poem never mentions growing "old," as it never mentions "death," is exactly the point. The h in "Holds" is functional, yes, but frail, is activated, but barely; it holds its lexeme in place, yet with the least insistence. Like everything the sonnet describes, this phrasing itself is mutable, wispy, dogged by its own negation—holding for a moment only insofar as it does not quite give out to "old." Establishing at the very start its semiotic "model" in the layered ambiguities, both grammatical and phonemic, of the first clause, the sonnet as paragrammatic variant—the sonnet as a conversion and expansion of a nest of timeworn truisms about the work of Time—would now, as we listen back over it one last time, be understood to go like this: "When I consider all growing things, and consider how in growing no thing can more than momentarily hold in—or hold on in, hold to—its own ripeness, how all things must therefore grow hold only to let go—in a word, to grow old —then I write poetry, this poetry, the text of you in youth."
The Shakespearean lyric text is thus dense not only with sonic patterning of all sorts, with lexical and grammatical ambiguity, with loaded rhyme, internal echo, even an anagrammatic streak at times, but dense also with what I have borrowed from Ricks's consideration of the "anti-pun" to call the antiphonal variant, spawned by a phantom enjambment between words. To sample just one more, earlier sonnet on the same theme of mutability, we note that sonnet 12 includes a lament for the cyclic passing of the summer foliage, "Which erst from heat did canopy the herd." The image of a "canopy," veil, or screen returns as an allusive trace in the next line's "summer's green all girded up in sheaves." This sonnet then closes with a tripled ligature that elides the twinned consonants g/'g, t/T , and 's/s across the sequentially enjambed gap between words. The result is again to depersonify the iconographic trope of Father Time—"And nothing 'gainst Time's scythe can make defense"—with an unfigural vernacular truism, the matrix under erasure: nothing gains time (against age and death).
Sonnets are of course, in verbal derivation, "little songs," sonnettes. In phonemic reading, however, enunciation as speech act is not the issue—but pronunciation as a linguistic fact decidedly is. Script, alone, cannot entirely exhaust its energies, and the unsettling excess—the edgy remainder—may operate as a return to differential origins from the seeming fixity of a written
wording. To isolate this reversion as a convertibility in the literary object itself, a functional ambivalence generated in processing the "object" as "text," is the work of the phonemic reading begun here.
"Porches of the Ear"
As we move from the ambivalent materiality of Shakespeare's lyric lines to a theatrical text whose phonic undulations would be at least to some extent disambiguated in the voicings of any given performance, questions of inflection and pronunciation become more prominent, though not finally determinate. How to do justice to Shakespeare's intonations and accents has long been a debate in the performance aesthetics of his plays. On the evidence of a remarkable parody of such debate in Henry Fielding, we are led to think of it as an "eternal" topos in Shakespeare studies. In the eighth chapter of Fielding's A Journey from This World to the Next , the deceased narrator comes upon Shakespeare, our deathless dramatist, holding court in the otherworld, arbitrating debates over "ambiguous passages in his works."[15] The Bard is at one point asked to adjudicate a dispute about the accentual stress and pitch with which Othello's murderous line "Put out the light, and then put out the light," should be intoned, and so interpreted, in performance. A parallel reading of the two clauses is suggested on one side, then an emphatic singling out of the second as "THE light," the light of lights. The narrator himself joins the debate to suggest instead "the light . . . THY light," but he is soon impressed with another reading "very sophisticated in my opinion"; without designating it as such, he is drawn to the idea of a paronomastic slant rhyme: "Put out the light, and then put out THEE, light," where the fuller voicing of the e in "the" elongates the pause between words into a virtual comma, allowing the article to thicken into a homonym for the doomed second person—as if in venomous travesty of a cliché like "light of my life."
The celebrated oratorial richness of Othello's speeches might well have provided further evidence of phonic contouring in this chapter, but it is Hamlet that most exhaustively thematizes the wiles of language, the unction and poison of speech, in company with a texturing sound play that saturates nearly every scene. The "phonetic and ideational interplay" (Booth again) prevalent in the sonnets is every bit as much in evidence, though less often critically glossed, in a dramatic text like Hamlet . Says the Prince, for instance, to Ophelia: "Nymph, in thy orisons / Be all my sins remembered" (3.1.88–89),[16] as if the monosyllabic "sin" could be purged by internalization within the very term for its remembrance in benediction. (And even here the paronomasia is not free from the transegmental drift of a widened internal rhyme ["orisons / my sins "] of just the sort the next chapter examines in terminal positions.) More often in this tragedy, however, the role of paronomasia is to encode
directly some overt and spreading contamination. This play about poison in the ear repeatedly submits its own phonemes to perverse dilutions and admixtures, the syllables split and spilt along the line. Rhetorician himself of these effects, Claudius would have his auditors intuit the leaking virulence of his nephew as the danger of one who would "envenom with his envy" (4.7.103), where the "etiologic distancing," as Hartman would call it, between imputed cause (desire, greed) and effect (pollution) is exposed in compression by the internal echo.
Again on the subject of a poisoning of ears, now literalized, the ghost charges that Claudius is the villain who "in the por ches of my ears did pour / the lepr ous distilment" (1.5.63–64). The syllabic figuration of these lines seems itself concentrated into a final phonemic juncture at the p/r of "leprous." Dell Hymes has suggested that poetry is often organized by a dominant nucleus of vowels transformed by consonantal variants.[17] Such a generative or transformational phonology, rather than grammar, serves to track euphonic and paronomastic effects back to a vocalic deep structure. This theory of the dominant nucleus may of course be subsumed to the notion of projected equivalences in Jakobson. A further example, with a doubled o sound more nearly equivalent in Shakespeare's pronunciation, appears in the internal slant rhyme of Hamlet's first soliloquy, where the sense of degenerate and unchecked luxuriance choking on its own excess is captured in a self-begetting thicket of sibilants: "'tis an unweeded garden / That grows to s eed. Things rank and gross in nature / Poss ess it merely" (1.2.135–37). The paronomasia of "grows"/"gross" may well be accounted for as a long o nucleus framed by gutterals and sibilants and then projected along the axis of contiguity — not only as a nearly perfect internal rhyme but as almost a full homophonic pun.
In moving from this general paronomastic terrain of Hamlet 's poetry to a specific crux like "A little more than kin and less than kind" (1.2.64), we can see the particular utility of the "dominant nucleus" model — with its transformational deep structure — for any partial utterance depending for its full power on what goes unsaid as well. In this respect, the relation of this model both to the "transformational poetics" of Hartman, with its overdetermined poles and elided middles, and to the invisible matrix of Riffaterre's semiotic expansions begins to take shape more clearly. Attempting to cajole Hamlet into accepting both his uncle's rule and his kinship, Claudius has just addressed the hero as "my cousin Hamlet, and my son" (1.2.63). Realizing that these blandishments have no effect on Hamlet's black mood, Claudius has then wondered aloud: "How is it that the clouds still hang on you?" (1.2.65). Echoing the uncle's cozening syllabic chime in "cousin . . . son," Hamlet has then rejoined, "Not so, my lord. I am too much in the sun" (1.2.66), a reply which not only mocks the heliocentric tropes of monarchy but, in its homophonic pun, seems to
suggest that the hero is on all counts too much in the role of "son" to suit him. Strategically lodged in just this larger field of phonetic reverberation is the famous ejaculation to which Hamlet now turns aside: "A little more than kin, and less than kind!" Taking "kind" as pitched between the meanings "natural" and "benign," between "of the kind" and "kindly," we thus overhear Hamlet mumbling at once about Claudius and about himself in a relationship too close for comfort, so near and yet so far: both incestuous, on the one hand, and mortally antagonistic, on the other. We are horribly more than cousins, he says, while I am far less than kindly disposed toward you, precisely for how far short you fall from the natural standards of your kind.
Then, too, as if the "more" carries with it almost a sense of prosodic quantity as well as spiritual quality, the weighted increase from "kin" toward what is here its virtual antithesis in "kind" passes invisibly through an unuttered but mediating third term. The complexities of the said circle the unsaid and entice it toward voice. Hamlet cannot name this transformation of the "kin" nucleus, nor even let slip a variant of its buried matrix, for such rancorous candor might consign him to final paralysis, spelling out the ambitious self-interest at the back of any revenge. Yet caught inaudibly between "kin" and "kind," or left in the lurch between them, lurks a ghostly overtone — lurks unheard, that is, by subversive reverberation, the tacit noun "king." More than "kin," "king" should be Hamlet's own title. Even in the silence between immobilizing polarities, the return of the repressed will out. Partly because the rhetorical schema of the line sets up the expectations of a guessing game ("What is a little more than kin and less than kind?"), we virtually await Hamlet's speaking the word "king" as a phonically clued solution to the puzzle. It may not seem accidental, either, that this threefold pattern recalls the triadic riddle asked of Oedipus by the Sphynx. The ultimate masterplot for Hartman's microstylistics is the elided maturation of Oedipus, the hero's internally truncated lifeline as an excluded middle between over-determined ends — in Oedipus's case, the overspecified terms of son and husband (of the same women). Hamlet's case, too, is figured by the extremity of the elided middle in a lexical form, the nephew effaced between a parent's death and a denied inheritance. Moreover, such an impetus is fostered by a phoneme (/g/) that never even appears in the text, a transformative cipher kept between the lines because it is interdicted by the psychology of the speech. If so much can be incident to so little, what then about a letter not interdicted so much as let loose in disguise, under cover of a dictional gap? Such is the fluctuant force field of an actually inscribed, though phantasmally detached, phoneme: resulting not from the blanking out of (forbidden) words but the activated and displaced blanks between them. A closer attention to this latter (transegmental) effect will follow some further discriminations among types
of Shakespearean sound play made operable across even a library copy as well as a theatrical experience of the "text."
Ghostly Play
Given the furious profusion in Hamlet of all sorts of lexical bucklings and permutations, it should be helpful to subdivide them on the way toward that most radical transegmental effect on which this discussion, as with sonnet 15, is eventually to concentrate. We will call their three chief manifestations, for convenience, "intralexical," "interlexical," and "supralexical" transforms. The least disturbing of the phonemic slippages, the one that respects most closely the lexical borders, if not the integrity of what they enclose, is the phonic tension that breaks down a word without breaking it open. To the examples already examined, we may add a double instance from the coda of Hamlet . About to become his own ghost in death, Hamlet commissions the retelling of his tragic history. Over the body of his friend, Horatio consecrates himself to this oratorical purpose:
So shall you hear
Of carnal, bloody, and unnatural acts,
Of accidental judgments, casual slaughters,
Of death put on by cunning and forced cause
(5.2.380–83)
Before the transposition (or metathesis) by which "casual" is switched to "cause," the monosyllable "acts" has already passed to its apparent reiteration, but only as a momentarily detached morpheme immediately expanded as part of a new epithet for random violence, "accidental." And just as "acts" carries a metadramatic overlay of the elapsed theatrical experience that has been comprised thereof, even as these sequential acts are about to be retold in narrative rather than theatrical discourse, so might the inner ear additionally mistake the second syllable (if there is one) of "forced" in a way that transforms the whole word into the punning "foresaid." What would thus be phonetically fostered is a metanarrative hint of the play's own previously written and performed incidents ("foresaid causes" already "put on," enacted, placed on stage by "cunning" — or theatrical conning) as they are now to be cast up into report. Indeed, the earliest cause of enmity and violence mentioned in the play is the usurpation of Fortinbras's territory — "all those his lands, / Which he stood seized [possessed] of" (1.1.88–89) — by Hamlet's father, the King. When these holdings are alluded to again fifteen lines later as "those foresaid lands" (1.1.103), the meter openly invites the pun on a disyllabic "forced." The homophonic twist on "foresaid" is thus a turn of
phrase that is arguably to be reversed five acts later in yet another retrospective allusion to all that antecedent bloodshed — beginning with the "foresaid" lands "forced" from an enemy of the state — that has led so inexorably to the play's funereal denouement. By even the differential minutiae of such a subtending phonemic webwork can a play's overall thematic symmetries be brought to light — and to bear on the text in production.
It is clear that "foresaid " is a complete homophonic transformation of the word's second syllable, whereas "acts -idental" is merely a temporary (morphemic) deflection. Still, in regard to their retention of the outer boundaries of the lexical unit — however unorthodox their internal syllabic manipulation — both belong essentially to the first type of transformation (intralexical) under analysis. Neither of these effects outplays the lexical frame that generates it. There is by contrast, with the next category to which we move, the tendency toward a momentarily breached word boundary. Though this distinction does not necessarily emerge from the performance of the text on stage, it has considerable ramifications for a theory of phonemic free play. There is at least the off chance that such a lexical breach is operating in Hamlet's climactic speech in the graveyard. Coming round to his long deferred echo of the "To be or not to be" soliloquy, there is the capping disyllabic sentence: "Let be." The complete brief text of Hamlet's transfiguring acquiescence of course goes like this: "The readiness is all. Since no man of aught he leaves knows, what is't to leave betimes? Let be" (5.2.222–24). Across the syntactic accumulation of this thought, and in colloquial continuity with it, flashes the momentary idiom "what is't to leave be ," as in to let alone . The thought holds, even as the syntax converts it into its corrective extension. No sooner said than undone, the second syllable of the fleeting verb phrase, "leave be," is absorbed into the adverb for early, "betimes." It is just then, in turn, that the very idea of leaving be is revoiced in the more capacious and relaxed alternative, "Let be" — where something may be coming as well as being put behind.
There is another way to upset the reading sequence without ripping apart a lexeme. Beyond (1) intralexical transforms — the single word either extruding another from a syllabic warp (the approximation of "son" in "cousin" or of "sin" in "orison") or twisting itself into a wholesale homophone ("grows"/"gross" or, in a single pun, "forced" as "foresaid") — and beyond (2) interlexical displacements — one word shearing off from another for some momentarily transformed phrasal attachment ("leave be =times") — there is the related phenomenon of (3) supralexical conflations — two separate words fusing as the syllabic components of a third, a finessing of the zero juncture that creates both a new aural impression and a new semantic expression. This third category of effect tends to be less fleeting and diaphanous than such a junctural shuffle as "leave be=times," gone almost before it is registered. An
illustration: after impugning Gertrude's "judgment" in her choice of Claudius (3.4.70), Hamlet in the next line refers to "sense" (for feeling, rather than intelligence) in his backhanded, punning acknowledgment, "Sense sure you have." With its almost inevitable elision (in the mouth of an actor) between s(e) and s , the phrase thus lets slip the very "censure" which the whole speech intends.[18]
We find another example of such lexical (con)fusion — somewhat more debatable in philological terms but present at least as a near miss, an antipun — in the opening lines of Hamlet's first soliloquy, "O that this too, too sullied flesh would melt, / Thaw, and resolve itself into a dew" (1.2.129–30). Backward to start with (if not simply pleonastic), since thawing would be expected to precede melting in any normal process (or at least to be indistinguishable from it), the second line's at best tautological series serves to invite another sense of "resolve" that dislodges the image altogether from the overdetermined context of deliquescence.[19] Conjured here, if only in words, is the resolution , even for suicide, that Hamlet cannot otherwise summon. Nor is the deepest desire expressed in the line further dimmed or appeased by the completion of the figure with "into a dew." Obviously, Hamlet speaks metaphorically, but the metaphor is itself double-voiced and returns a literalism from its underside. A reconstituted lexical bond resolves the scripted article and noun into a disyllabic homonym. With certain philological reservations about the precise anglicizing of the French word in Elizabethan English, the result is nonetheless that the wish fulfillment of a suicidal "resolve" commits the self to an overtone of life's final adieu.[20]
As I began this chapter by suggesting, "Shakespeare" in these investigations is not a proper name so much as the common term for a literary property: for a certain euphonic, polysemous roughening of the textual surface which has become the very norm of verbal wit in English — a virtual thinking in words and word sounds. Shakespeare is thus the password for that definition of the literary within which the writer by the same name wins his preeminence. If Shakespeare is a field of textual phenomena, rather than the source or hero of a private literary idiolect, then traditional stylistics does indeed stand in need of revision by a more impersonal categorization of perceptual (or receptual) effects, as with the thumbnail taxonomy of lexical and phonemic mutation here in progress. In this regard, to distinguish a Renaissance reading of, or attendance at, Hamlet from a modern-day counterpart (not equivalent), while locating both within a developing experience of the Shakespearean text, should only confirm our sense of Shakespeare as a receptual rather than a biographical category. On balance, there is certainly as much phonic density in the Shakespearean texts we read in their modern English "translation" as in the "originals"; it is just otherwise distributed,
located along different orthographic axes. Sounds may differ, but the sounded density remains, and remains in the form of an instigation to what we can only call thick reading . Since it all depends on how you say it, no matter whether out loud or to yourself, the "Shakespeherian Rag" is thus recognized all the more openly here for what it all along has been: a reading effect .
In the phonemic mutations discussed so far, we have noted three modes of morphophonemic transformation still bounded by the logic of lexical parameters: (1) intralexical — the homophonic syllable, the part echo, or the full homophonic pun; (2) interlexical — the momentary breaking-off of a phonemic cluster as free vocable; and (3) supralexical — the outright fusion of adjacent but lexically discrete morphemes in a new third term. In addition to these is perhaps the most common, potentially the most disruptive, in practice the most muted, and in criticism certainly the least discussed of all the phonic convolutions of Shakespeare's language — or of any other writer's. This is the transegmental drift that leaves at least one of the abutting lexemes no longer intact. Staying a while with Hamlet , we can now draw into the open some examples of this fourth, aurally diffusive category, as already noticed in sonnet 15 in the glancingly asserted "engraft (t')you (a)new" or, more faintly and fleetingly, "grows (h)old/s." This fourth type of (transegmental) mutation is characterized by the vanishing act of a single phoneme — backward or forward, by ligature or elision — in transit from one lexical unit to an adjacent one. It therefore constitutes an ambiguation of juncture at the level of the single phoneme rather than a mobile and adhesive morpheme or lexeme. Its exemplary leverage in this study derives from its refusal to respect even the morphemic, let alone lexical, borders on which traditional stylistics depends.
The spectral elocution of the father's ghost offers illustration within Hamlet 's thematic of defiled hearing. With well over half a dozen iterations of the call to Hamlet's audition — including "List, list, O list!" — massed behind it in the first scene, the ghost's indictment of Claudius climaxes the motif of hearing with the synecdoche "ear of Denmark":
'Tis given out that, sleeping in my orchard,
A serpent stung me. So the whole ear of Denmark
Is by a forged process of my death
Rankly abused.
(1.5.35–38)
In performance — given that our hearing, along with the hero's, has been whipped into febrile attention — we could barely help catching for a split second the drift, the rift, the junctural upset of "serpent's tongue." It is a spectral anti-pun from a ghost's own mouth.[21] Close upon, closing upon, this mere flick of the tongue is the syntax that infinitesimally shifts its ground and
shuts out the pun's very possibility, putting the second hissing sibilant back in its place as frontal alliteration. For a moment, though, in this metalingual irony of phonemic ambivalence, we eavesdrop upon the revelation not only of murder but, proleptically, of the forgery of report that follows — and for which the same villain is to blame. Hamlet is, after all, a play where death's sting is very much in the tongue, both in the surviving report that falsifies the crime and in the posthumous obligations that giving voice to it entails.
Beyond this reflexive ambiguity — a quintessential slippage in this play of ghostly phonemes — there are other sorts of aural conversions at lexical intervals. Conducing to reintegration in a new word, they do so within a plausible new syntax as well, as "serpent's tongue" does not. Says the ghost to Hamlet: "I could a tale unfold whose lightest word / Would harrow up thy soul" (1.5.15–16). Few actors could manage, or would trouble, to disambiguate the junctural slippage and ligature of "whose lightest" into "whose slightest." Equally difficult to keep from activation is the flickering counterphrase hovering over the next dire consequence of which Hamlet is warned by the ghost, whose story would "Make thy two eyes like stars start from their spheres" (1.5.17; my emphasis). Again a drifting, this time backward, of a sibilant, this time doubled, creates an auditory elision. From this telescoping of the syntactic gap is released, by an ambiguity of the dental d/t sound, the equally likely "stars dart": in a sense the momentaneous effect of the cause of "starting" in the first place. Phrasing becomes an etiology on the run, microscopically (or microphonically) contracted. The pressure of the unsaid matrix behind this strained kinetic phrase — the etymological link to "startling" — is not quite sufficient, that is, to suppress the transegmental alternative, since it too operates within the semantic paradigm of functional choices.
Upon the ghost's first exit, before he meets with Hamlet, it is Horatio who describes how "Th' extravagant and erring spirit hies / To his confine" (1.1.154–55). The erring of a phonemic ghost, too, in the root sense of wandering (hear errancy ), is also at play here. Eventual English teachers, among others, learn somewhere on their way to graduate school, usually on the occasion of an aphorism from Pope, that the word erring, if pronounced as in error, is in fact a mistake, since it should rhyme, for instance, with deferring . But this softening of the vowel sound had not yet taken place by Shakespeare's time,[22] and so the polysemy of this line from Hamlet is closer to pure segmental ambiguity than the educated theatergoer might today expect. "Extravagant and erring," given the etymology of "extravagant," thus comes bearing a redundancy that is partly done away with by the antiphonal drifting of the d into a reduplicating liaison — and hence into a transforming consonant in the unwritten but aurally unavoidable "extravagant and daring." Two paradigmatic axes are involved here, one linguistic, one contextual. The phonemic
alternation is allowed only because the lexical alternatives of "erring" and "daring" are also operable, in context, within the binary system of semantic and thematic choice. In their slippage they still fit the sense.
A close parallel in phonemic contour to this sliding recombination of dentalized phonemes returns us to the play's motif of the violated and defiled ear. The new king is attempting to poison Hamlet's reputation by accusing the Prince of poisoning Claudius's own. Not only does the usurping monarch claim that rumors "infect" Hamlet's ear (4.5.90), but he shortly adds that, as a result, the diseased Prince never hesitates "our person to arraign / In ear and ear" (4.5.93–94). Performed as prompt text for the "ear," there would again be an immediate doubling and liaison at the consonantal borders of the two particles, the preposition and the conjunction. An infectious drifting from word to(ward) word would thus discover in this emphatic, this tautological repetition of the noun "ear" the fractionally displaced paragram of an idiomatic matrix or cliché, "in near and dear." As never more exhaustively thematized in Shakespeare's work, the ear indeed is in dis-ease.
With wording itself out of joint, Hamlet, then, is a play whose preternatural quality is partly conveyed by the spookiness of its sepulchral echoes. If poison is to be poured in the ear, the ear must have porches. Since even the casual is causal, acts can't be accidental, which goes for linguistic acts as well. By its very nature, here its verbal nature, what envy does is to envenom. This is a play where cousins are cozened by the name of "son," where unkind kin vie for kingship, and in whose garden all that grows is, by acoustic contagion, gross. "Shakespeare" thus names a verbal field where words perpetually fertilize each other, sometimes in decomposition, crumbling and recombining, surrendering the envelope of their separate definition to mix, blur, and confound even their own lexical identity. One thing leads to another, lends to another. Paronomasia narrows to an interplay between (in the sense of across ) words, across their segmental breaks: again, wordplay as word-splay.
A Never-Fixed Mark
If paronomasia is the prototypical instance in Jakobson's definition of the poetic function, how then do these transegmental flashes and dodges either fit or revise that definition? One might say that they simply project comparison into combination, equivalence into sequence, at an angle of less deflection from the vertical, or paradigmatic, axis, so that the equatables or alternatives are overlapped within a single equivocation. But this adjustment of Jakobson's definition still fails to account for the unsettling effect in reading produced by these disjunctive lexical overlays. By contrast with such equivocal contiguity, the echo of "grows to seed" in "rank and gross," for instance, is separated almost as decidedly as end rhyme, falling indeed in a separate line. This sort of
chime quite audibly lifts words from sequence and stacks them on top of each other in the mind's ear, leaving the grammar intact. In such cases, the chain of succession in syntax remains the baseline upon which phonic recurrence is projected as a secondary feature. In the transegmental dislocations we have been examining, however, syntax does not remain a stationary grounding. With "stars (s)tart" (as a far limit of paronomasia in almost pure, undelayed doubling), if we hear them "start" and "dart" at once, we do so by an infinitesimal shift within the same grammatical collocation: "stars (s) = t/dart." By just this logic, however, the very axis of succession itself seems suddenly destabilized, the combinatory logic of semantic production no longer secure. At issue is not just a choice within the lexical paradigm, since "start" emerges as if by stealing a sibilant from the preceding plural. That its loss is not syntactically noted, that the plural stays put, does not mean that one axis has not encroached on the other. If straight paronomasia is the quintessential projection of equivalencies from the axis of selection onto that of combination, then such transegmental slippages might instead be read in part as the reverse projection of successivity back upon the axis of alternate selection. It is this spectrum of substitution, kept preternaturally active, that is now the operable though malleable baseline. Upon this shifting ground the previous security of grammatical sequence is mapped in all its sudden oscillation and vagrancy.
Whether or not the transegmental drift is taken to skew or directly to invert the coordinates of the poetic function as rendered by Jakobson, it does tend, as suggested earlier, to shadow the individual speech act (parole ) with its basis in language itself (langue ). As against signifying practice in other semiotic systems, language is based simultaneously on two stages of difference known to linguists as "double articulation." The phonic raw matter of speech is differentiated into a set of oppositions pertinent to a given language, as, for instance, the functional distinction in English between d and t . Upon this system of distinctions is mounted the difference between various bondings of phonemic matter into the units known as morphemes and lexemes, as in the functional opposition between "start" and "dart." This all takes place within the paradigms of the langue, before being further caught up in the metonymic linkage that bonds word to word in such a syntactic chain as the intransitive present-tense predication "stars start" (or "dart"). The options of the langue that continuously stand waiting for the selections of parole are ordinarily subsumed entirely to the final signifying function of any coherent grammatical utterance. That segmentation which undergirds the morphophonemic structure of the lexeme, in other words, is usually ignored — sentenced to suppression — once the lexeme passes into continuous utterance. If we take the transegmental drift to project this normative march of the horizontal, or
metonymic, sequence back into the paradigmatic axis, we must understand it to do so in a way that causes a temporary regression or breakdown, a reversion from selective speech to the dormant fluctuation, even turmoil, of language. "To hear with eyes" can in this sense be an audition of flux and irruption beneath and between the graphic signifiers on the page. The study of such disruptive moments becomes the monitoring of linguistic differentials as they impinge upon and disintegrate the sureties of parole, a study of langue itself — or "languageness" — as the return of the repressed.
It is a return to which the Shakespearean phonotext is endlessly hospitable. Among the widely noted puns on "Will" in the sonnets, there is this often-glossed moment from sonnet 136: "Will will fulfill the treasure of thy love, / Ay fill it full with wills, and my will one." Booth remarks on the "gradual revelation, increasing overtness, and mounting crudity of the pun on fulfill" (SS, 470), though not the play on affirmation as identity — the economy of an "Ay" for an I — that further crystallizes the self-nomination of "Will." Nor, in a transegmental mode, has anyone noted a related punning effect on Shakespeare's first name, Will's "will=full-"ness. As a lateral gambit of the text not contained by normal lexical borders, this effect emerges as a seeming impulse asserted from within that autonomous model of desire which is language itself. It is no accident of literary history that this telescoping process of one word from two finds its homage by reversal in Joyce's own later play with the last name of Will's wife, Ann Hathaway, decontracting it into a homophonic willfulness all its own: "If others have their will Ann hath a way."[23]
At least one such effect in the sonnets is even pitched explicitly between the claims of eye and ear, of script and presumed utterance. "Why is my verse so barren of new pride?" asks the persona of sonnet 76, lamenting that his lines are not given to "quick change," to "new-found methods, and to compounds strange?" Before even bringing the interrogation to a halt, however, the Shakespearean text has instanced its own quick-change artistry in the "compounds strange" of a segmental regrouping, as it asks why "every word doth almost tell my name?" (emphasis added). Though Booth notices in a later paradoxical figure from this same sonnet, "Spending again what is already spent," that there is a "muffled pun" (read: a muted junctural breakdown, otherwise an intralexical drift) evoking the "compound(s) strange," as it were, of "a=gain," and though he elsewhere glosses the verb "stell" for "to carve" in its relation to writing (to stell, to steel, to style [SS, 172–73]), neither his nor other commentary registers the transegmental "inscription" by which "almost tell" — that dead metaphor of vocal utterance — becomes (more accurately, in textual terms) "almo(st) stell." Here is the transegmental ligature of an entire diphone rather than single letter — and this carrying a suggestion about the
graphic rather than phonic imprint of poetry (a hint partly complicated, of course, by the aural form in which it reaches us).
We are listening, after all, to the same writer who, as dramatist, puts such a phonic drift into the voice of the chorus in Troilus and Cressida . After Ulysses's speech about the wanton and sluttish wiles of Cressida, the lascivious "language in her eye," a choric voice announces, by way of equivocal enunciation, "The Troyans' trumpet" (4.5.64) — where the expenditure of the sibilant breath famously blares out the phonemic and moral slur "s=trumpet." There is even an internal literary history, or intertextual lineage, ensuing from this morphemic joke within the sequence of Shakespearean drama. Two years after Troilus , in the next of the tragedies, Iago slips almost unconsciously into the same pun, uttered in soliloquy by a subliminal logic of association. Looking on at Cassio taking Desdemona's hand, he has begun hatching his adultery plot when Othello's arrival interrupts and refocuses his scheming. Putting himself on alert with "The Moor! I know his trumpet" (2.1.178),[24] Iago is also implicitly congratulating himself on having conceived his stratagem to recast the doting wife in the role of whore. This is, after all, the same "Shakespeare," the same punning field, that resuscitates the chronicle of the ancient "King Leir" as a privileged testing ground for that intensified blind willfulness which is "kinglier" only by tragic definition — a vein of wordplay, turning elsewhere on the "Royal l EAR" (1.1.139) of the king's attention, that is further exploited in Jean-Luc Godard's film version.[25] This is, as well, the same Shakespeare, the same homophonic field, in which a dialogue between Romeo and Juliet, where the notion of sinning lips is thrice on their lips (1.5.107, 108, 109), is followed by Juliet's renowned rhetorical question at their next meeting; as precipitated by the liaison of a sibilant and enhanced by the iambic emphasis on the second syllable, Juliet's phonemic lapsus will often be heard from the stage as the thematic distillant: "What sin a name?" This ambivalent lexical s/play asks as well as answers that deeper question about Shakespearean verbal pluralism: what's in a word? Nothing that cannot slip past in the happy fault of a redistributed lexical caesura, a dropped and recovered downbeat in the syncopated rhythm of the "Shakespeherian Rag." Even when not uttered from the stage, the written page has its own phonemic enactments, its own plays-within-the-play of reference. In that famous sonnet about love's necessary unalterability (sonnet 116) — to take a last reflexively thematized example — the very negation of a love so constituted rises to haunt all protestation from within the fickle signifying system of "a(n) (n)ever-fix'd mark," one mark after another on the track and trail of desire.
But why — one may ask again — should the author of Romeo and Juliet, Hamlet, Lear, or the sonnets be called upon to begin these proceedings?
Shakespeare serves to "authorize" the phonemic reading of this study not in his role as the poet who first posits the modern (lyric) subject — as in the convergent claims of Anthony Easthope and Joel Fineman — but rather as the writer and performed playwright alike who positions us before a certain kind of literary textuality in English.[26] Shakespeare has come to be read as if he taught us how to read, as if his texts are the wellsprings of those specifically literary inflections of language that are at once most prized and most often appraised in the processing of a poetic text, whether by reader or critic. The term "Shakespeare" thus stands in retrospect as the primary manifestation of that sustained and extreme verbal originality, that driven invention, density, and obliquity, which is literary language. With "his" lexical and syntactical eccentricity, concatenation of imagery, and sustained verbal opacity, "Shakespeare" names an exemplary case of that working of words beyond their referential service which has come to be called poetic. And, by circular reasoning, whenever we discover such literary "values" in a single text by Shakespeare, they only serve to confirm again his priority, his mastery, and his influence. The exemplary status of Shakespearean textuality resides therefore in the fact that such textuality defines for the act of reading not predominantly the modern discursive subject but, more decidedly, the modern subject-position of the reader, adrift amid a constant play of signifiers, a subject split, doubled, ambivalent, layered, elided, and in flux, moving forward only by doubling back, proceeding by reprocessing, reading by rewriting. If grammar, and certainly rhetoric, work to sustain this discursive positioning, at the same time there is something in what we might call the fact of the linguistic that also serves to contend with and contest the subjective grounding of utterance — equivocating it through the fluctuant wash of a phonotext polyvalent and unspeakable. The phonotext, in short, cannot be englobed or totalized as voice.
It is in this sense that the "enunciation" of discourse needs qualification by its always contingent "evocalization" as text. Poetry as produced language inflects and bedevils, strains and unfastens, poetry as discourse. The reigning ideology in any period may be one of containment, but the material base itself is always insurgent. This insurgency is what, in and of itself, we call literary, as well as what is often meant by the Shakespearean preeminence, or priority, within the history of literariness. In this closed conceptual circuit, the objectivity — the material basis, the inscribed signifiers — of a text, processed in the receptual "subjectivity" of reading, induces as secondary effect of their effects the very subject-positioning upon which the reading in the first (never fixed) place depends. Given what we have come to expect as readers of literature in the modern English tradition, it is precisely as readers rather than as psychological agents that the Shakespearean text reads us in the making, which is
another way of saying that it makes us in the reading. Stylistics has traditionally explored the "structuring" activity in only one sense: the authorial manufacture of the text. A concentration of the phonotext would, by contrast, entertain literary production as a construction of the reader, in the double genitive sense (both "by" and "of"): denoting text production on the one hand and subject formation on the other, a twain that never fails to meet—but never for more than a word or two, for one word on the verge of its relation to another. Shakespeare provides, then, less the literary-historical than the prototypical starting point for the present attempt at a phonemic reading of the transegmental drift—that structural biplay which is an exemplary instance of the fluctuations of literary discourse. This reading procedure, with its emphasis on the somatic as well as psychic dimension of the positioned subject, needs now to extend and deepen its reserve of examples before theorizing further the place of the voice body in the "subjective" space of reading.
2—
Rhymed Treason:
A Microlinguistic Test Case
Syntactically seditious, lexically anarchic: an arc of echo across unbonded, nonconfederate syllables—contaminating in the process that bastion of conservative poetics, the regularized forms of metered rhyme. Without in the least backing off from the claims of Chapter 1, we should now be able to watch certain of the verbal operations examined there as they get mobilized in a unique sort of rhyme, one engaged just fractionally before an echo is usually thought to set it. In play at such moments is the phonetically rough left edge (rather than merely typographically ragged right) of the rhyming cluster, an edge abrasive and absorbant: a lexical edging back. To move from the "unwritten" cross-lexical biplay in Chapter 1 to one of the most canonically inscribed of poetic features—end rhyme—should get the discussion, if only temporarily, back on common "stylistic" ground. But as the classic touchstone of poetic surplus, rhyme will also serve here as a test of phonemic surfeit, of segmental overflow.
We may begin with a radically unorthodox and entirely "modernist" rhyme, one that does not even adhere to the protocol of line endings. In Gerard Manley Hopkins's "Wreck of the Deutschland ," the line "She drove in the dark to the leeward" finds a rhyme for its last syllable, ward , only across the enjambment of the two subsequent lines: " . . . night drew her / D ead to the Kentish Knock."[1] This is a leeward as much as a wayward rhyme, one that will go any way the prevailing wind takes it. Driven here to the point of morphological mayhem is a dismembering rhyme that treasonably betrays both lexical and metrical borders, resting on what might be termed a double enjambment: first between words, "w=[h]er," then between whole lines, "er/=D." A rhyming cluster edging or lurching backward is common enough, as we shall see, but an echo toppling forward like this is rare indeed.
To begin with the spectacular exaggeration of the latter is meant to throw into relief the overlooked frequency of the former.
It happens that three of the most important contemporary essays on the question of rhyme—William K. Wimsatt's "One Relation of Rhyme to Reason," Hugh Kenner's "Pope's Reasonable Rhymes," and John Hollander's Rhyme's Reason —all highlight the logic behind the nonincidental coincidence of sound, the variable semantic overtones that rhyme sends into reverberation.[2] I certainly do not mean to discount the possibility of thematic links afforded and forwarded by rhyme, even though I will not here be pressing such points. This chapter is designed, rather, to register a break at times with rhyme's authorized grooves or tracks. Implied by "rhymed treason" is the kind of lexical insurrection that results from the overlapping, and so undermining, of word boundary in certain transgressive—namely, transegmental—rhymes. There are, it thus seems, more echoic constellations than normally admitted which need not be seen to be heard, nor even intended in order to be read. As a matter of fact, the whole argument from intentionality goes by the board if, within the undeniable return of acoustic material, we hear not only more than ever bargained for but more than can plausibly be absorbed within a given prosodic economy. Rhyme, that is, turns traitor to its own expectations, revolts against its own formulas, when its acoustic mass disturbs into recurrence more phonemes than the apparent verse contract requires. Rhymes generated by such systemic excess may be accidental, even counterproductive, but there nonetheless, audibly, they are. Thrust into the foreground by the force of echo, detached from the sole authority of syntax, such loosening through recurrence creates a fractional backsliding in a given line. Rhyme, in short, does not always toe the line ends.
More Than Meets the Eye
For the present purposes, the truancy of certain rhymes will help delimit by contrast a primary normative locus of phonological attention in the study of poetry, one too narrowly restricted to the regularities of meter and lineation. Then, too, rhyme offers an inviting experimental field on which to explore the sometimes discrepant work of graphemes and phonemes in literary (e)locutions—precisely because of its own long-acknowledged division of labor, at times, between eye and ear. From the "new"/"true" rhyming couplet of Shakespeare's sonnet 15, for instance, down through Yeats's rhyme in "A Prayer for my Daughter"—"yet not"/"distraught" (ll. 17–18)—there is a kind of rhyme that honors only the requirements of the ear, not the recognition of the eye. At the other extreme, of course, is the hypertrophy of lexically complete homophonic rhyme—called, after its more common French occur-
rences, rime très riche . An example from Chaucer, in whose Middle English such effects proliferated, is the homophonic and all but homographic echo in "The Legend of Cleopatra" of "liven may " with "rose in May " (ll. 612–13). By contrast, the nature of English orthography often prevents the instant recognition—unless pronounced—of such a subsyllabic homophony as Yeats's "not "/"distraught ." There is another way, as well, in which the graphic can mask or occult the phonic; homophony may go unnoticed in its full sweep because of lexical format itself, regardless of spelling. Certain phonic recurrences thus slink between ortho/graphic units as a surreptitious link. With the previously discussed tension between graphic and phonic signification in mind, consider now the common denominators of the following rhyming couplets, distributed across seven centuries of English verse:
For thus saith Salomon that was ful trewe,
'Werk al by conseil and thou shalt nought rewe.'
(Chaucer, The Miller's Tale, ll. 421–22)
KNOWLEDGE. Your Five-Wits as for your conselors.
GOOD DEEDS. You must have them ready at all hours.
(Everyman, ll. 663–64)
Sans-foy his shield is hang'd with bloudy hew:
Both those the lawrell girlonds to the victor dew.
(Spenser, The Faerie Queene, 1.5.5)
And lovers' houres be full eternity,
I can remember yet, that I
(Donne, "The Legacie," ll. 4–5)
Sees by degrees a purer blush arise,
And keener lightnings quicken in her eyes.
(Pope, The Rape of the Lock , 1.143–44)
Some hungry spell that loveliness absorbs;
There was no recognition in those orbs.
(Keats, "Lamia," ll. 259–60)
Came tamely back in front of me, the Drover,
To suffer the same driven nightmare over.
(Frost, "Our Singing Strength," ll. 44–45)
There are rhymes more or less apparent to the eye in every pair of lines except those from Everyman, where the imperfect mating (almost slant rhyme, even by contemporaneous standards of pronunciation) remains in a sense true to the eye only by being visible as a near miss. At first glance, the rhyming matter of these couplets would run as follows: "trewe "/"rewe, " "counselors "/"hours, " "hew "/"dew, " "eternity "/"I, " "arise "/"her eyes, " "absorbs "/"orbs, "
"Drover "/"over ." As such, the phonic consorts are not more exact than those in Shakespeare's "new "/"true " rhyme, for instance—just more visibly so.
But they are something more as well, something in addition, in lateral (and not immediately visible) accretion. They produce a different difference, as it were, between eye and ear. Rather than defying the eye, they simply offer more than meets it: a broader band of rhyme than appears at first glance. No handbooks or commentaries on rhyme pay any notice whatever to this accretive swath of rhyme, this magnetic force that pulls into its field stray phonemes. Yet it would surely seem that the audible—or, if unspoken, at least the aural—effect of these rhymes should properly be reschematized as follows: "trewe "/"nought rew, " "counselors "/"all hours " (or even "counselors "/"all hours "), "bloudy hew "/"dew, " "eternity "/"that I, " "arise "/"her eyes, " "absorbs "/"those orbs, " "Drover "/"nightmare over ." All but one of the rhyming pairs (where dy =hue actually "predicts" dew ) happen instead to begin with a lexeme that is answered by a transegmental phonemic cluster never gathering to a new semantic or lexical unit—but nevertheless urged upon the ear (even though unusable by syntax) by the preceding lexical model. Broadened from the rhyming minimum, these transegmental accords do not in any way render a more "perfect" rhyme—just a more dispersed one. They are not simply to be grouped with the condensed repetition of that rime très riche on "may" and "May" noted before in Chaucer's "Legend of Cleopatra." In that same tale from The Legend of Good Women, we also find examples of the more common rime riche . Such rhyme adds an initial consonant to the mere sufficiency of the closing vowel (or vowel-and-consonant) cluster, as in the parallel syllabic rhyme "usaunce "/"obeysaunce " (ll. 586–87) and even the subsyllabic recurrence of "atones " in "stones " (ll. 638–39). Between such "rich"—such fully invested—echoes and the minimal return of ordinary rhyme falls an additional accord—one unidentified in scholarship but detectable in the reading act. It falls there by falling between words, a richness whose enhancement comes by way of an extract from the adjacent term: a lexical lending as blending. In "The Legend of Cleopatra," as it happens, we do find such a compromise or median rhyme when mention of the queen's "purple sayl " is sounded against the "strokes" of the oars "whiche that wente as thikke as hayl " (ll. 654–55). This rhyme also includes a semantic redundancy capable of functioning in the larger sense of the line—as if, for instance, the oar strokes were exactly as rapid in their flailing as is the flapping of her windswept sail. Semantics aside, this cross-lexical echoism has passed beyond the formulaic densities of either ordinary or rich rhyme by falling between (as well as overshooting) them. It boasts a richness in excess of the minimal accord without being constrained by the delimited syllabic envelope of rime riche proper. It generates in this
way a destabilizing excess born of phonemic subtraction from the lexical precursor. Such a rhyming pattern, intended or not in a given case, is often nevertheless inevitable, given the nature of English "word processing." Or call it a rhyme only if it is taken as an intended aesthetic effect, an inadvertent echo otherwise: there it remains, between words, between them and the reader. There it is, after all, in Chaucer's "liven may "/"rose in May ."
The aesthetic acceptance of rime riche in both French and English poetry of Chaucer's day would in fact tend to encourage the sort of transegmental or assimilative echoes—as active rhymes—found in Chaucer's "for age "/"forage " ("The Reeve's Prologue," ll. 3867–68), where the charm of rhyme is not dependent on the avoidance of direct echo. This may in turn argue for the intentional force of Chaucer's "sayl "/"as hayl ." Aesthetically validated rhyme or not, however, the echo would be no less operable in the phonotext if the same line endings were found in Shakespeare, Milton, or Keats, poets in a modern English tradition not inclined to the symmetries of rime riche . Whether such later poets would have noticed these marginal phonemic assimilations and settled for them regretfully, or whether—in certain cases, from poet to poet and verse to verse—writers from Shakespeare to Tennyson, say, would have actively courted the increased textural possibilities provided by such a drift of acoustic iteration, this question need not escape the realm of sheer conjecture to justify consideration of the effects themselves. In roughly the place of rhyme, they offer more of the same, even if the same as different: formal rhyme lapsed (backward) to irruptive chime. It might be argued that such cross-lexical slippage is merely an aleatory offshoot of English morphophonemic structure, a risk of word division that has no formal place in English poetics between Chaucer and the modernist extravagance begun, roughly speaking, with Hopkins. Still, the inevitable slack within that locus of heightened phonemic sensitivity induced by rhyme might well serve to detect, once again, the roots of modernist disruption long before its systematic and canonical practice.
The issues raised by all such lateral play within rhyme begin in well-established theory and quickly move to test its limits. According to Jakobson's definition of the poetic function, rhyme shows forth as the most visible form of the "poetic" in a structural sense. To be sure, rhyme markedly projects equivalence from the axis of comparability ("profound," say, in its likeness to "resound") into that of combination (the sequence that laces them together as part of the same syntactic progression). Further, rhyme serves to activate the dead metaphor of axis itself as a real graphic figuration. The imagined horizontal armature of combination appears broken-down, pieced out, and stacked up, length upon length, metrically calibrated, until the vertical axis becomes visible (for the reading more than the seeing eye) at the right margin.
Granting that end rhyme thereby graphs equivalency as a vertical line of descent, at right angles to metered lineation, what then happens when this stacked axis of visible (or sometimes only phonic) accord becomes just slightly staggered, terraced, or tiered? Such variously telescoped echoes are certainly to be distinguished from the mere rhyming of a two-syllable word with the complete sound elements of two parallel monosyllables (usually called "polysyllabic" rhyme). The effect of a tiered rhyme, instead, is enhanced by the interlocking of two adjacent words at their border, whether for the augmentation or the concentration of the rhyme. The result is usually to widen the rhyme by attracting the closing phoneme(s) of a bordering lexical unit, by ligature; alternately, by elision, a contiguous phoneme may draw off its phonetic double (exact or approximate) in a way that isolates the rhyming matter into a more discrete node of echo. The linkage of words is performed, that is, either through an aural ligature—an acoustic holding over of one sound—or, instead, through a knotting up that bonds (in effect, by shortening) the phonemic span. In either case, a related musical term, the tie, would seem adaptable to poetic scansion as an invisible notation that overlaps and blurs words at their edges. Echoes generated in this way are the most purely phonic of all rhymes, defying even the intervals of script in order to forge an equivalence from the graphically discrete increments of combination.
Derrida's expansion of the category of rhyme to cover all the contingencies of iteration in poetry begins with a grammatological perception very close to the tenets of Jakobsonian structuralism. "Rhyme—which is the general law of textual effects—is the folding-together of an identity and a difference."[3] Jakobson might rather stress the unfolding, the displacement of identity into difference along the track of the consecutive. But from Derrida's perspective, rhyme is certainly an exemplary codification: an identity within difference that derives from the simultaneous orchestration of chance and design, of randomness and rule, precipitating from its first term a difference never predictable in advance, a deferral of return along the axis of advance. But a crucial aspect of all this eludes Derrida's comment, if not always his notice as a "practical" critic. The axis of syntactic sequence, whether subdivided into metrical lineation or not, also involves another crucial difference within sameness: the homophonic ambiguities of junctural distribution. In tie rhymes, the axis of contiguity becomes the axis of perpetual recombination, a lexical successor recruited in part as adjunct or increment rather than merely neighbor of the word before. Difference is thus generated from within the sameness of the lexicon itself (a word's or syllable's supposed acoustic identity with itself). The options of rhyme are transformed (whether widened or compressed) whenever the boundary matter of one lexeme is recruited in this way for a new morphological combination, whenever it is invisibly con-
scripted into the phoenmic mass, but not graphemic integrity, of an adjacent word—regardless of whether any new semantic melding results. With a single word failing in this way to retain its phonic matter untainted by acoustic contamination, its role as rhyming unit may well be either to annex or to exile a phoneme or two in order to maximize its sameness in recurrence. The differential shifts within the increments of a syntactic sequence are thus, in short, projected as the sameness-within-difference of a rhyming play capitalizing on just such fraying segmentation.
Shakespeare's verse deploys or exploits—or permits, or at least fails to prevent—such drifting rhymes in both initial and echoing slots. When "In the old age black was not counted fair" is rhymed two lines later in sonnet 127 with "But now is black beauty's successive heir," the closural adjective of the first line predisposes the third to a homophonic (nongraphic) rhyme not just on "fair"/"heir" but on "fair"/"successive heir, " the latter possibility based on the suballiteration of f and v. Shakespearean rhymes also frequently work in reverse, anticipating across two words a spread of sound heard only in retrospect as pattern, here again at the added distance of an alternating rather than couplet rhyme: "Thou art as fair in knowledge as in hue, / . . . And therefore art enforced to seek anew " (82). In sonnet 3, the expressed desire that the young man should reproduce so that his face should "form another" is contrasted in rhyme to that selfishness which would "unbless some m other." In this case, the transegmental drift of the rhyme is subtractive rather than accretive, signifying alternatively in the process—by way of that most loaded of psychological puns—both the woman he should inseminate and the genetic other and double he would sire. The arguable presence elsewhere in Shakespeare of that first example of a rhyming increment, that fricative adjunct in "fair"/"successive hair, " is entertained in a recent article by Debra Fried that offers one of the first signs in Anglo-American criticism of a willingness to surrender (at its edges) the hegemony of the word as a stylistic unit.[4] This is the case even though in her essay, "Rhyme Puns," Fried nowhere discusses the implications for rhyme at large of that poststructuralist "overdetermination, indeterminacy and phonemic play" (99) to which she is so steadily alert. On the subject of old age in sonnet 68, Fried seems at first simply to agree with Booth in an instinct "to audit 'hair'" as well as "heir" in "signs of fair" (97), but it is her own gloss on his remark that further identifies the mechanism of "elision" (97) by which this audition is encouraged. To the catchall category of elision (rather than liaison) she also attributes her most striking later example of a rhyme pun, the "stroke of eight" that clangs against "wait" in Housman's poem about a hanging, A Shropshire Lad 11, only to knell further with overtones of the "stroke of fate" (97).
The guiding thought of this chapter is that the demonstrated presence of
even semantically un motivated and irrelevant rhymes will provide a stronger phonotextual basis for assessing and theorizing the effects of those lexicosyntactic "puns" and other combinational duplicities more decidedly foregrounded in the reading of literary writing. This chapter is thus designed as the laying of a groundwork—but precisely in order to demonstrate the shakiness of the textual foundation, the indeterminacy of graphemic shape when phonemically processed. From line to line, the unchecked match of rhyme in the paradigmatic axis can unlatch the syntagmatic, unfastening the locks of juncture from word to word. As it happens, traditional phonological structuralism appears to converge with deconstruction upon just such eccentric echo—that special case of rhyme which begins in departure from itself before projecting its difference elsewhere for iteration. The particular usefulness of these tandem rhymes in following out such lines of convergence between formal patterning and aleatory free play—following them straight into the previous impasse of stylistics, into questions of "voicing"—is that these (like all) rhymes speak, as it were, for themselves. They isolate the phonotext in operation. The cross-lexical slippages in Chapter 1 often had to be subsumed to meaning in order to emerge at all from the undulation of English phonemes. In the case of transegmental rhymes, however, meaning need not be invoked for the effects to be functional. Their randomness is thus dependent on no referential or thematic likelihoods except the undeniability of phonemic recurrence itself. What the ear hears in rhyme requires no surplus of meaning: this is the general rule. If transegmental effects can thus be readily admitted in association with already rhyming syllables, where no syntactic or semantic demands constrain the ear's acceptance of such phonic transformations, then their wider existence outside of rhyme, latent at least, must be allowed. The following chapters will elaborate on this corollary proposition after the present chapter lays the necessary groundwork. It must do so by situating its specialized demonstrations against the larger backdrop of received opinion about both the phonetic routine and the semantic machinations of rhymed verse.
Echonomics
Rhyme's reach, its breach, its treason: an accord syllabically heretical, (s)cryptic in the sense of being suppressed by writing. Such an echo is experienced only by voicing, either on waves of audibility or in the echo chamber of the mind's ear. In all his extensive scholarship on verse echo, including "polysyllabic rhyme," John Hollander tends to disregard, outside of comic verse, what we might term polyphonemic rhyme, operating as it does without syllabic check.[5] The closest his own evidence draws him to such a phenomenon is with the off-rhyming couplet from Geoffrey Hill taken up in an
important recent essay: "Patience hardens to a pittance, courage / unflinchingly declines into sour rage."[6] Quite apart from echo, the interest of the first two lines for Hollander rests with the way in which they implicitly develop a linguistics of their own generation through those "fabulous etymologies" by which poetry often operates: what he closes his essay by calling "chronic synchrodiachronosis" (133). Just as the ablauting of "patience" into "pittance"—with "forbearance shrivelling into its own pit"—has a metaphonetic dimension, so with the "diachronic" implications of "courage" as it "declines" to "sour rage," as if the adjective-noun phrase "were a suffixed, /au/- grade form, as it were, reflected in some latter day satem language" (129). Nothing, however, is said by Hollander about another (at least as likely) transegmental anti-pun—or antiphone—of "sour rage": the slightly less assimilative but semantically operable "sour age."
Hollander's own rhymes in the self-exemplifying verses of Rhymes's Reason do incline to such transegmental effects. After reading dedicatory verses that rhyme "I fill a " with "ancilla, " we later encounter Hollander's explicit attitude toward polysyllabic rhyme in a reflexive pronouncement: "A serious effect is often killable / By rhyming with too much more than one syllable" (14)—as in "This dimeter / Would limit her, " or in such later rhymes as "So pure a "/"caesura " (18), "seem a "/"ottava rima " (17), "a dime "/"paradigm " (18), "upon it "/"sonnet " (20–21). In extreme form, "polysyllabic" rhyme tends toward what Hollander refers to in The Figure of Echo as rime équivoquée (from the French tradition of the rhétoriqueurs ).[7] Illustrated by Hollander with Théodore de Banville's "perfectly punning" lines, "Dans ces meules laques, rideaux et dais moroses, / Danse, aime, bleu laquais, ris d'oser des mots roses" (32), this is what Walter Redfern in Puns calls "holorhyme," the complete homophonic nexus—again illustrated from the French: "Gal, amant de la reine, à la tour Magne, à Nîmes / Galamment de l'arène alla, tour magnanime."[8] Redfern adds in further demonstration a triple rhyme from Hugo—"mémorable "/"même au râble "/"mais mort, hâble "—along with the risible polysyllabism of Byron's "pukes in"/"Euxine" (100). Short of the drastic equality of the equivocating echo or holorhyme, the drifting or sliding rhymes that play to ear over eye—or, rather, that play off eye against ear—may be said to span a gap in two directions, down and across . They are thus found bracketing the horizontal as well as the vertical axis, cutting loose from the syllabic paradigms of standard rhyme as they break with the lexical modules of the syntagmatic sequence.
Before Hollander's Rhyme's Reason, Wimsatt's influential essay had also consigned polysyllabic rhyme to the category of humorous extravagance, to the "double or triple rhymes of a Butler, a Swift, a Byron, or a Browning." His interest lay instead with rhymes more subtly "logical." Building on Wimsatt's
approach, Hugh Kenner, in "Pope's Reasonable Rhymes," distinguishes between "incongruous rhymes" and those "normal rhymes" that seem more indissolubly "mated."[9] Kenner argues that in early Pope the former are reserved for satiric effect, the latter for "the realm of law" (82), asserting those "banal congruities" (83) that carry the force of received truth. One might take as an exception the early satiric passage from The Rape of the Lock, which Donald Wesling, in The Chances of Rhyme, has singled out as a supreme example of "chiastic rhyme."[10] The disarray of the pandemonium described is nevertheless organized by close semantic correlates. "Where wigs with wigs, with sword-knots sword-knots strive," it is inevitable that "Beaux banish beaux, and coaches coaches drive" (1.101–2). The semantic distinction between purpose and sheer impetus implied by strive versus drive is parodically blurred within the rhyming format by the reciprocity of a sibilant elision—"s=(s)trive"—and a sibilant liaison—"s=drive." This interplay is then reechoed two lines later with the summarizing "Sylphs contrive it all," subordinating the crisscross purposes of Pope's preceding couplet to a nucleus of first cause in "contrive "—the "hypothetical morpheme" (Hollander's useful concept), or homonymic root, of drive and strive together.
Pope, master of the rhyming couplet, is the recurrent test case in criticism for interpretations of semantic and phonic accord. From the present vantage, when such transegmental slippage serves to inflect a rhyming pattern with some kind of referential bonus, questions of intentionality seemingly ruled out by the contingencies of English phonemic structure return intriguingly to the fore. Yet all one needs to allow is that the active lexical ambiguity and thematic force of such a tiered rhyme in the following example from Pope's "Essay on Criticism," for instance, is produced in the phonotext as we read, however the words that allow its lateral play may have come to be there on the page in the first place. The subject is Homer as source of inspiration: "Thence form your judgments, thence your maxims bring" (1. 126). From Homer take your lead: this is the imperative logic organizing both halves of the chiastic line, with the second half inverting the placement of verb and object after the adverb. The result is a plosive transegmental ambiguity across the drift of sibilance: "thence your maxims (s)p/bring." The suspicion that what is being described is a spring or fountainhead, a fluent source, a wellspring of wisdom, is soon confirmed by the fulfillment of this cross-lexical hint in the rhyming line: "And trace the Muses upward to their spring" (1. 127).[11] Furthermore, if this undoubted tier rhyme between "s=bring" and "spring" is allowed by double grammar to supplant effect with cause even in the reading of the first line (the sense of "flow [from]" replacing "draw [from]"), then it would only ratify—at the level of rhyming logic—that very aesthetic "maxim" later to be articulated by Pope: "Men must be taught as if you taught them not, / And things
unknown proposed as things forgot" (ll. 574–75). Forgetting is here the very sign, in reception, of what is partly suppressed in the written text. The rhyming word that merely repeats the implicit by surfacing it—that spells out the unwritten vocable cloaked by its preceding scriptive alternative—can therefore be read to reverse the combined letteral forgetting of both an entire hidden "matrix" (a cliché like "the well of inspiration") and in particular the verb itself, "springs," that connotes such a source phrase. When that forestalled verb actually emerges into the letters of the text for the next line's rhyming slot, it is, like wisdom itself in the formula of that later precept, inculcated as if merely recalled to mind. It appears, in other words, as something we full well knew to—but immediately forgot to—expect, something we might well have deduced from the phonetic as well as the didactic context. Upon its arrival, we take it as given, given already by the phonic uptake of the preceding terminal wording. Pope's transegmental rhyme thus instances here—yes, whether he knew it or not—the whole didactic logic of his rhyming style.
Recurrent Tendings
In retreat from a premature and compromising semantics of the transegmental, however, in which only those echoes would be heard that are put to work, let us return now to the formative stages of that rhyming tradition in English out of which Pope develops his more systematic and masterly effects. In the "General Prologue" to The Canterbury Tales, Chaucer, portraying the cumbersome headgear of the Wife of Bath, matches the mention of this burden with the described red of her stockings. He does so in such a way that the transegmental anticipation makes her costume seem all the more of a piece, the finery "upon hir heed " suited to the "hoosen" of "scarlet reed " (ll. 457–58). Of the Miller, along with the fact that "he hadde a thombe of gold, pardee, " we hear that "a blew hood wered he " (ll. 565–66). Moving into "modern" English, we find several kinds of transegmental rhymes in Spenser. There are those, like the already mentioned "bloudy hew "/"dew, " that seem like a phonetic analysis by decontraction, among which should also be included "brood "/"were wood " and "be (h)ealed "/"concealed " from The Faerie Queene (1.5.20, 29). Other transegmental rhymes can of course entice us with a semantic evocation. In The Shepheardes Calender, for instance, there is the hint of a "rhyme pun" in the way that "lyftes him up out of the loathsome m yre" (1. 92), though anticipating the rhyme word "admire," looks back to "so hie" at the completion of the preceding (and, we now see, incremental) couplet, which yields to the new rhyming pair by the addition of an /r/ to its recurrent vowel sound. The effect on the transegmental instability of the succeeding couplet is to suggest that "the loathsome" can operate as a generic noun for that out of
which one can only be raised "(h)igher": lifted, that is, "out of the loathsome (m)yre." Again we confront a semantic appropriation of syntactic form, managed by transegmental adhesion or assimilation even while motivated by rhyme: an anti-pun sprung by eccentric recurrence.
One inevitable result of this chapter's collection of evidence is to suggest how often, how inevitably, this will happen. Transegmental rhymes in Donne, for instance, can be as inconsequential as "doth us / thus " from "A Valediction: Of the Book," but we move closer to an active semantic drift in "A Fever." Beyond the title's homophonic overlay of a single adjective, afever, the speaker's "persever" rhymes with his insistence that he would rather "owner be" for "one hour" of his beloved "than all else ever " (ll. 26–28). What needs primary emphasis here is that, quite apart from any functional lexical severance and regrouping, the ligature of the s, permitted by the elision of the e, at once widens and tightens the rhyme. It does this regardless of whether we hear the phantom repetition of the word sever . Neither does the effect (given the transitional and not entirely regularized status of then and than in early seventeenth-century comparatives) depend upon whether one hears a functional double grammar in the expressed wish to own her an hour, and then (thereafter) to sever all else, all other ties. If the transegmental audition of rhyme seems to open the floodgates of such syntactic as well as phonemic rearrangements, to trigger major ambiguities by the least pull on a single word, this no more certifies than it denies the existence of the widened rhyme in the first place. It needs only to be heard to be believed, not recovered by meaning (any more than does rhyme in general). Precisely because the undisputed existence of rhyme in no way depends upon thematic correspondence, we are approaching the problematic operation of the transegmental drift within precisely this area of overdetermined (but not semantic) stress, the manifest (but not necessarily meaningful) sound play between lines.
To be sure, following Jakobson again, we can acknowledge that rhyme is merely a special case of the poetic function in this regard. Phonetic recurrences do not respect syllabic—even lexical—boundaries, since even the simple matter of alliteration is, in its own way, the cryptic anagrammatizing of a line. Returning to Donne with the stylistic work of David I. Masson in mind,[12] we find that certain not uncommon pararhymes, often associated with a more immediately perceived terminal rhyme, approximate the "circumsyllabic sequence" noted by Masson in a paronomastic shift like Pope's "Why feels my h eart its long forgotten h eat ?" Related to the "circumsyllabic" paronomasia—or accord riche intérieur (170)—explored by Masson in such a quasi-anagrammatic bracket rhyme as Donne's "give "/"prerogative " ("A Valediction: Of the Book," ll. 44–45), or the more complex example from his Holy Sonnet 6, "purgd of evil "/"devil " (ll. 13–14). In such tercet rhymes as we find
in the preceding century in Herbert's "The Sacrifice" or Crashaw's "Wishes," the strain on invention and variety leads, as one might expect, to an unusual ingenuity in the recruitment—or felicity in the accident—of proximate phonemes within the rhyming span. An obvious ligature in Herbert's text gains hold on the phonic orchestration of "on high "/"to die "/"companie " (ll. 233–35). In Crashaw's "Wishes," one notes such transegmental augmentations (followed by dilutions) as "lye "/"mortal Eye "/"Destiny " or the more obvious phonemic declension of "grows "/"Morning Rose "/"being owes " (ll. 34–36), with the homophonic bond between the first and second lines reduced to a circumphonemic bracket in the third. Again, this density of syllabic and phonic overlay can obviously spill over into semantic play. In Crashaw's "Wishes," rhyme enters the semantic field almost by way of a tautological circle:
Life, that dares send
A challenge to his end,
And when it comes say Welcome Friend .
(ll. 85–87; Crashaw's emphasis)
In the interplay of liaison and elision, one can hear—even see—the reduplication as either "dares send "/"his end " (allowing for the slight difference of the s of his as a voiced /z/) or "dares send "/"his end ." In the latter case, especially, the reflexive gauntlet thrown down by life to its own death is captured by the sense that our "daring to end" is the be-all and end-all of the circumlocutionary notion of "daring to send" an invitation to one's own death in the first place.
Rhyming Pares
No doubt the most widely recognizable example of such subtractive, reductive, or distillant rhyme, where phonemic matter progressively falls away to expose the homophonic core, is George Herbert's famous tercet rhyming in "Paradise." From the first stanza forward, the poem blatantly capitalizes, so to speak, on the sublexical increments of its rhyme. Announces the persona in gratitude to God, "I GROW" like those trees "in a ROW" that thus to "thee both fruit and order OW." The remaining triadic, eroding rhymes go as follows: "CHARM"/"HARM"/"ARM," "START"/"TART"/"ART," "SPARE"/"PARE"/"ARE," "FREND"/"REND"/"END." In Vision and Resonance, John Hollander terms this process "metonymic" echo, which would seem to imply the synecdochic detachment of part from whole. Herbert's poem makes "typically brilliant use of metonymic rhyming sequences to stand for the conceptual sequences they describe, the central device being one of seeming to extract each rhyme of a tercet from its preceding one, the corresponding trope in the poem being one of pruning" (130). Hollander stresses "phonetic
derivations" but admits as well the presence of "graphemic or scribal ones," since the reader "cannot audibly chop c from a ch cluster any more than he can prune w down into v by removing half of it" (130). What, on the other hand, the ear can do (though Hollander does not mention this) is actually to register the ampersand joining "hand & ART" in the central stanza: "I START"/"and TART"/"& ART." This would serve not only to vocalize it (silently) as and but also to soften the extracted "art" into a more nearly homophonic rhyme with the preceding line: d=tart/d=art . In resistance to this, one might well imagine that the conjunction, unlike its appearance in the preceding line (or in two later instances), is here rendered nonalphabetically, nonsyllabically, to avoid just that dental accretion which would spoil, by overcrowding, the clean descent of the rhyme.
Elsewhere, however, the de-escalation of the rhyme is not so strictly policed. A ligature may indeed, though rendering slippery a single rung in the contracting ladder, work to highlight its governing structure. In the move from "judgments S PARE" to "prime and PARE" in the next stanza, for instance, the elision institutes the pruning back of the verb, the paring away of its lexical excess, even before the compound verb "prune and pare" is dropped into the rhyming slot. And the elision operates in such a way that the "metonymical" link emerges almost as an etymological association—synchronic and diachronic at once—as if paring were the slowly disclosed first cause of anything spare . In the closing lines of the poem, another transegmental augmentation of the end-stopping "end" (as in Crashaw's "Wishes") also subserves rather than subverts the logic of the poem:
Such sharpness shows the sweetest FREND:
Such cuttings rather heal than REND:
And such beginnings touch their END.
Read phonotextually as well as visually—in other words, read rather than glanced at—the last line's opening "And" touches its own "END" by the diacritical variation of a single letter. Further, the closing evocalized rhyme "REND"/"theiR END" returns the last verse to the already mistaken idea of rending (actually a nurturing). Lexical slack becomes spiritual lapse: the transegmental overtone of the phonotext staging here, at the poem's metalexical climax, a miniature ritual of resurgent doubt exorcized by a prelapsarian faith in God's spiritual gardening. The idea of rending, that is, returns only as a phantom antiphone, denied at the lexical level by the visible or graphic logic of the whole poem, itself a teleological model of paradisal ordering and destination. Such is the return of a repressed religious doubt under the erasure of formal coherence. It is a spiritual hesitation put in its place by the last
fluctuant move of a grand, malleable, and pervasive design, a Creation in little.
Whether paring away to a phonic module or modulated by adjacent matter into a broader base of rhyme, such jointures are a recurrent, an irrepressible feature of rhyme—not only down through Pope but on into the next century, where the hegemony of rhyming verse begins to give way, regular echoes becoming marginalized as comic or satiric extravagance. Well before the Romantic decline of neoclassic rhyme, however, between Herbert's extreme experiment and the regularization of rhyming couplets in the eighteenth century, falls Milton's masterly variety in verse forms and acoustic schemes. In "Il Penseroso," for instance, the thematized acoustics of the rhyme "noise of folly / most melancholy" (ll. 61–62) seems to justify the phonemic self-consciousness of the elision of fricatives, purifying the rhyme to "of (f)olly "/"melancholy ." Even the blank verse lines of Paradise Lost do at one point happen to incur an exact lexical recurrence—though not a rhyme, exactly. The moment is also acoustically thematized in a way that releases a transegmental irony. Hollander's The Figure of Echo contains a fine extended gloss (43–44) on this passage in which Sin cried out "Death " as proper name in one line and two lines later, by inversion, "back resounded Death " (2.787–89; Milton's emphasis). Writes Hollander: "Hell's return of the word is the sound of revulsion from caves whose hollowed emptiness" becomes "a physical locus of echo," no longer pastoral and affirmative. The sonic contours of the scene capture all this and, I think, more. The resounding caves, filling their void with the name of all mortal voiding, do not register solely in the exact chiming of "Death "/"Death ." Taken as exposition, this is the sign of echo more than its work, but Milton in fact gives both within the immediate sound play of his phrasing—even if we don't hear the potential ricochet of "dead" in "resounded ." Given the dental assimilation in both halves of the rhyme ("out D eath" and "resounded D eath"), the second, answering phrase begins to stutter with its own inward echo. What we may well finally hear is the rumbling and ungrammatical resound-d-deth, a formal if garbled present tense (as in Shakespeare's "debateth with decay") for the eternal re-verb-eration of the abstract substantive Death, its deafening process in and across time.
Echoic Cleavings
As we move forward from Donne, Milton, Dryden, and Pope—each authorial surname merely the place-name here for the site of certain phonemic contingencies—into the frequently comic or satiric echoism of Romantic poetry, we come upon the exemplary syllabic ingenuity of Thomas Hood's light verse. There is the overt polysyllabic and cross-lexical spread of "dishonour "/"left on her " or "pitiful "/"city full " from "The Bridge of Sighs." In reverse, the
rhyme word in Hood can of course be anticipated by a collision of two predecessors: "morning c loud / so sweet and loud" ("False Poets and True," ll. 8–9). By means of ligature, the conflation of the voiced and unvoiced stops produces a covert emergence of the adjective from its denominated source. Oppositely, by elision, the abab rhyme in "Queen Mab" captures, partly through such sonic dispersion and recurrence, the ubiquitous magic upheaval of earth, sea, and sky: when "raging flames comes scorching round " and, two lines later, "serpents crawl along the ground " (ll. 30, 32). The first rhyming cluster could genuinely be heard—in its semantic, its syntactic, context—as a homophonic alternative for its actual scripted form; as verb phrase, that is, "come scorching ground" would fall into place in the larger grammatical frame. No less suggestive, even though not passing through any such moment of semantic availability, is a tiered rhyme from one of Hood's noncomic poems, "The Workhouse Clock: An Allegory," in which the persona wishes "that all the Good and Wise / Could see the Million of hollow eyes " (ll. 73–74). The contrast between the blinkered vision of complacent morality and the emptied stare of malnutrition and despair seems actually to hollow out the rounded first phoneme of "wise" on the way to the contrasting "w=eyes." In a graphic play between signifier and signified, these "eyes" thus appear visibly severed on the page from the w of the supposed "wisdom" that might appreciate and help appease their deprivation.
Byron's rhymes are often deployed in such a comic mode, whether in the explicit self-reference of Beppo, with its polysyllabic rhyming of "person" and "for rhyme, to hook my rambling verse on " (st. 52), or in Don Juan, with the proleptic rhyme of "all ears! " with "some leers " (9.78).[13] Wordsworth would not seem to be this sort of poet, and he isn't, nor even a poet particularly inventive in his manipulation of rhyme, and yet a similar anticipatory echo haunts—there is no other word for it—a late sonnet, "The Column Intended by Buonaparte for A Triumphal Edifice in Milan." As the attentive soul "hears combats whistling o'er the ensanguined heath," the couplet is rounded off only by an eye rhyme in "What groans! what shrieks! what quietness in death!"—a rhyming pair that nonetheless reveals the very sign of "death" fluttering visibly across the windswept "ensanguined (h)eath ."
According to Donald Wesling's account, in The Chances of Rhyme, of the rise of modernist poetics, Romantic writers are situated at the irreversible crossroads of the rhyming enterprise. Opposed (like many of the Romantics) to the notion that "rhyme is the whole being of poetry" (132), Wesling nevertheless studies it as a signal ingredient of poetic transformation over time, a limit case for euphonic recurrences and regularized schematic patterns. Wesling's is as much an essay on modernism as it is on rhyme, using the latter as a touchstone for the more dispersed and experimental forms of
"markedness" in the poetry of the last two centuries. If, on Wesling's showing, rhyme is most productively understood as a special case of literary "defacilitation," then it may well appear that transegmental echoes—where the text is roughened to the point of lexical disjunction, retarded to the point of double take (when the effect is noticed at all, that is)—is a special case of defamiliarized devising. Transegmental accords expose the whole manufactured fabric of rhyme by drawing upon the phonemic additions or subtractions always incident to it—but rendered more graphically extreme in such cases across the zeroing out of the scriptive space between words. By being less overt than ordinary rhymes, by being in fact mostly invisible to the unaided (the unauditing) eye under the normal constraints of syntagmatic or lexical expectation, the echoic recurrence based primarily on liaison or elision thus lays bare the entire machine of rhyme's artifice. Since Wesling's book deploys rhyming practice as an index to modernist ferment, its commentary does indeed naturally gravitate to Romanticism as a watershed moment for the "chances," both local accidents and historical fortunes, of rhyme. If in Romanticism rhyme gradually loses its definitive hold on the poetic imagination, the transegmental drift further loosens the syllabic autonomy itself upon which rhyme has usually, in principle, been centered. The tendency in transegmental rhyming, more than in its polysyllabic cousin or even in certain rare forms of rime équivoquée, is to denaturalize the lexeme itself, even the syllable—to disperse both into their phonemic minims, their linguistic constituents. This microstructure of echoism thus falls into line with the other eccentric and elaborately reticulated sonorities of Romantic practice.
Keats's Lamia , one of the handful of self-conscious Romantic experiments in rhyme, was written under the immediate influence of Dryden's rhyming couplets. Though most of the paired rhymes sound more neoclassic, more cleanly paced and clipped, the following couplet precipitates a curious transegmental ambiguity; it describes the supernatural dispensation by which Lamia's beauty is "unassailed," in her invisible doings, by "the love-glances of unlovely eyes / Of Satyrs, Fauns, and blear'd Silenus' sighs" (1.102–3). In the grammatical scattering of the second line, the intended sense is probably that Lamia is safe from the stares of such unlovely eyes as those of satyrs and fauns, as well as being spared the sighs of the drunken Silenus. Yet the misleading serial grammar of this line—along with the strong association of blurred or bleary vision (as well as of dull-witted drunkenness) in the adjective "blear'd"—encourages, at least subliminally, an elision of sibilants resulting in the extraction of a more exact echo: "eyes "/"Silenus' sighs ." This tautological echo, after all, would only operate to establish cause for the graphically stated effect, since it is the escape from all eyes that liberates Lamia from the insult of anyone's lecherous sighs. In a famous passage from Keats's earlier attempt
at rhyming couplets, Endymion, the rhapsodic and rather vaporous description of the "Cave of Quietude"—
O happy spirit-home! O wondrous soul
Pregnant with such a den to save the whole
In thine own depth.
(4.543–45)
—is redeemed by the telescoped sibilance ("wondrous s oul") that releases the transegmental rime riche (or its near miss) on oul and (wh)ole . Hearing this right off as a kind of redundant homophonic pun reveals from the start the sense of totalizing fecundity (or "whole"-ness) infused within the spiritual term "soul."
There is a more dramatic, and more typically Keatsian, threading of such effects in the "Ode to a Nightingale," turning on a rhyme that has been much debated in Keats criticism without any account of its transegmental dislocations.[14] The knelling of "forlorn" that works to "toll me back from thee to my sole self " serves in the process to defy the fancy as a "deceiving elf " (ll. 71–74). In the transitional context of "magic" and "faerylands" from the preceding stanza, one may (half) hear "soul's elf" for "sole self." Echoing this, the exact belling rhyme in "deceiving elf"—that castigated and excluded alternative—both endorses such a directly polarized anti-pun and at the same time rectifies it in a single stroke of denounced imaginative escapism. Against the sensuous drifts of the ear, the inscribed rhyme may be taken to reconsolidate a genuinely sole or solitary "self" as antonym of "elf." The speaker's asserted subjectivity is construed as an integral being rescued from the invasive—here homophonically obtruded—threat of a preemptive double, phantasmatic and deceptive. Such an audition of the phrase explores in effect the underside of just the sort of critique leveled by Donald Reiman at "repetition of sounds" in even the "greatest poems" of Keats and Shelley. In a chapter on Shelley's style, Reiman alludes to Keats's alliteration in "sole self," quite apart from any hovering pun on "soul," as "a bit heavy-handed."[15] Such heaviness is, on my hearing too, exactly the first impression created by the phrasing—and not just by its written sibilants but also by the heft of another s in the gap of homophonic transformation between the words. This is an extra burden that works to fracture the collocation under its own phonic weight—and by shearing force to generate a momentarily ambiguous slippage.
When Wesling moves forward from the Romantics, the lack of attention paid by his overview of rhyme's "chances" to such elusive phonemic displacements prompts the unqualified historicizing of a sentence like this: "The offkey consonant rhymes of Dickinson, Owen, Auden, and the later Yeats are clearly deliberate dishevelments, attempts to cut the sound sweetness of most
high Victorian rhyme, except Browning's" (120). Browning, however, cannot so easily and singly be exempted from the post-Romantic nostalgia for euphonic chiming, partly because of the phonemic mutations his prosody shares with the other more stately or harmonious Victorians. In "Youth and Art," Browning rhymes—or his text does, arranging the echo without authorial veto—"rankles" not just with "ankles" but with "her ankles " (ll. 42, 44), a rime riche obstructed only by juncture. The same ear—ours, if not Browning's—for the embedding of one lexeme within the span of another has, two stanzas before, generated a rather more Keatsian word play. Reminiscent of the manipulation of lexical subclusters in "Not to the sensual ear, but, more endear'd" from the "Ode on a Grecian Urn" is Browning's "For spring bade the sparrows pair," as if—beneath the transegmental alliteration here—the natural imperative of such creatures were the homophonic verb form instinct in their name. So, too, Browning's colloquial dexterity with bold polysyllabic rhymes may be enhanced by a transegmental nudge, as in this from "A Grammarian's Funeral": "what it all meant "/"by installment " (ll. 106, 108).
Such tactics—or accidents—readily lend themselves to the contrapuntal harmonics of a high Romantic legacy, even while they afford a different sense, and a broader history, of "deliberate dishevelments" than Wesling intends when he lists Browning as a modernist forebear. There is a phonic disarray, a discord within lexical sequencing itself, that is (paradoxically enough) native to rhyming metrical accords. It is to be detected, as we have been seeing, from Chaucer through Pope to Byron and Keats—and beyond. We find it in simple form, semantically neutral and unobtrusive, forming no gratuitous new lexemes in process, in a rhyme like Arnold's "smoother reed "/"shall heed " from Thyrsis (ll. 78–79). The elision of the liquid r and the suppression of the aspirated h smooth the very strain of the rhyme to a purer homophonic core. In American verse of the period, Dickinson's irregular rhymes can readily tap into this phonemic instability. In "My life closed twice before its close" (J 1732), her text mounts, or at least permits, such an effect in portraying those mortal separations that, when they "twice befell, " are "all we need of hell, " that last portentous noun seeming to yawn from within the verb of simple occurrence, "befell," while making good on its inactive etymological underside of plummet. So can Tennyson, "sweetest" of the high Victorians, demonstrate this vein of phonic shifting, even within the regularizing constraints of the In Memoriam stanza. Tennyson's persona subscribes there to the belief of Hallam, that former poet of "divers tones, " that "men may rise on stepping stones " (ll. 1–3). Beyond the already "rich" echo beginning with the consonant t, the special aptness of this rhyming tie (or tiered rhyme) rests with the lingering, though entirely nonmorphemic, hint of "tone" in "stones"—apt, because the
very agency of human betterment in this elegy is allegorized as the surviving inspiration of the dead poet's "silent-speaking" voice.
Lines like this suggest why any periodization of rhyme should be pursued cautiously. Certainly it is too easy to limit the notion of "dishevelment," as Wesling tends to do, to a modernist aesthetic—or even to exclude the poets of high Victorian sonorities from textual disturbances incident to inevitabilities in the language itself. Tennyson rhyming may at times sound like T. S. Eliot avoiding it, or Shelley like Stevens. The Romantic energies of oblique rhyme are much in evidence as they inflect the complex scheme of the second stanza of Adonais, which unfolds from an interrogation (addressed to Urania) about the conditions of death: "where was lorn Urania / When Adonais died?" (ll. 12–13). The answer is conveyed by a set of rhymes that seem so deeply contaminated by the matrix verb "die" that they proceed almost involuntarily to reveal its third-person singular form, first by means of a transegmental adjustment of juncture and then by a syllabic disintegration of the lexeme. Urania sits "with veiled eyes, / 'Mid listening Echoes, in her Paradise " (ll. 13–14), and there, a line later, "Rekindled all the fading melodies, " with rhyme an aural form of such rekindling. Two centuries before, Donne's "Who though from heart, and eyes / They exact great subsidies " ("A Valediction: Of the Book," ll. 43–44) displays a similar ligature securing a widened spread of rhyme. It does so, however, with none of the semantic charge that Shelley's lines manage at once to bury and to detonate in those echoes from Adonais, when the past tense "died" seems to saturate the transegmental rhyme scheme with supernal traces of a still present dying.
Even more crucial to the thematic resolution of its text is the closing aberrant rhyme of "Mont Blanc." In the speaker's apostrophe to the absconded power source of the mountain, the title's potential pun on "Mon Blanc" (for the imaginatively impregnable blank of nature when received into the mind) is at last thematically—and transegmentally—foregrounded:
And what were thou, and earth, and stars, and sea,
If to the human mind's imaginings
Silence and solitude were vacancy?
(ll. 142–44)
Even before the rhyming word, the emptiness that threatens is felt in the lurking specter of "dearth" in the first juncture, "and earth. " Adducing the final echo to illustrate how the highly irregular and displaced rhymes throughout "Mont Blanc" subserve the thematic emphasis on that "sound no other sound can tame," those voices issuing from the heights and recesses of power, William Keach, in Shelley's Style , asks whether the trisyllable "vacancy" can
"belong in the rhyming sequence with 'thee' and 'sea.'" His answer: "It both does and does not: the '-cy' suffix rhymes with 'thee' and 'sea', but imperfectly because it is rhythmically unstressed and because it is attached to the root vacan(s) ."[16] The poem's last word thus "seems both to yield to and to resist the rhyming power of the compositional will" (200). The "imperfect" rhyme might on this account be said to introduce a dissonance into the whole consort of rhyme, a precarious reverberation from its own subjunctive fear ("What if?") of a spiritual nonaccord with the physical universe. But Shelley's closing lines are framed as a question, to which one possible response is that the mind does indeed rhyme with the universe. And so does "vacancy" rhyme in its own way as well. Keach characterizes it simply as "unstressed," unlike "and sea," a trisyllabic word that would get no terminal emphasis in the normal (nonversified) pronunciation of its last syllable. But more than the imposition of an iambic measure helps to secure Shelley's final accord, perhaps the single most arresting transegmental echo in the whole Romantic period. In the first of the two would-be rhyming lines, that is, there is an encroachment not just from phoneme to phoneme across an ambiguous syllabic break but from word to whole word. Emerging there—from within the very vacancy of spacing as transcended alone by reading —is, after all, the full, the overfull, rhyme of "stars and sea "/"vacancy ."
Though this effect may sound closer to a polysyllabic rhyme out of Byronic satire than to lyric high seriousness, it is deployed by Shelley as part of the unsettling urgency of the closing question. Neither the textural nor the metatextual resources of rhyme are exhausted in elevating the imperfect of one word to a metaphysical crux, to a problematizing of language in its power to order phenomena. Since the problem is itself bipartite, concerned with the interrelation of mind and nature, that component of the rhyme which establishes the ground of the latter ("earth, and stars, and sea") should be given equal weight with the noun of potential subjective decimation ("vacancy"). Phonemic undecidability has thus located the crisis, rhyme courting the danger the text's fluctuations can move to cure. It is in this sense that the topography of the rhyming phrase describing the natural landscape enters upon the inscribed metadrama of the closing rhyme. A disjunction between sight and sound privileges yet again the power of unseen harmony in the poem. In so doing, it anticipates the unorthodox rhyme word to follow ("vacancy") by exercising the lexical blank ("and-sea") for meaning. And it is precisely the graphonic force of this syllabic play, the intervallic transformations of its phrasing, that makes Shelley's largest metaphysical point for him. In the sweep of the mind over "stars, and sea," the very syntactic interval or break—requiring "conjunction"—rules out the facile fusion which would chime too closely with an imaginatively impoverished "vacancy." It is thus in
Shelley that we come to hear how Mon(t) Blanc remains perpetually available to be filled in by imagination's various engagements with the gap. Blanks represent hiatus, not void, a secret source in nature mated to a waiting hollow in the mind, a source imagined yet again as fecund, reverberant, eternally renewed.
Rays of Iteration
Hopkins's rhyming practice, as part of his larger system of grammatical torque and lexical strain, may be taken to mark one point of transition from nineteenth-century sonority, or its comic opposite, to a more irregular modernist texturing of the line. In a literary history conceived apart from influence, Hopkins leads through the early poetry of Joyce to the gyrating phonotext of Ulysses and Finnegans Wake . At the same time, in a way that recalls the starting point of these chapters in a Shakespearean "hearing with eyes," James Milroy's study of Hopkins introduces "gradience," its most comprehensive term for the poet's concatenated sound effects, at the conclusion of a chapter entitled, with a phrase borrowed from Hopkins, "Read with the Ear."[17] The term "gradience" is accompanied by an example directly pertinent to the transegmental effect, both in rhyming patterns and elsewhere. Its subsequent exposition, however, offers an unusually compact example of the artificial limits imposed, even in linguistically oriented criticism, upon hearing with the ear. The example, from stanza 26 of "The Wreck of the Deutschland, " is "The down-dugged / ground-hugged / grey," and Milroy uses it to illustrate his claim for "gradience" as "a kind of extended 'rhyming'" (148), the later term including internal as well as terminal echoes—assonance, consonance, and so forth. If a potential auditory recurrence actually threatens the lexical integrity of the graphic sequence, if it operates out of phase with the lettered sense, it is regularly suppressed. Hence the second hyphenated phrase, "ground-hugged," is said by Milroy to "abandon" (148) the d alliteration, when precisely what seems most Hopkins-like about the mounting sequence and its gradient escalation is the alliterative identity-within-difference, the incremental variance, provided by the ligature between d and the consequently less aspirated h in "groun-d=hugged."
Only in Hopkins would one tend to find such a phrase in a syntagmatic rather than paradigmatic relation to its alter ego, each collocation the phonic paragram—or internal slant rhyme—of the other as actively manifested in the text. In the case of "leeward / drew her / D ead," the dismembered rhyme with which this chapter began, Milroy is enticed only by the expectancies, the exigencies, of rhyme to find congruence around the acoustic edges of the scripted lexemes and lineations that produce this perverse (transegmental) enjambment. By contrast, supplemental rhyming matter at the other end of
terminal phrases—that is, occurring between lexically banded terminal syllables and the preceding phonemic (but sublexical) material of a given line—slips past Milroy without comment. Such effects thus miss their chance, in Milroy's hands, to illustrate a larger phonic principle of Hopkins's verse. In a mode of "gradience" reversing the manner of Milroy's previous example, there is a typifying internal rhyme in "The Starlight Night." From the two words "diamond d elves!" is thereby generated the possessive collocation "elves' eyes," where the internal echo is either precipitated by the former phrase or "delved" from within it. Either way, it is produced by an elision at the double d juncture—as if the "inscape" (Hopkins's famous term for the inner contours of both language and things) of "delves" is discovered as the very haunt of elves, playing phantasmatically within its verbal borders—Keats's "s/elf" in a new and pluralized phonemic habitat. The effect in linguistic terms remains a transegmental slippage between, however, rather than an inherence within: a swallowed d here, as above a forwarded d in "ground-hugged." Hopkins's end rhymes, as well as such internal echoes as we have just examined, also benefit from a similar phonemic motility, including the bidirectional drift from "The Escorial" where "Gothic grace " rhymes with "engemming rays ." The assimilative liaison of c with g releases "race" to a closer slant rhyme with "rays," while at the same time the preceding g from "engemming" slides both out from under its velar bond in ng (/h /) and across the lexical hiatus to expand ("enrich") the rhyming nucleus just distilled. It is again as if the radiance of "rays" were somehow part of the inscape of "grace" to begin with.
The metaphysical suggestion there is achieved by precisely that auditory physics which Joyce, in A Portrait of the Artist, will call the "rays of rhyme" (218)—and which he will exploit by means of transegmental ingenuities from his early poems through to the embedded rhymed verses in Finnegans Wake . The satiric passage in Ulysses in which Bloom compares Shakespeare's ingenuity to the routine rhymes of ordinary poets, who in effect gull us with such dull rhymes as "gull"/"dull" (Ulysses, 125), exemplifies a negative standard for Joyce's own less prepackaged rhymes. In "A Flower Given to My Daughter," Joyce's early text sets in echo to "Frail the white rose and frail are " the line "Whose soul is sere and paler ."[18] This tendency has mutated by the time of Finnegans Wake into such full-blown polyphonic rhyming in his italicized verse inserts as "philosopher "/"top of her " (47.1–2), "from on Hoath "/"upon Oath " (175.14–16), "story ends "/"orience " (418.29–30), or the more blatantly false rhyme "red her"/"feather" (383.12–13). At one point in the Wake, the semantic pressure of rhyme actually triggers a phonemic slippage just before the terminal syllable, when "At Island Bridge she met her tide " is answered a line later with "The Fin had a flux and his Ebba a ride " (103.1, 3), the latter inevitably eliding to "his ebb a ride."
In between Joyce's early poetry and the late interpolated verses of the Wake comes the famous villanelle in the Portrait, its preparation laid in the younger Stephen's earlier bursts of compositional free play and phonemic progression. From the first, Stephen interrogates his own half-conscious effusions. In a daydreaming manner he has concocted some doggerel lines: "The ivy whines upon the wall / And whines and twines upon the wall " (179). But immediately he asks himself, "Who ever heard of ivy whining on a wall?" The reader "hears" more than this, however, and hears it in a literary-historical context that expands our sense of the verbal dispensation within which Stephen's inchoate word-spinning is to be understood. Hopkins's poetry had not yet been published, so there is thus no question of direct influence. Along with Lewis Carroll, however, Hopkins is the great forerunner of the Joycean portmanteau term. As with the transegmental logic behind Hopkins's portmanteau wording from "Inversaid," the phrase "turns and twindles " arranged to reveal its appropriate overtone of "dwindling," so too with Stephen's "And whines and twines upon the wall ." Beyond the probable free association of "whine" with the "vine" of ivy, dental ligature parallels dental elision to generate across the combinatory logic of syntax, and even of lexical articulation, the approximate equivalence of "d=whines" with "d=twines."
It is the same ear for covert phonic accords beneath scripted patterns that culminates in Stephen's morning composition of his villanelle, including the tandem rhyme of "will of him " with "seraphim " (217; Joyce's italics). Stephen's self-consciously shaped exercise in rhyming schemes is characterized in its final rendition by what his accompanying stream of consciousness prose calls "liquid letters of speech, symbols of the element of mystery" that "flowed forth over his brain" (223). The contradictory signals of "letters" and "speech" track the graphonic play not only of the villanelle but of Stephen's whole verbal imagination—half incantation, half inscription—as it is captured here in its transit from mental conjuration to the act of transcription. It is in this way that his poem "sent forth its rays of rhyme: ways, days, blaze, praise, raise" (218). That last word is heard circling round in its own homophonic rhyme to the very metapoetic trope of "rays" initially devised to figure the echoic process itself. It is, moreover, this same radiation of a sounds (ai, a, ai ) that seems finally to lie behind the transegmental paring back (through elision) of that rhetorical question four times refrained: "Are you not weary of enchanteD Days? " The poem's self-examined array of rhyme is one mark of a poetic indulgence whose overcoming in the later style of Joyce is here virtually glimpsed—for a split second, a split monosyllable—in the technique of its own phonic deconstruction. Not for later Joyce the measured luxurious recurrences of such lapidary rhymes, with their enchanted and enchained string of incantatory a 's.
As early as the Portrait, Joyce steps back from the celebratory securities of this aesthetizing view, even while he employs the transegmental mechanism that foregrounds it in the elision of "enchanted (d)ays." His text manages, if momentarily, the play between graphic and phonic signification, between typography and sonority, that begins to expose and unravel the over-schematized rhymes of the villanelle. His junctural ingenuity serves at the same time, in a passage of more mysterious mellifluity yet, to offer a liquefied lettering of speech in a covert vocalization: "A soft liquid joy like the noise of many waters flowed over his memory and he felt in his heart the soft peace of silent spaces of fading tenuous sky above the waters, of oceanic silence, of swallows flying through the seadusk over the flowing waters" (225–26). Here are the rays of an interior slant rhyming that has naturalized the strained phonic gradience of Hopkins within the lyricism of a distended modernist syntax. It is in a prose-poetic passage like this, rather than in the anachronistic finesse of the villanelle, that Joyce begins to cut those stylistic grooves that will produce the graphonic cacophony of his later works.
"A Matter of Nice (S)pacing"
The great follower of Joyce in this phonemic hypertrophy of verbal texture is not Beckett, his heir in so many things, but Nabokov. There is the character (and novel) whose name is a telescoped phonetic transcription of the metalinguistic account of his very conception: punning (punnin') as Pnin . And the subtitle of a later novel blends transegmentally with the main title in a manner that summons in advance the mirror wor(l)ds of the plot: Ada; or, Ardor, as if the first two words together, name plus conjunction, spelled the verb form "Adore." In the text of Pale Fire —half poem, half novelistic gloss—a phrase like "limp blimp"[19] plays less obviously, but no less functionally, with a collapsed junctural irony. Beyond the internal echo of the phrase, its abutting rhyme, is the reciprocal (and transegmental) subtraction of immediately adjacent phonemes, which in their very articulation can be said to take some of the plosive air out of this "limp (b)limp" of a phrase.
The closeness of this collocation, for instance, to the quasi-onomatopoetic word chains in Hopkins—as might be continued, say, with simp, skimp, pimp, primp, pimple, dimple , and so on—begins to suggest a common denominator between the two writers (along with Joyce) in their sense of textual increments. James Milroy relates the manifold devices of Hopkins's internal accords and "complex phonetic play," including "alliteration, vowelling-off, skothending , assonance and internal rhyme," to "the old word-game whose object is to progress from one word to another . . . through a series of words formed by changing one letter of the previous words" (105). Sometimes called "metagram" (to borrow the term we will find Raymond Roussel adopting in
allusion to just such games) and related to classical rhetoric's "metaplasm" (the transformation of a word by addition or subtraction of a letter), this is exactly what Nabokov in Pale Fire calls "word golf," illustrating it by the challenge to transfigure "live" into "dead" in just five moves, passing through "lend" (262). This is not only a metalinguistic allegory for that easing over from life into death which guides the novel's whole plot trajectory, nor in addition simply a parable of the "lending" and borrowing of letters that indeed negotiates the transformation between these monosyllables, but also a deeper-going reflexive commentary on the textual conditioning of any thematically momentous transit: as always, in some sense, a mere play on words, a cumulative interchange of signifiers lent out within the economy of difference.
More explicit play with juncture—not just with such eroded borders as in "limp blimp," but with the actual segmentalizing and recombining of terms—might be called a metaplasm of the blank itself, a plasticizing of the lexical gap. By such means, for instance, the obscure word "grimpen" becomes in the course of a given line the textual tool itself, a "Grim Pen" (l. 368). And for a latter-day instance of Hopkins-like "gradience," we find on the first page of Pale Fire "the gradual and dual blue / As night unite s the viewer and the view" (ll. 17–18). Phonemic play is there doubly thematized as the graduation of visual tonality in one line and as unison through nondistinction in the next. Such is the inscape of coincidence (of lexical coincidings) in Nabokov's anything but fortuitous chiming. It is related to the prose play in the preceding novel, Lolita , on one word differentiated into two, a game that deconstructs the very psychoanalytic model of the erotic situation: "The rapist was Charlie Holmes. I was the therapist—a matter of nice spacing in the way of distinction" (2.1)—or for that matter, of nice distinction in the way of spacing.
In Pale Fire Shade and Kinbote, the poetic and editorial genius initially bifurcated for analysis, seem ultimately to converge—at least to the extent that each aestheticizes the verbal accident, the happy fault, the balletic slip. Both revel in a delectation of the verbal weft which is "based upon / A feeling of fantastically planned / Richly rhymed life" (ll. 968–70). The allusive twist here on rime riche notwithstanding, the actual rhymes of Shade's poem evoke something more subversive and unconstrained than "Echo's fey child" (l. 968). There is a weird fatalism as well as whimsy at work, which turns reading itself into a kind of paranoia. It is an apprehensiveness curiously matched by Kinbote's replotting of the poetic text as his own prolonged flight from death into textual immortality. To submit to the text as an occasion of paranoia is no longer to trust your senses. Words can't be relied on to stay in their scripted place. Their constituent phonemes may at any moment contract new allegiances, forge new words, or if not words, then unprocessed but palpable new sound configurations that subtend or overarch the lexical bound-
ary without stabilizing any alternative phrase. Single rhyme words in the poem of Pale Fire may extend themselves backward in this way, in tandem rhymes reminiscent of Hopkins—though with less mannerist contortion, more comic snap. Nabokov's approach to the mystical coincidence latent in rhyme or echolalia is primarily through a transfiguration of the mundane. His rhyming inscapes are for the most part vernacular rather than baroque. Both phonetically overdetermined and liberated at the same time from the predictable boxing off of the prescribed syllabic unit, such treasonous rhyme quickly puts the reader on nervous alert. A given rhyme may already be upon us, so to speak, before we realize it, as in the thematized paranoia of an erotic snare when Shade, confronted with an unknown woman's "freckled hands, that rapt / Orchideous air," immediately "knew that I was trapped " (ll. 771–72). The context is of course heightened by that portmanteau "orchideous," a parodistic coinage by which the effect of revulsion is recognized as inextricable from its hideous floral cause. Otherwise thematized, the complex echo can be timed to a scene of mirroring in which the slippage between words is reciprocal, impossible to pin down: "the mirrors smiled, / The lights were merciful, the shadows mild " (ll. 361–62). The rhyme gets the better of itself coming or going. Either an elision extracts "mild" from "smiled," or, if a liaison is instead felt to thicken the sibilant in the first line, then another liaison may work to extend the rhyme backward in the second. Such undecidable phonic drifts are only the more pertinent in this case for rendering ambivalent the exact dimensions of the echo, since smiling is simply a metaphor in the first place for the ameliorated mirror image in soft light.
Less perfect thematic matches may also result from interlexical phonemic slippage. These may destabilize further the reading effect, given that their phonetic overdetermination is neither confirmed by the scripted letter nor possible to shake off altogether. In the following pair of examples—"dance "/"its stance " (ll. 913–14); and "stress "/"woman's dress " (ll. 951–52)—the closeness of dental phonemes connives respectively to extend and contract the rhymes (and in both cases to intensify them—in the first case, by elision; in the second, by liaison) across adjacent but gaped sibilants. In another place, a dental sound may itself provide the shifting piece of the subliminal phonemic puzzle: "no doubt, "/"left out " (ll. 307–8). A minimal ambivalence of dental sounds can elsewhere abet a classical instance of what Wimsatt, Kenner, or Hollander might all agree to call rhyme's reason, despite its unorthodox procedures. To wit, the question about "what dawn, what death, what d oom / Awaited consciousness beyond the tomb?" (ll. 175–76). In another ambiguity of dental articulation, again associated with death, a drifting t functions in a lexically seditious mode in Shade's false "syllogism," the contention that (his italics) "other men die but I / Am not another; therefore I'll not die " (ll.
213–14). At such loaded moments, the very otherness of language—every signified sliding into the signifier of yet some other referent—betrays the self into exactly the excision it would protest against. The couplet following this serves to gloss such a metalinguistic condition more directly than might at first appear, unless one is reading already with an unnerved ear:
Space is a swarming in the eyes; and time,
A singing in the ears. In this hive I'm
Locked up.
(ll. 215–17)
With that threefold repetition of the I sound (or /ay/ diphthong) in the preceding syllogism, Shade might well be punningly implying that the space of identity is "a-swarming in the I's," just as time is "a-singing in the ears." Given the stress on first-person pronouns, there is also, with "In this hive I'm," the momentary phonemic expectation of "In this I've . . ."—what, come to believe? This to one side, though (as, indeed, it is in the horizontal axis of syntax), there follows that last contraction into the monosyllable "I'm," a phrasing dislodged by elision from its antecedent rhyme: the actual predication of identity thus drawn forth from the enforced conjunction of time ("and t IMe") with space. By the very trope of rhyming inscape, as it were, the self is herewith imaged as enhived by the joint envelope of space and time, shape and duration. The textual self is similarly localized by the specific linguistic processes of eyed shape and heard duration—in short, by the graphonic nature of textual manifestation.
This habit of phantom lateral play finds a phonemic free agent ready (as the text's own anagrammatic formulation phrases the movements of a ghost at one point) to "sidle and slide" (l. 554) at will. The penchant for transegmental edgings and dodges comes to a climax marked by rhyme at exactly the pivotal moment of Shade's strangely displaced elegiac poem, in which the poet himself dies momentarily in the course of mourning his daughter. In so doing, Shade confirms his metapoetic instinct that the cosmos is not chaos but organized, rather, along the lines of those verbal differentials and hidden echoes that make for life's "contrapuntal" poetry (l. 807). On the brink of extinction, his is an epiphany of "topsy-turvical coincidence" (l. 809), of sameness within difference, a reticulated "web of sense" (l. 810) that makes "ornaments / Of accidents and possibilities" (ll. 828–29). Precisely at this turn, this recognition of interlinked contingencies, Shade's poem dips into its only illicit polysyllabic rhyme. With "possibilities," that is to say, the poet rhymes "Sybil, it is / My firm conviction," setting up a testament of faith whose declaration is interrupted and unfulfilled. Its point is nevertheless obliquely made in this self-instancing happenstance of artificially cohering
rhyme. In the strained lexical disrespect of such farcical rhythm, Shade has fortuitously achieved what he has evoked a few lines before as "Some kind of link-and-bobolink, some kind / Of correlated pattern in the game" (ll. 812–13). His radically cross-lexical and low-comic rhyme on "possibilities" merely exaggerates into the open—into the multiple syntactic openings it bridges—those sly elidings that have intermittently characterized his rhyming habits and that are now erected into the logopoetic microcosm of a world to which, after almost leaving it, he finds himself restored.
It is left to his clinically paranoid commentator, Kinbote, to enter upon a closer analysis of cross-lexical effects, here those found in a variant manuscript passage of Shade's poem framed on a canonical Popeian rhyme of mortal imperatives. It is a traditional rhyme as seasoned and deep, if not as closely homophonic, as doom/tomb . "Do objects have a soul," he wonders, "Or perish must / Alike great temples and Tanagra dust?" (231). Though Kinbote does not mention it, there is yet another dental elision in "an (d T)anagra " that helps foreground all the more clearly the "anagra mmatic" consciousness at work in his subsequent decipherings. Reshuffling this very line, Kinbote targets instead the next lexical juncture to find the name of the assassin buried in the subsyllabic ash heap of "Tanagra dus t?" Mocking the "pedestrian reader"—one slavishly dedicated to metrical feet?—who may balk at this over-reading, Kinbote wonders rhetorically "how many such combinations are possible and plausible," offering two that are not: "'Leningrad us ed to be Petrograd?'" Or the trisyllabic: "A prig rad (obs. past tense of read) us? " (231). Neither work, the one because its phonetic form (the long u of used ) does not submit to metamorphosis of this sort, the other because the requisite a in rad appears in a word no longer in the langue and, hence, unsoundable in any parole . Like the Nabokovian text itself, whose surface tension Kinbote's dementia is established to travesty, his paranoia does set certain limits to its provocations.
Even the editor and anagrammatical exegete who finds Gradus lurking in the syllabic modules of "Tanagra dust" is bested by the bizarre alphabetical transcriptions of the poet's daughter, Hazel. She has spent time in a haunted barn engaging in coded conversations with a "roundlet" of light (188) that answers intermittently to her recitations of the alphabet, selecting out an apparently random sequence of letters. "The barn ghost," writes Kinbote, "seems to have expressed himself with the empasted difficulty of apoplexy"—some kind of lexical dysfunction?—"or of a half-awakening from a half-dream slashed by a sword of light" (189). In Nabokov's version of The Wake, this language of dreams, its condensations and displacements, is reduced to an alphabetic shifting and regrouping, a mass of "meaningless syllables which she managed at last to collect . . . in her dutiful notes as a short line of simple
letter groups." They are given by Kinbote as follows: "pada ata lane pad not ogo old wart alan ther tale feur far rant lant tal told" (188). We can find intact lexemes like "lane," "pad," "not," "old," "wart," "tale," "far," "rant," and "told," though in context they don't register as units of signification, as words; they are contaminated by the nonce syllables around them. In explicitly linguistic terminology, Kinbote insists: "Divisions based on such variable intervals cannot but be rather arbitrary" (189). He does, however, sense that "some of the balderdash may be recombined into other lexical units making no better sense (e.g., 'war,' 'talant,' 'her,' 'arrant,' etc.)." The urge to "seek a secret design in the abracadabra," with its echo of "seek" itself in the adjective "sec ret"—the admitted need, one might say, for the phonic to submit to the symbolic Law of the Father as "pa data" (rather than "pad ata")—is followed by the confession that "I abhor such games; they make my temples throb with abominable pain," an admission which involves not only the frontal rhyme of "ab hor" with "ab ominable" but an additional part-echo based on the intervalic redivision—the transegmental drift—of "throb with ab ominable."
Despite this "pain," and the apparently contagious sound play of Kinbote's own retrospect, "I have braved it," he writes, "and pored endlessly, with a commentator's infinite patience and disgust, over the crippled syllables in Hazel's report to find the least allusion to the poor girl's fate. Not one hint did I find," nothing at all "that might be construed, however remotely as containing a warning, or having some bearing on the circumstances of her soon-coming death" (189). She in fact dies exactly midway through the poetic text by drowning in a partly frozen lake, "a crackling, gulping swamp" (l. 500), whose very designation would require, if orally rendered, either elision or a gulping exaggeration of the repeated g sounds. Moreover, the lake in which Hazel dies is located (by way of an aural rebus) in the alphabetic void—not even a functional lexical gap—between the towns of "Exe" and "Wye" (l. 490) when pronounced as the phonetic names of adjacent alphabetic characters. The "variable interval" guiding Kinbote's researches into the ghost's code is thus in and of itself the clue to Hazel's fate, and precisely in its semantic contingency, its nonsignifying mobility. The "certain sounds and lights" (l. 346) that the young girl tried to parse into an articulate message, calling up perhaps the auditory and visual collaboration of ordinary language, thereby offers a clue to the textual allegory of her end. She has suffered a vanishing into and between script, a death by letters—merely because of their phonemic and graphic constituents, shifting and imponderable as they may be. The clue was there in the impenetrability of the alphabetic code itself. Suicide or not, her death as represented is made to depend on the sheer accident of phonemic sequence. Whatever stabs we may take at it—as, for instance, when thinking that we perhaps hear an admonitory "do not go" or "not to go" in "pad not ogo,"
or finding by elision the closural note of "tale all told" in "tale . . . lant tal told"—remain undecidable.
Even if we suspect that this fateful message is a premonition not of her own death but of her father's, by Gradus's bullet in Shade's own neighborhood lane—"Pa da ata lane pad not ogo "—we are still responding to a carefully staged allegory of textual excision. In something of this sense the unevenly numbered last line of the poem, "Trundling an empty barrow up the lane," passes away into absolute blank—unless, as the textual notes suggest for different reasons, we return again to begin rereading the 998 lines leading up to it. It is, to be sure, a Joycean circular echo like Molly's "Yes " looped back to "S tately" or, in the Wake , "the" attached back to "riverrun." But the circularity of Shade's poem defies all rules of proximate rhyming accord, so that "lane" carries us back to the poet's projective identification with the death of a bird against his window in the opening line:
I was the shadow of the waxwing slain
By the false azure in the windowpane;
The false "azure-in" is what gives false "assuran(ce)" to the bird so slain. Other such phantom mirror words hover over this passage as well, including the grand homophonic accident that allows us, in the antiphone "waxwingss lain," to hear pluralized the designation of the species (as well as the waxen "wings" for which it is synecdochically named). This same sibilant drift also releases by elision (and grammatical license) the past participal of "lie," as in the wings of a flying body having "lain" against the windowpane upon fatal impact with it—with the additional passing suggestion, in an anti-pun on the participial form of "lay," of the wings as s/laid out. In the retrospective rereading urged by Kinbote, this transegmental displacement of "waxwings (s )lain" would thus only free up a more perfect rhyme with the long-deferred closing "lane."
This is again the "division" of "variable intervals" which was the real secret message of the barn ghost's code, its cryptogrammatology of the scriptive break. And as with daughter, so with progenitor: an author's "issue" or product (represented partly by Nabokov's own brainchild in Shade himself) passes from supposed existence into textual sequencing, from living voice to the material processes of graphic and phonic succession. In lateral play here is the "interval" (Nabokov), or, as we will see, brisure (Derrida), or signifying gap (Lacan), that floats meaning over the void—the void between words—and listens there for the voice of the other, surfaced in permutation. It is in this Nabokovian sense that literature and the self it enshrines, whose doom it seals, whose immortality it assures, is merely so many marks aswarming before the
eyes, marks whose traces are at the same time singing upon the inner ear—a silent singing variably assigning its notes to meaning.
In something of a Nabokovian reflexive mode, poet-phonologist John Hollander, in the twelfth stanza of his recent poem, "Summer Day,"[20] offers an exemplary instance of what we might designate as a "phonemic trope" in rhyme—otherwise, his "figure of echo"—a braiding of end rhyme by which "Twilight and meaning, darkness and rising hope," having "Stretched out across my path a twisted rope," are followed, after a night of sleep, by a day in which "I trip over last evening's trope." The peculiar twist of such a "rope," beyond the proleptic etymology of troping as turning (as well as the etymologically false association of "twilight" with "twist" or "twill") is that, by liaison of the preceding participial /d/, the word's original appearance phonemically prefigures—as if by the logic of a midsummer night's dream—the dentalized alternative t = rope .
With dozens of examples gathered behind us, we can begin to recognize the full leverage offered by graphonic eccentricity upon the whole structural logic of rhyme. In this light, the linguistically as well as thematically "ironic" rhymes of Hopkins, Joyce, and Nabokov may be read all but explicitly to dismantle the artificial protective barrier in the previous English tradition, at least after Chaucer, between the low comedy of polysyllabic recurrence (Byron, Hood) and the lyric sanctity of tastefully restrained echo (the Shakespearean "line"). In turn, we can finally arrive at a definition of rhyme itself, in its play of discrepant matings and matched divergences, that actually includes, at its lower limit, the category of transegmental dislocation. The peculiarities of cross-lexical rhyming may in this way be found to reenact the general nature of terminal accord across the intervals within , rather than between, phrases. Implicit in this understanding, to begin with, is the nature of rhyme as an exemplary, even definitive, case of the poetic function. Taking the term "rhyme" in the broadest sense of recurrence and symmetry, phonic included, one could well characterize poetry, indeed literature as a whole, as "rhymed discourse." On this view, transegmental reading, insofar as it reads a certain (telescoped) logic of rhyme, might carry with it a rather pointed understanding of the literary condition itself as a reading effect.
All traditional rhyme, as normally conceived, puts difference at a safe and stabilizing distance, removes the recurring syllable or phonemic cluster at least as far away as the next line, and often farther. Rhyme thus plots either partial echo or phonic ambiguity, both aural factors, as graphic ones. It inscribes linguistic similarity and difference so that, up and down the page, gaps loom; the very loom of the text gapes; distances need to be crossed, convergences secured. For rhyme to work along this vertical axis, the ear
must overcome the compartmentalizations of the eye. So, too, with the horizontal axis in the event of transegmental rhymes, as well as the case of transegmental drifts per se as a liminal variety of "self-rhyme." That is, in graphonic slippage apart from end rhyme, lexical rather than linear spacing—prone to Nabokov's "variable intervals"—is defied and finessed by new hearings of a phrase, evocalizations that construe divergent vocables from the visible given of script. When the structure of terminal echoism draws on such transegmental effects, then, it does not just enhance or spread out the possibilities of phonic accord; it explicates the very mechanism of rhyme within a single reading span. When, for instance, an isolated word(ing) rhymes vertically with another, and then some , that something more, that overplus, that increment borrowed by lateral proximity, accrues to a rhyming difference within the lexical structure of either phrase—a difference from itself as written.
To hear how this works is to encounter the physical medium of rhyme: that median field which provides the space of transference between one line and some next (or later) one by which—not in which—the chiming effect is to be activated. For chiming is after all an effect, not a fact. Rhymes are not read; rather, rhyming is a way of reading. This second chapter has hereby progressed to the point where we are thrown back upon the question that preceded even the first: Where do we read? Rhyme makes the answer unusually clear. It is not found on the page but, rather, founded through the activity of the inner ear. Rhymes are less induced upon than produced within our "reading," where such reading is taken to be a process and not a received text. In a word, a double one: reading effects . Part of that process, in the processing of any kind of literary text, involves our passive auscultation, a listening in on our own latent articulation. Treasonous one may say, certainly transgressive, offbeat—if transegmental rhymes (by any other name: tie rhymes, tier rhymes, terraced or tandem rhymes) operate apart from or beyond visible repetition (as in mere "eye rhymes," say), then they prove the rule of the ear once again. But of the inner ear, the decoding and remembering ear—as the instrumentation of a sequence rather than the organ of a unified perception. What end rhymes prove first of all, then, is the differential linguistics of all textuality. No single inscribed word can rhyme in and of itself. It must appeal (rather than peal) across a distance defined by difference even when processed under the sign of similarity. That distance, however, though visible as space, is essentially blank as text: it is merely the space between lines. In the vertical axis of the textual plane, the line one might draw between two rhyming termini may best be thought to inscribe the base of a triangle whose apex lies above the text at the site of reading. This is a reading which therefore only exists through the
differential relation between these two points, separated cleanly by the space of lineation.
Less cleanly, because overlapped upon the same space, appears the contracted auto-echoism of the transegmental drift. Rhyme's constitution by difference—by the relation of one grapheme to its close phonic alternative—is intensively played out in the miniaturized compass of the transegmental effect, whether as part of a larger rhyming pattern or not. The relation of "rover" to "drover," "drive" to "strive," "end" to "send," "round" to "ground," "eyes" to "sighs," "rays" to "grace," "rapt" to "trapped"—to reconvene a few of the terminal densities scanned in this chapter—would amount to much the same relation as obtains in rhyme even if only the first of each pair were actually to appear (as graphism) anywhere in the text, the other left to an overtone contingent on some impinging phoneme. With enunciation thus encumbered by contiguity, such an adjacent, dislodged phoneme—catalytically activated at the level of the phonotext in the generation of a new though unwritten word—could therefore be said to precipitate the rhyming of a single lexeme with its own silent, layered evocalization. Alongside the traditional modes of end rhyme and internal rhyme, then, is the additional category of interstitial, or intervalic, rhyme: sprung to begin with from a word's resounding off itself in the slack—become play—between it and its neighbor. Such is the word in play with that difference from itself imposed by phonetic—become phonemic—context.
Mediating between two such hearings of a single phrase, negotiating their indeterminacy, triangulated by them as site, is again the reading body, medium of textuality as a phonic if not auditory phenomenon. Such triangulation resembles what Chapter 6 will borrow from Joyce to call "polysyllabax." In sum, reading alone voices rhyme, including that always imperfect but no less rich rhyme instituted between inscribed and evocalized signifiers in the same phrasing. The task of the next chapter is to work away at the conceptual objections to just this sense of text production, this "heretical" reintroduction of phonism (not, to be sure, phonocentrism) into the differential traces of textuality. Hearing reading voicing would otherwise have to take place over and above the distracting murmur of theoretical dissent.
3—
The Ear Heretical:
A Theoretical Forum on Phonemic Reading
It is time to stand back from examples. The theoretical plateau of this chapter follows from a study of loosened end rhymes as a limit case for phonemic drifting. Such lateral accretions produce an instability within programmed iteration that redefines the marginal units of recurrence. In the usual case of an aural, whether or not oral, reading of rhymed poetry, the reader, so to say, looks to the sound of rhyme. We know where to look; metrically speaking, one can count on rhyme. If one hears there, here and there, more than expected at the ends of lines, it is because rhyme has edged back into the precincts of a language whose combinatory progress is less predictable in its rhythm of equivalences. This is the textual space of a less codified phonemic succession, the jagged terrain occasionally inflected in its own way by the transegmental drift. We need now an account of that unstable terrain—and that continuous potential for inflection—within the broader parameters of its theoretical field. The remaining chapters will then turn to a selection of textual moments in poetry and prose that should thus be more fully understood to read the structure of a certain structure—namely, language.
But that structure has itself warranted, in approaches associated more or less indirectly with literary analysis, many quite different readings, ones whose points of opposition as well as of potential consensus (at least on the ground of phonemic versus graphemic form) need to be followed out rather closely. From descriptive linguistics to structural linguistics—or, in other words, from phonetics to phonemics; from structuralist linguistics on to deconstruction; next, in turn, from the deconstructed or decentered text to the split subject of psycholinguistics; from this sense of the "unconscious structured like a language" to the renovated "semiotics" of a subliminal matrix or genotext somehow actively in touch with the bodily rhythm of speech production (a looping back, as it were, to articulatory phonetics at a new level of
poststructuralist speculation); and on from this recorporealized linguistic substratum to the polymorphous pleasures of the read but no longer primarily interpreted text; this is the plan of a chapter that needs to ground itself at the start in the rigor of phonetic taxonomy.
The language of linguistics registers with keen, even colorful, precision the microturf of phonetic skirmishes flaring up as one speech sound, refusing to settle (in) for mere contiguity, struggles for dominance over another or induces its mutation. Concerning the boundary disputes and border conflicts between words, phonetic specificity tends to breed both terminological histrionics and clarifying analogy. Ballistic movements, morphemic onsets and offsets, open-vowel attacks, explosion, release, rests, constrictions, stops, residual breath streams, beat strokes and back strokes, false divisions, taut versus loose nexus, intrusions as well as occlusions, "diphthongizing diminuendo," even "vanish" used as a noun: these are just some of the terms that monitor the ceaseless interrelationship and relativity, the mutually implicated biplay, the virtual dramaturgy , of aural sequencing. In one extended survey of phonetics, for example, the factors most directly bearing on the literary phenomenon of the transegmental drift are enumerated by R.-M. S. Heffner under the heading "fusion," including such subcategories as "dynamic displacement," "doubling," "reduction," "omission," "glides," and "linking."[1] Yet the very rigor of distinction in phonetics is devoted to accounting for the functional "arrest of fusion" in a given communicative instance. Phonetic study is therefore less open to issues of cross-lexical vacillation than is the related consideration of morphemically pertinent phonemes . A case in point: in illustrating his claim about the disambiguating "initial element," Heffner writes in his subsection "Boundary Marks" that "no English form begins with [
According to the narrowest precision of phonetics, there would nevertheless be operative differences even between "engemming rays" and "engemming grays" that would rule out all active undecidability of lexical juncture. Such boundary adjudication, so to speak, is what linguists label phonetic "analysis," the ear's ascertaining of morphemic structure, its pinning down of word breaks. According to the sheerly formal alternations of phonemic structure, however, the pertinent alternatives — by the very logic of potential pertinence — may be felt to impinge more energetically upon the given scripted phrase. As we process the phonotext in silent reading, that is, the paradigm is kept in an overcharged and unstable relation to the ongoing
syntagmatic sequence. Though ambiguity can most often be hammered into submission by phonetics, in phonemic reading, by contrast, the weaker alternative remains something of a felt option even when — and often because — such reading is not constrained by one or another probable oral inflection. In moving from the descriptive linguistics of phonetic sound to the structuralist basis of phonemic differentiation — in short, from description to classification — we therefore find that the formal play of signifying units is more potently mobilized within the allophonic leeway of the latter system, a system, unlike phonetics, whose discriminations do not exhaust the uncertainties upon which poetic writing thrives.
Apart from critical junctures of lexical ambiguity, phonetics is certainly quick to admit that its calibrated speech sounds often overcode any functional difference, as language, between the "stimulus-response value" of given linguistic units.[2] But at another level of language production, of verbal "stimulation," the received "value" of literary phonemes becomes a matter of more freedom and more interchange in the circuit (rather than mere sequence) of stimulus and response. Literary language induces its own aural (if silent) feedback, not so much to thicken itself into "literariness" as continually to "reanalyze" the medium called "language" on which it is staked. In phonemic reading as practiced here, then, what is read is language itself, its deepest transformative logic, not just its given textual manifestations. In the semiotic terms introduced and reviewed in the Prologue, such reading engages genotexture as well as phenotext — in particular, the mutable phonotext subtending all script. We read language held in never more than partial check by meaning. Or put the other way round, so as to anticipate the Derridean turn, we read literature as the continuous critique of meaning. Not as a methodology, exactly, but rather as a working experiment, we read it here — as is seldom done — quite literally word by word, often word for (another) word.
What follows first in this chapter is intended mainly to concentrate theoretical issues scattered across or merely tacit in the previous two chapters. Derrida on Saussure's "general linguistics," Starobinski and Lotringer on the special case of Saussure's anagrammatic speculations — these complementary matters will offer the structuralist touchstones, both normative and eccentric, by which to consider, among other things, the unexpectedly related hypotheses of diachronic linguistics about the historical effect of junctural undecidability. From this anchoring in the formal linguistics of these separate schools, we can then move forward to the linguistic implications of post-Freudian psychoanalysis, Kristevan semiotics, Barthes's "grain of the voice," and other diverse ventures in avant-garde writing and reading — each as they bear in turn upon the elusive but illustrative function of the transegmental drift, that exemplary limit case of phonemic reading. With all this in mind, we
will finally circle round to Derrida again to find in the graphonic interplay of just such a drifting aural signifier a further sense of that "supplement" so crucial to the deconstructive understanding of differential formations. Getting straight exactly what the transegmental phenomenon takes by way of clarification from this whole complex intersection of discrepant and often conflicting methodologies, as well as what it yields in connection with the most debated issues in the contemporary understanding of the textual condition, is the combined business of the present theoretical overview.
Hearing a Difference?
Oscillating ambiguously not only between words but between text and reception, the "dyslexia" of the transegmental effect consists in an aural activation that impedes mere scriptive processing. Derrida's frontal assault on the primacy of voice in language might well seem to render any reading-with-the-ear a theoretically groundless pastime. If one recognizes, however, why the phoneme need not suffer the same fate of banishment from textuality as does the voice, such an acknowledgment secures a considerable phonotextual foothold within both the practices and the axioms of deconstruction. From such common ground it should be apparent how deconstruction in no way necessarily prohibits the reconception of style, of voicing (not centered and authorial but, rather, textual, receptual) at the site, along the very track, of those irreducible differences upon which a science of grammatology is also mounted. So far from being at odds, deconstruction and postformalist stylistics are inherently compatible, both in theory and in application. This is because deconstruction, in retrieving literary language from phenomenology, gives it over entirely to the regimen of a textual analysis.
An essential fact about both primary and secondary articulation — that they must be understood as sign systems derived from sound pattern rather than as a transcription of actual speech — is what makes possible the Derridean intervention at the very starting point of Saussurean linguistics. According to Saussure, as summarized by Derrida, the utterance of a vocable does not produce signification through sounds per se, but through a "sound image," which appears in the authoritative English translation of Saussure's phrase "image acoustique" as "sound pattern."[3] Highlighted here is an aural configuration with no contradictory hint of the visual. Still, Saussure's emphasis leads not so much to a phonational grounding for language as to a phonotational one. It is for just such reasons that Derrida is drawn to Hjelmslev's "concept of form which permitted a distinction between formal difference and phonic difference, and this even within 'spoken' language" (57). It is that Derrida introduces the notion of the trace as the very formation of such differential form. Functional — that is, formal — opposition emerges in the paradigmatic
axis not by a phonemic difference in sound but by the unheard trace of the alternative. This trace bears the sheer form of difference, emptied equally of sound and of content, dematerialized. We do not, we cannot, hear the difference itself, yet with it alone, not with the sounds that may be thought to determine it, does meaning rest. Linguistics is thus concerned no longer with the pertinent opposition of phonic matter but with the semantic pertinence of opposition itself. This is difference not as between presences (sounds, for instance), but difference as the perpetual deferral within a network of signifying alternatives and alternations. One of Derrida's several summary formulations of this issue: "By definition, difference is never in itself a sensible plenitude. Therefore, its necessity contradicts the allegation of a naturally phonic essence of language" (53).
Acoustic production — material, substantial, full — only becomes language, speech or otherwise, with the introduction of breaks, fissures, cuts. In other words, absence is the catalyst of the system of traces that converts the activity of sound into the formal pattern of speech. "Spacing (notice that this word speaks the articulation of space and time, the becoming-space of time and the becoming-time of space) is always the unperceived, the nonpresent, and the nonconscious" (68). In phonetic terms, one can think of spacing, the exemplary nonphonic marker, as the switch point between the "offset" of a terminal phoneme and the "onset" of its successor in the next lexeme. Whether "inbetween" a blank and a letter or "between" one letter and its alternative, linguistic difference is famously spelled différance by Derrida, evoking not only the formal "differential" of the trace but also, since the coinage is itself unsayable as different, its perpetual "deferral" of presence or voice — its operational basis, rather, in an "arche-writing" (68). It should be obvious how central this notion of the marked break in speech is for the present phonotextual investigations: this "joint" that (in Derrida's remarks on "la brisure" [65]) is both fracture and hinge, rupture and juncture, split and pivot — what I want to call the "double bind" of juncture.
Of Grammatology has thus deconstructed the metaphysical premises of Saussurean linguistics by introducing the proto-scriptive gap in order to erase the misleading metaphysical gap between the sensible (phonic) and the intelligible (semantic). All becomes properly semiotic, a tracery of signs, a differential notation. But in the sense that such a deconstruction is taken to reconceive language in a graphic rather than oral image of itself, it must be admitted that grammatology leaves out a theory of textuality as a reading effect . To include such considerations of speech in textual reception would reinstate, would indeed reinscribe, the phonemic stratum. It would be returned not as a dimension of full presence, never more than the mere trace of speech in writing. Rather, phonemics would be reinscribed within textuality in
such a way as to account, for instance, for the power of Derrida's own entirely writerly coinage, différance . We might say that the antilogocentrism of this term renders it not a full neologism but rather a neographism , being already the homophonic double of an existing word which haunts it "under erasure."[4] Coined in part, as Derrida explains in Speech and Phenomena , to cover those two simultaneous differential chains that are mutually activated in language, the phonemic and the graphemic, the term différance enacts the formal interchange "between speech and writing and beyond the tranquil familiarity that binds us to one and to the other, reassuring us sometimes in the illusion that they are two separate things."[5] Contrived, therefore, to illustrate the fact that written characters — as integers of meaning — have no inalienable bond to sound, still the coinage derives its effect from our tracing an apparent (alphabetically marked) difference back to its emptied vocal center. Its whole force as pun, as reading effect, thus depends after all upon invoking (evocalizing) the phonemic determinants of language — if only in deferral and under erasure.
Let us grant that, once grammatologically rethought, literary language is always more "writing" than speaking. Nonetheless, at a text's receiving end, the "productive" (Riffaterre's sense) "presence" (the common, not the metaphysical, sense) of the reader, as well as the potential or "passive" (the linguistic sense) presence of an articulating voice, together emphasize the "reembodied" nature of textual inscription in the work of enunciation, whether latent or out loud, whether stopping short at the mere rules of pronunciation or bringing them to voice. A close reading of Derrida's différance — as a scriptive detour within phonic equivalence — thus leads us not so much to the confirmation of his theories as to the source of the original confusion they are meant to reduce. To understand textuality as text production , in other words, is to isolate the phonocentric temptation. Script has so often been mystified as transcribed voice, as encoded speech, in part because it seems to meet a speaking voice halfway. I don't mean halfway between author and reader, the text as silent interface. The voice of the author has nothing to do with, or in, the textual circuit. Reading, rather, provides a tacit halving of the distance between someone else's text, that someone long gone, and one's own voice. We can read back from text to voice in only one direction, not back (behind it) to origins but back out from the page into the locus of the text's production as evocalized. That words on a page make us almost hear the sonic "patterns" conveying their signifieds as the productions of a latent voice, if only our own, offers one probable explanation for the lure of logocentrism — as well as the motive for deconstructing it. Yet, as we shall see later, the proliferation of Derridean critique in literary studies tends to undermine this illusion of voice without fully acknowledging what prompts it in textual (rather than meta-
physical) terms. In this sense, what the principle meaning of différance itself defers, often deters, from consideration is the role of the phoneme in that uncertainly fissured and fused stream of signification which is the read text.
This is a tendency rather than a tenet of the deconstructive agenda, however, since Derrida himself does make room, as noted earlier, for a certain oppositional play "between speech and writing" under the rubric of différance as well. This same interplay or tension will surface in his recent commentaries on Joyce. More than elsewhere in Derrida's work, the Joycean inference of "voice" (indeed, of multilingual pronunciations) within the very deferral of any such voicing by script will allow the return of a certain phonemic response to literary language — though certainly not a phonocentric one — into the field of verbal analysis. This is exactly the aspect of Derrida's practice as a reader that has been too little understood and pursued in the Anglo-American adaptation of his theoretical views — with a notable exception in the work of Geoffrey Hartman. One does need to begin any overall account of the "phonotext" roughly where (that is, in Of Grammatology ) Derrida begins his critique of "phonetic writing" (23), parting company with the Saussure of a remark like this: "The linguistic object is not defined by the combination of the written word and the spoken word: the spoken form alone constitutes the object" (31). The present study would take exception to, and so take off from, this same statement in Saussure — but not so as to follow Derrida to his full reversal in arguing for the exclusive priority of "arche-writing." To proceed toward a receptual theory of the phonotext requires a more limited contradiction of Saussure. Concerning the writing at issue here, often an extreme (but exemplary) case of literary textuality, I would claim that, whatever the "linguistic object," the textual object, produced as read, is exactly "defined by the combination of the written word and the spoken [that is, speakable] word," though unclouded by what we have already seen Derrida reject as that "tranquil familiarity" which can delude us into thinking that writing and speech remain readily distinguishable at base. Rather, their bond is a never stabilized compact, the shakiness of its terms being exposed when the usual maintenance by suppression it involves — by which all sense of the spoken is subordinated and contained by script — is suddenly abrogated in the act of reading. The scriptive warranty of lexical autonomy may then frequently be breached, words rent by jostling divergences, syntax itself unraveled in the slippage of difference.
Given the curious transegmental effects (including tandem rhymes) so far examined, this book might well seem determined on a course of very special pleading. Why the insistence on putting forward not a collaborative stylistics of diction, syntax, and phonetics but, instead, a highly particularized approach to the phonotext as a system not only of differences but of intervals —
intervals as the sine qua non of difference? In all this, the specialized evidence is meant to prompt larger considerations. The transegmental drift simply exacerbates to the point of unprecedented clarity a leading fact about all language, a fact that much literary phrasing is meant in other ways to foreground as well. Language lives, and breathes, on the skid. I don't mean primarily to be holding out for the dulce , even the utile , of the transegmental drift, an effect so dubious and ephemeral, in fact, that the words constituting it could be said either to take it or leave it. That the transegmental trace must be heard to be believed, however, is exactly what suggests a theoretical utility to these slidings and elidings between words. Such a purely differential effect can gain a corrective purchase on the very deconstruction that would evoke "difference" (however spelled) in the process of subordinating its phonic seductiveness. If the present field of exemplification seems unusually narrow, then, it is because I am trying to drive a wedge. I want to assert enough leverage on the deconstructive critique to free up its analytic power from the single-mindedness of its anti-Saussurean (and hence, ultimately, antiformalist) bias. The linguistic refinement displayed in Derrida's apprehension of textuality, deconstruction's unprecedented grasp of the play of literary language (play in the sense of alternation, oscillation, give)—indeed, its naming of this play as the founding condition of the trace—ought to be, though has so far not proved, widely enabling for an intensive reading of the literary text. In fact, one line of Derrida's thinking could lead, more or less directly, to what I have been characterizing as phonemic reading: a continuous response to those traces highlighted in literary language as a drift of functioning differentials.
Anagrammatology:
Words Sup on Words
In assessing the Derridean corrective of Saussure, one must include Saussure's notorious divergence from himself. Alongside the Course in General Linguistics ,[6] there is the Saussure who tentatively put forward, but never into print, his very special theory of linguistic relativity in the putative "anagrams" of ancient poetry. Jean Starobinski has edited Saussure's notes on the anagram in Words upon Words , and Sylvère Lotringer's extensive review of this volume follows out even more fully the irreconcilable divergence between this never-completed work and the Course , in light not only of deconstruction but of post-Freudian psycholinguistics.[7] In this commentary, Lotringer proposes to show that general linguistics, with its deference to normative signification, could emerge only if "armed" against the implications of the anagram—on guard, in other words, against "the irruption of the signifier on the scene of writing" (2). Lotringer is able to seize upon the implications for self and subjectivity in the discontinuities of Saussure's method for reasons having to do in part with Saussure's recurrent emphasis of the proper name, usually of a
deity or hero. This is the "hypogram" hiding behind the verse line or hovering above it—or, in Starobinski's formulation, lying under it (Les mots sous les mots ) or, yet again, in the words of his translator, mounted upon it. The poetic line thus unfolds under the aegis of that name, sponsored by a "presence" beyond the text. In Lotringer's ironically loaded, logocentric phrasing, the divining of the name is regularly its divinizing (5), the accessing of a plentitude both linguistic and metaphysical. Saussure's first full-scale example is "Taurasia Cisauna Samnio cepit," which he advances as "an anagrammatic line containing the entire name of Scipio" in the rearranged phonemic dyads "Ci sauna" + "cepi t" + "Samnio " (16). This is an anagram, if at all, in no exhaustive sense, for many of the line's letters, hence sounds, fall away in the face of the erratic logic of recombination. What results is Saussure's specialized notion of the "hypogram," an absent "theme word" that deliberately organizes the line from without , like an absconded prime mover.[8] Causation, intention: these were crucial for Saussure. As they increasingly slipped beyond proof, he despaired of shaping his studies into a definitive work, withholding them finally from publication. The fear was always that the verbal interplay, unmotivated, decentered, if you will, would lapse to rhetoric, to decoration, to mere "style." In this context of a heavily worked if cryptic agenda, the danger signal was always the specter of asemantic sound play , a superficial density characterized by the term "paronomase" (18) which Saussure retrieves from its contemporaneous obscurity only to relegate again to inconsequence. Up to a point, nothing sounds more like Saussure in his charts of phonic recurrence than the work of David Masson, cited in Chapter 1, on "free phonetic patterns" in Shakespeare—"free" from semantic obligation, "free" therefore from the snares of intentionalism. Saussure, however, tried to recover these snares as a safety net, underwriting with the logic of causality an otherwise willful gaming.
In linguistic terms, all Saussure insists upon in order for the hypogram to serve as an appreciable function of the verse, rather than just the reshuffling of single alphabetic characters in the narrower sense of an anagram, is that the shifting sublexical unit of measure be diphonic rather than monophonic, composed of at least two sounds in sequence. This insistence, however, is the very point of attack in poststructuralist commentary: the supposedly vulnerable underside of Saussure's instinct for textual permutation. In regard to this sticking point in any allegation of a phonocentric bias, it must be remembered that Saussure lays his stress upon paired or linked phonemes more on account of their credible perceptibility than their acoustic primacy; he emphasizes their reception as language , unfolded in time, an enunciable linear procession—however interrupted—rather than a sheerly graphic puzzle. It is for this reason, a lineated duration, that Saussure insists on the diphone: the smallest
registrable unit of sequence. On the other hand, the sense of progression clearly does not imply a continuous or unbroken flow of sound, merely the necessity of one "sound image" appearing (determiningly) before or after another.[9] Based on the diphone as the smallest increment that could effectually enter into covert bondings without surrendering itself wholly to the random, the result nevertheless remains a linguistic order irregular, ridged, interruptive. Since nothing in the principle of the diphone prevents the hypogram from overriding grammar and even morphology, one might take the anagram's aberrant rewording to mark the transgressed limit of that signifying continuum based on external open juncture, on lexical closure and punctuation, on decisive spacing and breaks. The law of the diphone could never, in other words, succeed in normalizing the anagrammatic moment within discourse; rather, the diphone anchors itself, just barely, against the exploding surface of that discourse.
For Lotringer, however, as for Derrida, the elevation of the diphone as first principle seals Saussure's pact with the textually conservative nature even of the anagrams. On this showing, the diphone, once insisted upon, becomes a policing agent: "The inter-position of the diphone is a summons to the linear order which will have its consecration in the Cours . The Anagrams weren't published: linguistics was born of that exclusion. We would suggest that Saussure's reasoning unreason proposed in fact the suppressed foundation of all that he elaborated subsequently" (8). But how conservative can the diphone be if it conserves only itself, relinquishing all else we take for granted about speech? How conservative, when it resembles so closely the polymorphous language of schizophrenia that Deleuze finds, for instance, in the displacement of the "difficult" consonants r and l of the English word early across a number of French expressions used in connection with time by his "schizophrenic student," including such anagram-like dispersions as "suR Le champ" and "dévoRer L'espace"?[10] Though stationed amid the directional stream of speech, the diphone is an overreaching agent. Its clusters mark, by stepping off, the spaces that need to be neutralized in order for the anagram to emerge. Once a free agent, in other words, the separate phonemic unit of the diphonic sequence enacts a textually disjunctive function, one whose possibility depends on the graphic apprehension of the very sequences it overleaps, the spaces or gaps whose ordinary structuring presence it only confirms by momentarily ignoring.
The general bearing of all this on the phenomenon of the transegmental drift should be growing clear. In such lexical slippages, a wording discrete and consecutive is tightened to the ambivalent segue of equivalence, the different fashioned on the trace of the same. Like the Saussurean diphone, the linguistic unit at issue in the transegmental effect—we might name it the
duophone , constituted as it is by ambiguous phonemic ligatures—also calls the syntactic sequence into question. Yet the transegmental drift, whether eliding a letter or sliding one forward (or back) along the lexical chain, does, through its mutation of the diphone itself as double integer, also fuse a new unit on either side of its formerly bonded phonemes. Words break apart, syntax breaks down, but still no single phoneme—unless it also comprises in itself a lexeme (as with the article a )—ever breaks away entirely. The "deconstruction," though powerful, is not exhaustive, since the results are held within the realm of a reading effect, a factor of process. They must remain in play within the very system they assault, eroding it without removing themselves from it, a structure of oppositions passing in and out of lexical pertinence.
This "duophonic" nature of certain unexpectedly active synchronic oppositions, however, finds a curious parallel, as it happens, in the field of historical or diachronic linguistics. The contingencies of articulation—in particular, its lateral instability—make it prone to shifts over time in the same manner that it is prey to flux in a given utterance. Diachronic linguistics is illustrative here through just this homology with textual play. Otto Jespersen's landmark study in linguistic evolution, Language: Its Nature, Development, and Origin (1922), privileges voice over writing in this regard as the more fluid agency of historical transformation. Jespersen pays special attention to the infantile stages in the development of symbolic language, "sketching the linguistic biology or biography of the speaking individual" as in certain cases it models the development of the entire lexicon over time.[11] For our purpose, his most interesting studies of children's verbal fumbles, the "lapses and blendings" of semiliterate speech, concern the segmental ambiguities he classifies under a subheading called "Word-division," the hyphen proving its own point about phonemically indeterminate juncture. "Children will often say napple for apple through a misdivision of an-apple " (133), in what amounts to a transegmental blending truncated by aphaeresis. Or there is the pure transegmental drift of "some-ice" mistaken for "some mice." French examples include un tarbre , starting from cet arbre , and "ce nos for 'cet os,' from un os " (133). The junctural decision making in these cases, erroneous or (over time) transformative—the effort at what linguistics elsewhere terms "disambiguation"—is designated in Jespersen as "analysis." Lifted to the historical stage of language change, as "meta-analysis," it is rewritten by a transegmental fusion of Jespersen's own: "I have ventured to coin the term 'metanalysis,' by which I mean that words or word-groups are by a new generation analyzed differently from the analysis of a former age" (173). The British place-name of "Riding," for instance, "as a name of one of the three districts of Yorkshire is due to a metanalysis of North Thriding," meaning "third part" (173). Jespersen in-
cludes a further list of such lexemes as napron, nauger , and numpire that eventually, through recurrent use preceded by articles, gave up their first consonants to become, as in "an (n )apron," the words we have in modern English. One of his examples, the transformation of a naddre into an adder , can in fact be checked against that poem by Chaucer which provided early evidence of tandem rhyme in Chapter 2. In the tale of Cleopatra from The Legend of Good Women , we find the archaic version of the noun nested in an alliterative sequence of internal rhymes: "An on the n adderes gonn e hire for to styn ge" (1. 699). And for a contemporary example of language change in the making, through the faulty analysis of consonantal "doubling," there is the solecism "advance degree" for "advanced degree," which I have actually seen spelled out in an advertising circular from Ford Motor Company (as if the Ph.D. were a way to getta head).
What is so distinctive in Jespersen's sense of that "metanalysis" by which we have come to lose such a lexical form as nadder , and with it, for instance, the chance for the particular consonantal alliteration of Chaucer's literary use of the word, is the role played by childhood—in other words, by marginal literacy—in linguistic change. Posied on the verge of fully mastered articulation, aware of grammatical bondings between articles and nouns, say, but not yet in control of vocabulary, the child, in Jespersen's view, is likely to find itself precariously adrift in the (transegmental) flux between diction and syntax. Within the diachronic "biology or biography" (8) Jespersen sets out to chart, we could say that ontogeny not only recapitulates but actually precipitates the phylogenetic transformations of language. In Julia Kristeva's terms (soon to be explored), the lingual vestiges of "semiotic" (or prearticulate) motility are glimpsed here in their still ambiguous and incomplete assimilation to the "symbolic" order, caught in vexed lexical junctures on the speaker's way toward full segmental mastery. The tricks of the ear thus take priority over the written "tongue," troubling its lexical reception, motivating its historical transformation.
As suggested in the Prologue, it should therefore be possible to conceive poetry, literary textuality, as the resynchronization of such a potential for diachronic change. Poetic writing, dependent as it regularly is on the phonic contours of language, invites a running junctural "analysis" kept open from word to(ward) word. Literature thus foregrounds without fossilizing those paradigmatic scrambles—sometimes historically determinate—that get disposed along the syntagmatic line. The tendency toward lapse and blend, toward phonemic lag and syllabic overlap, is always present, regardless of its force for change within the etymology of a given word. It resides in the differential synchrony—and syncopation—of a written text still haunted by the subarticulate undertones of a potentially more disruptive transformation, a
junctural shifting supposedly forestalled by its very institution as text. The reader's registration of a transegmental effect depends upon this continuous latency of mutation, rarely of full lexemes but rather of diphones (triphones, and so forth) abutting each other in syllabic collisions or sundered by juncture. To recognize this is a way of acknowledging that in certain cases the structuralist axis of syntagmatic succession seems to compact and replicate the axis of historical change over etymological time. Diachronic shifts thus have their textual counterpart in synchronic drifts. In this way, the defamiliarizing ambiguity of articulation in the phonotext performs its own "metanalysis" in process, rewording script by silent listening.
Despite the implications of diphonic—or duophonic—juncture for diachronic linguistics, it is still the structuralist process of differential signification to which the anagrammatic excursus of Saussure retains its most revealing relation, if only by default. The time has certainly come to say it: one can finally believe neither Saussure's evidence for the anagrams, by and large, nor the theory of their generation. What is alone persuasive is the emergent linguistic principle of "consecutivity" by which he hopes, and fails, to constrain both evidence and theory within the limits of credence, within a system rather than a random field of play. The limits of sequence placed upon the syntagmatic latitude of the anagrams thus make for a somewhat retrenched, but ultimately more progressive, account of reading practice as text production.[12] Jonathan Culler has written: "The interest Saussure's work on anagrams has provoked comes in part from its practical dramatization of the tension between finding and positing meaning."[13] For Lotringer, a more provocative debate would have to do with who is responsible for either the finding or the imposition. Which is a way of asking about the theory of the psychic apparatus behind such verbal play: whether in fact the Saussurean anagram did come closer than the Course could have permitted to a subject split wide by the discontinuities of the signifiers through which it assembles meaning. This is also obviously an inescapable question, psycholinguistic rather than phonemic, for any emphasis on the reading effect . Upon whom? Upon what kind of subject?
Puns Spun:
The Weave of the Unconscious
For all the kaleidoscopic discomfitures of its signifying energy, Saussure's anagram is still, from a poststructuralist vantage, found burdened with an unexamined complacency about the signified. In Lotringer's form of this objection, the "Saussurian anagram inhibits the circulation of desire by constituting a nominal reserve in which the subject remains caught" (9), a reserve of nomination, the sacred or at least mystified name. The specter of a constrained or inhibited subject, or reader, thus spurs a necessary deconstruc-
tion of the proper name. Such a deconstruction sees its fullest statement in Derrida's essay on the name of the French prose poet, Francis Ponge. The study appears under a portmanteau title that, in French and in English, happens to illustrate a transegmental principle of conflation itself undermining the integrity of the name. In Signéponge/Signsponge , the French for "signed Ponge" loosens by liaison the word éponge , while in English the elliptical syntax "(he) signs ponge" releases in fuller grammar, again by a kind of liaison, the aphoristic clue to the whole project: any sign will sponge off essence, will seem to absorb presence.[14] As regards the tensions between phonemes and graphemes in the work of alphabetic writing, it is typical of the distance between Derridean practice and theory that the localized punning implications of his title, its fortuitous disruption of juncture, carries even more telling suggestions for the free play of the signifier in inscription than does the book-length meditation on aleatory and irrepressible nomination that accompanies it. For unlike the case of the graphological coinage, différance , with its complete homonymic collision, the differences among signe/éponge, signé/ponge, signé/éponge , even "signé, eh! Ponge " (101) do not so much depend upon as precipitate the graphic differentials or blanks, let alone the orthography of doubled or inserted letters, by which the spread of punning is determined. In such paradoxical, such contra-dictional, graphonic biplay, it is the imaged "brisure" upon which all punning activity "hinges" as heard . There is a moment in Ponge's own manipulation of a silent phoneme, as glossed by Derrida, that gets to something of the same linguistic depth. The unavoidable imbrication of description with inscription is so often foregrounded in Ponge that his own wordplay follows suit in the (transegmental) conversion of "nature" to "gnature" (122)—a name for the textually engendered world that, even as name, bears peripherally with it, under never more than partial erasure, the dismembered morphophonemic trace of its materialization as sheer "sig=n=ature," sheer sign.
The onomastic basis of the Saussurean hypogram, by contrast, privileges the "nominal" to the point of mystification. It makes axiomatic the wholeness of the reading subject—according to Lotringer—before the reintegration of the maimed name of the Other, reified and often deified: the hidden name of hero or god that "solves" the anagram. The argument concerns a displacement of implied presence forward from, as well as out behind, the text. To piece together the name is tacitly to name our own relation to the logocentric basis of the text—our relation as subjects, as subject to it. We participate in the text as interpreters before the mystery, dedicated in advance to the reproduction of the phonic substance of its dismembered body; our subservience is that of acolytes at a sacrament, a linguistic transubstantiation. Otherwise, we as readers would confront in such texts the explosion of all names, all words, in a
leveling polysemy of the trace, a war of competing emphases, the fray (in every sense) of wording itself. In his one explicit allusion to the Grammatology , Lotringer registers how Saussure's claims for the anagrams, if pressed, would "in effect wrest the phonic substance away from that logocentric tradition (Derrida) which elevated voice to the ideality of meaning" (3). Derrida has himself pursued this toppled status, this dethronement of voice, not so much as a tumbling into the unconscious but rather as that fall, slip, lapse, that original and perpetual discontinuity, which, in Lotringer's terms, too, defines the unconscious: in Of Grammatology he writes, "Within the horizontality of spacing, which is in fact the precise dimension I have been speaking of so far, it is not even necessary to say that spacing cuts, drops, and causes to drop within consciousness: the unconscious is nothing without this cadence and before this caesura" (69).
Here of course is the dreamwork of Freud, its condensations and displacements, reread through linguistics. And at the heart of that linguistics, at the base of the Saussurean revolution (according to Lotringer and Derrida both), is a Janus-faced subject—conscious and unconscious at once, seemingly unified and ultimately split—whose contradictory positioning is brought to the fore in the reading of anagrams. In arranging to refer away textual motivation to a sacred or mystic nomination, to foist off on the name of a ritualized and transcendental signified all the deferments and deflections of what Lotringer calls the "phonetic gram" in its dance along the line, Saussure's attempted redemption of reference does no justice to the fractured nature of the text that produces this name. It is a text whose operations, aberrant and oblique, are in fact the staging of that name's coming (up) to consciousness from the words under words. The break into speech is thus a discontinuity at its point of origin, founded as it is on a lacunary slide. In resistance to this, the discoursing subject of the anagrams—and, by intersubjective extension, the reader as well—is shielded from the slippages of language by the Saussurean sense, as phrased by Lotringer, that "in the linguistic game, the subject of the enunciation does not have to be unconscious—language is unconscious for him" (5).
The psychoanalytic premise of Lotringer's critique is, more or less directly, Lacan's "agency of the letter in the unconscious." It leads us to the Lacan of verbal practice as well as theory. Christian Metz, in The Imaginary Signifier, compares the polysemy of language to the unbinding of psychic energies in Freud and then offers a punning passage from Lacan's Ecrits, never before translated into English, about the word as a nodalization of various meanings. Though not stressed by Metz, these are meanings that intersect the phonemic rather than semantic level of the word, unfurled there by segmental permutations. "'The word,' according to Lacan, 'is not a sign but a node of significa-
tion.'" As instance, Lacan gives rideau for "curtain," subsequently stretching it out "as pun" to "Les rides et les ris de l'eau [the wrinkles and smiles of the water], and my friend Leiris mastering [LeiRIS DOminant ] these glossolalic games better than I can."[15] Walter Redfern, in his intermittent diatribe in Puns against the self-regarding cryptograms of French avant-garde literature and theory, castigates none other than Leiris for the quasi-anagrammatic conversion, in Mots sans mémoire, of syllabe to sybille .[16] It is, however, the sibylline power of the syllable that is just the point of such a memoryless mot, bon or not, its power to unveil new meaning from its own folds, its own refolding. Two pages later Redfern has turned his attack against "Lacanism," singling out the notorious (and typifying) "homophonic" logic (in French only, he objects) by which "le nom du père" and "le non du père" are related, let alone "Les Non-Dupes Errent" (80).
To claim, however, that the mind works in puns—in slips, parapraxes, glissades, elisions, and the like—is not to say that this can be demonstrated through the stylistic equivalents of these fluxions and functions. Lacan's rhetoric in this regard can never be more than evocative, never be conclusive. At the same time, since the play and displacement of verbalization is the very scene of the psyche for Lacan, punning does not merely package an argument but, rather, enacts it. The Name of the Father, for example, institutes the Law of the Father, the interdict, the psychically enunciated "no" to desire; hence le nom slides over into le non . In this paronomastic parable (it could never be more), the name is heard as a negation in the very instant of its designation. After Lacan and Derrida together, one could propose that all attempts of the nom to cover (for) an inevitable nonidentity (with) offer a deferring to naming that is never more than a deferring of the named. Taken this way, still a homophonic pun cannot prove the point, nor even argue it. It can, though, foreground the instabilities and splits attending upon language—and thus attending any attempts to prove in or with language the existence of a subject otherwise constituted.
This is the usual force behind the effrontery of Lacan's vaunts, including such famous punning manifestos as l'être seen comprised entirely of lettres or the portmanteau coinage lalangue, which is not another name for la langue but instead for something else, something other and prior, to which, in a nonname, lalangue gives voice. It is no longer the langue, even the language in which it is written, no longer one among many langues (French, say), but a shaping of sound previous to the fall into Babelized difference, a lingualism unspecified by syntax or lexicon. In lingual execution as well as in reference, a pun like lalangue bears closest comparison to the semiotic, rather than symbolic, stratum of language in Julia Kristeva's unorthodox distinction, a phonatory impulse before logical articulation. Indeed, the systolic nature of
the coinage—that vibratory, palatalized effect produced by the syntagmatic telescoping of article and noun—catches perfectly the sense of a sound poised precariously on the threshold of referential or symbolic language. Before exploring the Kristevan parallel in more detail, we might simply note that lalangue, as almost a word, seems both to share with such punning as nom/n a phonemic basis and with l'être versus lettre to tap into a transegmental process that performs the deconstructive logic of signifying slippages. This is the metonymic skid that never allows signifier to catch up with a fixed signified.
But what does it really mean to predicate meaning in this way? Text production: the inscription, the con scription, of readers into the tread and skid, the separations and parings, the irregular restless s/pacing of langue . This is langue in its perpetual and never-completed realization as parole , as systemic sons made over into specific sens . The matter admits, however, of a more rigorous linguistic description than is usually offered. In mishearing—which is to say, in reading—éponge in signé ponge , for instance, or lettre in l'être , writing falls, stumbles, back upon its own double articulation. Primary articulation is voluntary, "stylistic"; secondary articulation subsists precoded, built into the word shapes of the langue . Upsetting this clear distinction, however, transegmental drifts transform, in the process of merely choosing, the lexical units they deploy; spacing becomes itself a module in the secondary stage of articulation, not just a partition after the fact at the primary level. A two-staged account of reception as well as of enunciation must therefore be developed, a double system not only of message production but of text production. If reading is construing, then we might call this reading function, in its reactivation of the twofold articulation of language, a double construction . It operates by the graphonic tension which in effect plays secondary off against primary articulation long after the former would ordinarily be silenced by lexical selection: call it the return of the repressed, of unconscious and involuntary phonemes, within the awakened but not orally engaged "voice" of the reader.
Phonocentrism is superseded entirely here by a decentered articulation. The uttering subject is gone from the authorization of the text, zeroed out at the point of origin. At the point of destination, then—or, rather, not then but instead (there is no prior to the advent of reading-as-producing)—at the point of destination, a decorporealized phonemic base is rematerialized by the still latent body of the reading—but the reading what ? the reading self? the reading subject? the reading mechanism? This reading function processes inscription not as enunciation but, rather, as language in action. Writing is revealed as sheer writing, poetry or prose as a text , but a text formally constituted at its base, rather than at any tenable source, by differences between phonemes and graphemes as they bear on the double articulation of the system they jointly
establish. Whereas Derrida demystifies Saussurean linguistics, even its glossematic revision, by deconstructing the "metaphysical" distinction between phoneme and moneme (or, in other words, between the ontology of the sensible versus the intelligible), I would want to locate elsewhere the brush of meaning with sense , with corporeality. At the locus of text production as a site of reading, texturing there the double construction of the semiotic fabrication, the dualism of mind/body returns already stripped of its metaphysical load. It returns not as a dualism at all but as a functioning difference within the (already contradictory, idiomatically deconstructed) "mind's ear" of the reader. It returns there as the very operation of the phonotext.
Metagrammatology
If the theory advancing a psycholinguistic foundation for the text produced in this way carries conviction, then the condition of such textuality must take precedence over any of the vicissitudes of literary history, including the championship of modernism in poststructuralist theory. Like early Foucault, like Barthes and Kristeva, Derrida in Of Grammatology is an inveterate advocate of modernist writing, even though he would seem to join Hjelmslev in departing from the retrograde devotion to the phonic stratum in its first great theoretical movement. Such is that Russian formalism which, in its "attention to the being-literary of literature, perhaps favored the phonological instance and the literary models that it dominates. Notably poetry" (59). On the other end of the spectrum from this implicitly logocentric formalism and the traditional texts to which it is in service lies the radicalized and irreducible textuality of a modernist writing practice: what Derrida follows Hjelmslev in calling, by a strained nesting of his own grammar, "the purely graphic stratum within the structure of the literary text within the history of the becoming-literary of literality, notably in its 'modernity'" (59). Derrida admits that this attempt at a "recaptured parity of substances of expression" taps a "vulgar" rather than purely differential "concept of writing" (55), an intuition "popular" (59) rather than scientific or formal, and that it holds no real interest for his own level of analysis. He retains the concept of the scriptive in "arche-writing," he admits, "only because it essentially communicates with the vulgar concept of writing" (56). This communication has in fact been too effective at times, for in the widespread extrapolations from Derrida's work, the graphic insistence, the relentless stress on the inscribed (too often the "vulgar" sense as well) trace, has often held sway among his followers, closing their ears to the silent but effectual forms of acoustic difference and deferral in literary writing. Against this tendency, the transegmental drift offers itself as a more revealing syntagmatic instance of that "'dialectic' of protention and retention" (67), that liminal waver of anticipation and holding
over, which founds the trace as more than a tracery of letters and which can be silently heard—beyond the grounding of any Voice—to keep language awake on the productive edge of wording: a continual syntactic emergency sprung by that many a soundless slip between (always somewhere between) page and lip.
Derrida's preferential sense of the modernist telos of literary history can best be appreciated by the example of one of those defiantly experimental "writers," no longer an author in any traditional sense, who left an inaugural imprint on French modernism. We turn, then, to a writing agent all but fanatically nonmetaphysical—and irreferential—in his practice but one whose activity nonetheless qualifies more than confirms Hjelmslev's stress on a "graphic stratum" favored in the modernist program. The fiercely self-conscious "process" of Raymond Roussel's early texts begins in the structural logic of phonemic equivalences. Instead of settling for a homophonic unity of effect, however, where one sound cluster would be heard in another's grammes, the Rousselian experiment spreads out the graphically masked phonemic replicants along the length of an entire text: a case of writing as sheer process. In the quintessential form of these early texts, a cryptic statement initiates a narrative line, and the only logic by which the text then moves forward is dictated by the—desire is the wrong word, except as it is internal to language—by the impulse, let us say—to end exactly but unrecognizably where it began. Subtending the sequence of lexemes as a syntagmatic fluctuation, where the spaces between words can be recruited, bridged, syncopated, redistributed at will, this superimposition of "process" upon sequence—of perceived equivalence within the deceptive autonomy of orthographic difference—serves to isolate the very production of textuality as against its manifestation as script.
An entire "plot" is often in this way erected on a transegmental model—where, in a sense, it takes a whole text to bridge the gap between words. To give a concocted English-language example for economy's sake, imagine a short critical text that begins "Joyce, he annoys!" and then works itself round to a celebration of "Joycean noise." Or, indeed, a French sequence beginning with "Vraiment rousse elle!" ("Truly a readhead, she!") wheeling round to "rêve Raymond Roussel" ("dreams R. Roussel"). The idea is not to generate a set of discrete puns but to pun on the very failure of discreteness in language, to capitalize on the verbal penumbra that hedges round every supposed integral vocalization, a phantom trace at once evoked and revoked. The split nucleus for such word play may occasionally be formed by a phonemic variance within a single vocable. This is all it takes in Impressions d'Afrique (1910) for Roussel to construct an entire narrative around verbal clues inscribed on a billiard (billard ) table as they set in motion the story of a
notorious plunderer (pillard ). Such random free play strikes to the heart of the arbitrariness of primary, not just secondary, articulation—as does Paul de Man's deconstruction of a similar billow/pillow rhyme in Shelley.[17] Forging a textual chime, the phonemic differentials testify at the same time to the referential dissociation between signifier and signified. The phoneme as trace becomes in this way a synecdoche for text production at large.
For his first and last extended venture in literary criticism, Michel Foucault, in his recently translated book on Roussel, studies the "process text" as underscoring "the sovereign role of chance in the interstices of language," an "effect" quite removed from "the order of style."[18] Foucault sees the "process" as a kind of obsessively proliferated rhyme, but his commentary takes no notice of individual couplets in Roussel's verse texts, where the premises of homophony are repeatedly tested. It is no surprise that such lexically dismembering rhymes in La doublure as "d'une voiture"/"ouverture," "saccades"/"des arcades ," and especially "second doigt "/"endroit"[19] come very near to the tandem rhymes discussed in Chapter 2. Foucault is interested in the more elongated as well as displaced chiming patterns which make the "process" texts seem bracketed by a concatenation of "antiwords" (33), each the "negative copy" (30) of the other, no original given privilege. In this respect, the "process" bears resemblance to the materialized visual pun on scriptive punning itself in Marcel Duchamp's anagrammatically titled film Anemic Cinema (1927),[20] where a spiral of script on a rotating disk—including such self-referential verbal transforms as the chiastic "esquimaux aux mots exquise"—literally put the spin on words. This later surrealist play with the hallucinated junctures of an alter language is indeed very close to Roussel's experiments in the "process" text. Duchamp's pseudonym, with which he "signed" Anemic Cinema, was Rrose Sélavy (for a swallowed "Eros, c'est la vie"), and he closes a brief pamphlet by that title with a Rousselian deconstruction of "literature" into "Lits et ratures."[21] Writing about the wordplay of Robert Desnos on the word éphémère, the surrealist Aragon also sounds remarkably close to Foucault on the molecularization of modernist phrasing. According to Aragon, Desnos has "pored long over these syllables which ring out like a legend. . . . He has descended philology's silk ladder in his search for the meaning concealed by this word at the heart of its fertile images."[22] Desnos is then quoted on the conversion of the trisyllabic éphémère into its alphabetic rebus, "F.M.R.," and then, in a more Rousselian mode, into its homophonic doubles, "Les faits m'errent" and "Des faix, Mères" (103). The English translator's version of Aragon's subsequent commentary is especially ingenious: "Ephemeral, F.M.R.L. (frenzy-madness-reverie-love), a fame really, ever, merrilly, Effie marry Lee; there are words which are mirrors, optical lakes toward which hands stretch out in vain" (103).
In the strict early form of the "process" narrative, Roussel also gives, like Desnos and unlike Duchamp, each form of the wordplay. Both halves of the vocalic nuclear fission are thus (di)splayed to view, so that recognition is itself demystified, trivialized. A phrase like "Eut reçu pour hochet la couronne de Rome" is never left to insinuate its own dismemberment and reassemblage; the text moves instead toward some quasi-narrative justification for eventually spelling out "Ursule, brochet, lac Huronne, drome" (43), a phrase dispelling its phonic double even as it replicates it in different scriptive shape. Roussel himself explicitly saw the "process" as related to the game of "metagram" (25), where words are gradually transformed by the replacement of single letters (in rhetoric, the metaplasm). In Derridean terms, his texts expose the local "sponging-up" of phonemic material from one lexeme to the next in a serial displacement along the syntactic chain. So much textual space has elapsed between these two avatars of the same phonotext, we might say, that the bizarre rime très riche has become impoverished and nonnegotiable, neutralized by distance. The text at such points reads like a primed but undetonated pun, a play so free there is no payoff.
Closer than Duchamp's puns to Roussel's "process," then, are those distended echoic "glosses" of Leiris which were no doubt being alluded to earlier in that homophonic homage by Lacan. In the 1939 text, "Glossaire, j'y serre mes gloses" ("Glossary, where I store my glosses"), collected in Mots sans Mémoire, Leiris's alphabetic list of matrix terms is submitted to vocalic and semantic expansions of the sort that translate "VOCABLES" itself into "vos câbles, pour échapper à vos caveaux" (113). Often there is no strict homophony, even at the start, but rather a dilation of the syllabic matter across a very free semantic paraphrase. The latter strategy is often organized around a phonic bracket, as when "JOUISSANCE" is virtually anagrammatized into "jusqu'où (en jouant) se hissent mes sens?" (94). Or the pattern may be controlled by metathesis or chiasmus, turning "LIMON," for instance, into "mon lit" (96) or, less precisely, "LAGUNE" into "une langue" (96).[23]
As a matter of fact, Roussel's "process" may best be understood as an experiment in commutation at the very crux of linguistics. Paradigmatic "equivalences" or pertinent oppositions, ordinarily resolved by the very choice of words—and then set in motion along the axis of grammatical alternatives—are instead kept afloat at syntagmatic peril. This binary impertinence virtually undoes the semantic regime of language, reshuffles it into a dreamlike virtuality. Anticipating the vocabulary in The Order of Things used to describe the modernist literary revolution, Foucault sees the "process" as one wherein an initial grammar is "pulverized" (43) and reconstituted in the "syllable-droplets of a polyphonic language" (60). Roussel himself puts it this
way: "I decided to take a sentence at random from which I drew images by distorting it, as though I were taking them from the drawings of a rebus" (41). The most systematically "processed" example of these "homonym sentences" (17–18) offered by Foucault is the nucleus (extracted from a folk song) "J'ai du bon tabac dans ma tabatière" transformed into "Jade, tube, onde, aubade en mat à basse tierce" (41). The phrase "I've got good tobacco in my tobacco pouch" thus spills over and out into the serialized non sequiturs "Jade, tube, water, mat, object, to third bass"—by no logic of plot, narrative, or signification, only by a phonemic filiation internal to language, strained and attenuated at that. But the stratagem is clear. The syllabic spin on one word sets up a phonic chain reaction that spirals along the nexus of syntax, relooping its linkages. The vowel sounds that end the first half-dozen words, despite the (silent) scriptive consonants that round off half of them, open an oral force field that seems to magnetize the largest aggregate of subsequent material capable of morphophonemic bonding into a new lexeme. By ligature, the attracted phonemic particles are either held over (as if geminated into renewal) for the next lexeme—"J'ai du b-" splitting its dental sound into d/t to generate "Jad(e) tub(e)"—or displaced backward: "-on tab-" become part of "onde aubade."
In the irregularities and risks of these regroupings, what dictates the pattern of recombination is the internal logic of langue as it spews up parole —but also (and this neither Roussel nor Foucault discusses) the power of reading itself to motivate the reconstitution of the chain. The reader can participate in the metagrammatic game only if his or her listening ear allows the inaugural dilation of one lexeme—reduced suddenly to mere phonemes—to linger for completion by the rest of the lexical material it now needs. Absorbing all it can at once, the reading ear must fend for itself from then on, submitting to the impoverishment of sense in its effort to catch up, to complete each phonemic move with a syllabic clustering capable of building supplemental words. The initial insolvency, that is, the expenditure of one word in augmenting its predecessor, forces each new sound into a phonemic parasitism, by which it leeches what it needs from the paradigmatic disarray in order to keep up the syntagmatic momentum. In all this, Foucault's interest in Roussel seems as exemplary as it is peculiar. Well before his exhaustive investigations into the discourses of culture, Foucault here submerges himself in a radical undermining of the culture of discourse. He studies in Roussel the human mind's very acculturation to language as signifying function. Roussel becomes the exception that only proves the rule, recirculating the ingredients of meaning without closing them off in an interpretable text: a mere procession of sounds slotted randomly into meaning, one by one, one upon the other, overlapping, slapdash, contiguous but maddeningly inconsequent. In narratological terms, the
collapse into the oversameness of metaphor, of likeness, as postponed by metonymical dilation, would here be a collapse into the same once more, only different.[24] The metagrammatic variants, like a reverse Rosetta stone, give not one message in two languages, with the resulting spur to decoding through cross-reference; rather, the "process" text gives two phonemic messages in one scriptive form, waiting for a second inscribed formation to confirm the doubleness by seizing upon its alternative spelling—as upon a long-lost twin. This reunion scene, however, has no textual site, is consolidated in no single verbal space. With its leveling of reading effect into textual function, of impact into sheer process, it is a pun split wide and gone flat.
From a psycholinguistic perspective, there is certainly no chance for a stable subject-position in Rousselian text production. No reader can anchor the lexical and semantic drifting in a way that could possibly naturalize its doubleness. Indeed, no imagined speaker could be thought to intend both versions of the metagram in a single utterance, as a single subject can, for instance, be thought to mean a pun. The enunciating subject of the pure "process" text is split from inception, decidedly ruling out the phonocentric myth of authorial presence, a voice behind the wholesale syntactic shakedown. In "ordinary" written language, the link, say, between lac Huronne and la couronne, distanced from the possible friction and perceived frisson of a pun, would be a negligible fortuity of "phonetic writing." Any junctural undecidability would ordinarily be "disambiguated" by context without even coming to consciousness. But when we enter upon Roussel's hypertrophied textuality, with its unique strain on, and analogy to, mental processes, we enter a different verbal regime. What is distinctly literary about the metagrammatic games of Roussel is precisely the metalinguistic torsion within discourse they produce, a vertiginous instability that refers every word away to the variable procedures of wording. Roussel's metagrammatic maneuvers, straining the very notion of wordplay to its limit, nevertheless explicate once and for all the force of transegmental punning within a scriptive system. We may call it, once again, the return of the phonemic repressed.
For Foucault in Death and the Labyrinth, Roussel's homophonic chains, as they double and travesty each other, are likened not to phantom voices but to "mirror" images, as were the ephemeral words of Desnos by Aragon. In this they stage the relation of death to life; they inscribe, that is, by being no more than script, the relation of simulacrum to vital reality, of verbal ghost to the living body of utterance. True to Foucault's double title, the labyrinth, become a hall of mirrors, becomes a tomb: what Derrida would see as the underlying "thanatopraxis" of all textuality exhumed and laid open to view in a quintessential modernist instance. The implacable face of every text becomes a death's head, a death said. At the origin of the "process," writes Foucault, is "a
repetition that is always anticipating itself; it acts as a mirror in relation to death[,] . . . life repeated to death" (162). If the mirror within the labyrinth is to serve fully as a figure for the depersonalizing (which is to say, killing) otherness in language, however, it can be only in a more or less Lacanian sense. The mirror is that figure of self in specular alienation from itself used to mark the transition from the imaginary to the symbolic, from a presumed plenary fusion of body and voice into the mere traces necessitated to express, and in the process surrender, just that union, that plenitude—to externalize and thereby deconstruct it. Script materializes the lacunae in speech that are at once created by and make for the activity of consciousness. Foucault comes close to such a perception about Roussel when he writes, in a way that calls up both Derridean and Lacanian presuppositions, that "it's the reserve from which words flow, this absolute distance of language from itself, which makes it speak. . . . It is the very edge of consciousness. It shows that at the moment of speaking the words are already there, while before speaking there was nothing" (39). The differentials by which language operates, if they include spacing as well as paradigmatic alternatives between phonemes, are therefore a matter of successive and conceptualized breaches sprung from an inaugural gap. They take part in a metanarrative on the very nature of language and the irrational, utterance always stretched thin across that abyss from whose silent mumbling emerges the pressure of alternatives, the ceaseless Voice of the Other. Deconstruction and psycholinguistics here meet again on the divisive ground of the rent and emptied center, as marked out—and, in Roussel's most illustrative feats, overleapt—by transegmental inscription.
Roussel's affront to ordinary poetics, to rhetoric, to style, his joint confounding of message and code within a strenuously scriptive (as well as, or because, obsessively homophonic) semiosis, invites some attempt to align his eccentric textual practice with the study of variation-by-"paragram" in Riffaterre's Semiotics of Poetry . Roussel's "process" is, as Riffaterre might say, certainly one of "conversion," but so relentless and unchecked that it undermines the principles of variation on which a normal, communicable semiosis of poetry is based. Roussel's texts are ferociously literary, letteral, without being in the least traditionally poetic. Unlike the paragrammatic conversion and displacement of the poetic text, Roussel's play with the metagram, having no rhetorical intent, avoids no cliché (in Riffaterre's sense) because it courts no original idiolect, aspires to no style. It displays the "languageness" of the literary in a way that forbids the poetic: a parole randomly sorted into langue and then metamorphosed beyond easy recognition. The metagrammatic text is systemically overdetermined but referentially arbitrary. Rather than generating a Riffaterrean paragram, it spawns ultimately a sentence that is more like an anagram (or hypogram) of itself —but of itself as different, alienated at
base, uprooted at the point of semantic implantation and narrative departure. Signifying force in Riffaterre's semiotics comes from a tension between what is literally said and what is repressed by that phrasing. Roussel's texts, by contrast, short-circuit such a signifying system by extruding everything as surface play. The variant appears in the text along with its seed phrasing, each a variation of the other but without a real matrix—except in the pure phonemic substratum—and no model to establish priority in the forms of conversion. Semic variants at the level of primary articulation are reduced to the mere function of secondary articulation, polysemous and irresponsible, ungoverned by any semiotic principle in the message at all, only the sequential code of the langue . Sequential—and irreversible.
This is the inarguable common ground of a Riffaterrean semiotics and a Rousselian metagrammar of literature. It is related to what Paul de Man remarked in his omnibus review of Riffaterre's work as the latter's conservative devotion to the "phenomenality" of the text as experienced, "its accessibility to intuition or cognition."[25] Pointed out here is what separates Riffaterre's "semiotics" from the extremity of Saussure in his anagrammatic phase. "What Riffaterre has done is to re-lexicalize Saussure, and no amount of emphasis on the mechanics of the procedure will undo the weight of this gesture" (26). To put it another way: only by extending the diphonic to the scope of the lexical could Riffaterre succeed in (de Man again) "assimilating, through the mediation of the reader, phenomenal intuition to semantic cognition" (22). In line with de Man's critique, there is a rare moment of phonotextual—indeed, transegmental—audition in Riffaterre which confirms the primacy of the lexeme in the very process of allowing its momentary disintegration. A sonnet of Mallarmé's is the subject, including the line "Avec ce seul objet dont le Néant s'honore."[26] After two pages of commentary, Riffaterre observes the sole instance of cross-lexical homophony I have found noted anywhere in his work, "a pun to top it off"—or, in Ricks's terms, an anti-pun, since a complete alternative grammar is never activated by it. Riffaterre hereby registers the way "Néant s'honore sounds like néant sonore " (18), so that the idea of nothingness taking pride in objects is doubled by the idea of "sonorous nothingness." The fact that sonority itself at the level of the phonotext, silent enunciation, is what catalyzes this phrasal self-transformation, this evacuation of one phrase by its double, goes unmentioned. Typically, though, the threatened referential anarchy of this liaison is fended off by the immediate reassertion of a lexical alternative to the dissolving reflexive phrase, one that not only shelters the overall sense from disruption but renders it thematically efficient.
Faced with this signal departure from Riffaterre's semiotic norm, one might infer from it what Lotringer did from Saussure's excursion into ana-
grammatic reading: that an entire system (poetic semiotics here, linguistics previously) is based upon the exclusion of just this sort of textual permissiveness. If Riffaterre were to hear texts consistently "produced" at this level of phonemic free association, the paragrammatic structures of his method might soon find themselves in irreparable disarray. De Man's point about a "re-lexicalizing" retreat from the Saussure of the anagram is thus clinched here by counterexample. I stress this sense of Riffaterre's referential conservatism because it helps, in turn, to situate Roussel somewhere about halfway between Saussure and Riffaterre, committed to uninterrupted linearity but not to lexical stability, permitting the counterintuitive intrusions of phonemic latency upon visible script even within the relentless phenomenality of an audited text. In response to such texts, then, we need a "phenomenology" of the listening eye: a theory of the reading body in its silent activation of the voiceless but evocalized text.
Bodying Forth
Since Roussel's dramatic breakdown in the semiotic premises of signification returns us to the undulations of the language itself, it cannot help but raise issues of vocality upon which language is in one sense based. The "process text" therefore facilitates our commuting between two very different definitions of semiosis, that of Riffaterre and that of Julia Kristeva. On her account, semiosis has a decidedly somatic dimension, originating as it does in the maternal core or matrix, which she calls the "chora," a zone of vocal production that precedes articulation by providing its raw material. Demarcated thereby is an amorphous stratum of vibratory sound production, a phonation without phonemics, a lingual impulse before language. It is the origin and fund of all speech, rather like the "reserve" described by Foucault, and it is even materialized in scripted language as a vibratory pulse, a phonic counterpoint, in certain of the definitive modernist poets Kristeva studies in Revolution in Poetic Language .[27] In characterizing their experiments, she uses the very term, "pulverizes" (51), that Foucault applies to Roussel's language; it names for Kristeva the way in which the breakthrough of the semiotic into the symbolic disintegrates the logic of the signifier, let alone the signified, and drops the symbolic order back into that roiling phonic source that language exists to regiment. It is something of this heaving, sieved depth that Roussel's "process" may be found to extrude—and rethink—on the very surface of his texts. Such a momentarily disclosed and arbitrarily marshaled assertion of raw signifying energy is therefore a disturbance at the threshold of language and the threshold of consciousness at once. It is activated, I stress once again, through the body's as well as the mind's encounter with the text—through a reading effect that passively evokes even without engaging the phonatory
apparatus by which script passes (not back but over) into speech. This Foucault, though ignoring, has moved us into position to understand.
Before we pursue further this receptual effect, it is important to look again, and warily (as we began doing in the Prologue) at the historicizing strain of—and strain on—such linguistically oriented readings of modern poetic and prose discourse as those of Kristeva and Foucault. Literary history tends to fade away when the focus is too narrowly aesthetic, charting a monolithic development climaxed by a radical upheaval at the end of the nineteenth century. But what else might explain a Roussel or a Joyce, a Mallarmé or a Pound, besides the inherent aesthetic mechanism of language itself, delved to new depths? An indirect (and, I find, partial) explanation has recently been offered by Friedrich Kittler, philosopher of media history, who traces the decline of Romanticism to the invention of phonographic storage media.[28] After centuries of uncontested and unexamined discursive preeminence, the medium of print was suddenly challenged on precisely the ground Romanticism held dear: the power of literary language, that privileged subclass of print, to evoke and incorporate the sounds of nature in textual evocation (or even simulation). With the beginning of phonographic technology, argues Kittler, literature lost not only its recording or storage monopoly but also its unquestioned supremacy as a representational medium. The sounds of the world, including voice itself, could now be far more closely approximated by a mechanical mimesis. Kittler thus recounts the history of a competitive and embattled advance in communications media, one innovation superseding the transmissive capacities of its forerunner, with post-Romantic print surrendering much of its representational authority to phonographic recording. Literature, Kittler claims, fled from the arena of its own dwindling prestige into a more resolute bookishness, a foregrounded textuality, an abdication of mimetic obligations, especially as regards the speaking voice—fled, in short, into the brandished textuality of early high modernism. But to Kittler's revisionist departure from Foucault's autonomous teleology of language theory, there is yet another counterresponse in return. Kittler may well be right about Derrida's failure to account even tacitly for the place of phonography in the origins of structuralist (differential) linguistics, with its dependence on the minute calibrations of language's phonemic materiality. If phonographic recording can be said to have lent empirical substantiation to the division in modern linguistics between differential values and sheer acoustics, this alone may justify the present theory of phonemic reading even after a full absorption of the grammatological critique. But is the historical causality as linear, after all, as Kittler suggests? Is it necessarily the case that literature as a whole, faced with the phonographic erosion of its mimetic empire, necessarily beat a retreat into an opposite, mute, purely graphic conception of itself? Where was
the competitive spirit to compensate literature's anxiety? Is it not likely that modern literature instead attempted to preempt its partial eclipse by sound recording with its own syncopated phono/graphy?
In the case of this study's first example, as drawn from J. L. Austin, if "I stink" were to be given along with, or in place of, "iced ink" in a phonographically recorded utterance, ambiguity could either be maximized or prevented by the voice assigned to its articulation. Writing is different. It institutes a different difference. On the page, a transegmental provocation like "iced ink" is set by inscription; it cannot be ambiguous in the same way as it might be in a recorded speech. It is definitively inscribed, decided—but, then again, not quite. The phonic momentum that interrupts its lexical contours and aurally redistributes its spacing is not there in the phrase, not scripted. But it exists—in the phonotext. It is brought to bear only in the ungraphed phonism of reading, as it could never be so obliquely in sound recording. This is a way of acknowledging that the evocalization of transegmental alternatives is not sonic at all but phonemic. And such evocalization is only one among many ways—if perhaps the one with most theoretical edge—for texts to face down the phonographic advent through a stereography (one might say) of their own, with one word inscribed, its homophonic double copresent as the trace of a lateral displacement. Surrendering its supposed hold on the body of the world's sounds, literature makes an adjusted claim on the passive body of a reader voicing: voicing, that is, under erasure.
As an inevitability of "phonetic writing" that tends to be recruited by the opacities of literary style, such phantom stereography is to be found, of course, in Shakespeare as well as in Joyce or Woolf. As a shoring up of literary vocalization against the mimetic triumph of sound recording, its late nineteenth-century exaggeration may well be read as a specific stratagem of modernist inscription—recruited, concentrated, and explored as never before. And this, quite possibly, from within the newly defensive presumption of a nonphonocentric, purely text-based aesthetics of inscription. The modernist gambit would then read like this: lest phonography preen itself on a wholesale advance over literature, it must be noted that there is a more supple and striking (e)vocalization—bivalent rather than simply ambiguous—available to the scriptive text than can be found in the straightforward naturalism of acoustic recording. On this reading of media-historical anxieties at the level of material signification itself, literature did not so much back off from phonography as attempt to trivialize and overshadow it with an increased reliance on its own textual graphonics.
At this stage in surveying the theoretical implications of phonemic reading, questions obviously begin to regroup again around the idea of the reading effect and the reading subject. If speech—that is, language—does not origi-
nally reside in the body as talk and become dematerialized when given up to script, then is it the other way around? Does writing alone vest language with the only substantiality to which it can ever attain, the material substance of script? If speech is not the voice of the body but, rather, a special case of writing, then is writing the sole body of the voice? Opposing sides of such questions are debated by Denis Donoghue under the headings of "epireading" and "graphireading."[29] The first coinage comes not from epi but from the Greek epos , for speech or utterance. "Epireading is the reader's form of compensation, making up for the tokens of absence and distance which he finds in written words" (98). Nothing could be farther from the Derridean position, except that the "tokens of absence" are at least recognized as such—before the attempt is made to overcome them. The restoration of text to its human context in "epireading" includes the tracking back of script to its origin in an authorial voice, as well as the leading on from referents into the world they call up. Epireading thus "interprets experience in terms of voice, speech, utterance, logos understood as action and for that reason dreads the reification and idolatry of language" (151).
To this method Donoghue contrasts the preoccupation with graphic signifiers, with text as scriptive system rather than mediated voice. It is a tradition that for Donoghue runs (down) to the present day from its most illustrious culprit, Mallarmé, through (quite selectively and variously) Derrida, Barthes, de Man, to its last gasp in Lucette Finas. In this tradition, on Donoghue's understanding, the author is so violently excised from every textual consideration that the vacuum thus opened draws off the reader as well.
Though not mentioned by Donoghue, there is a particularly clear instance of the former position in John Vernon's Poetry and the Body ,[30] which locates its titular conjunction at what seems to me the wrong end of the semiotic continuum between signification and reception; it thus makes the link between textuality and materiality more metaphoric than it need be. Indeed, Vernon's resistance to the tradition of Mallarmé, in favor of the bardic line of Whitman, recalls Donoghue's characterization of Mallarmé's own resistance to vocal plenitude: "If the word were to be defined as spoken, the sovereign author could not suppress himself in its favor; he would be dragged into the poem with every audible breath" (155). A deconstruction of the authoring (never mind the reading ) subject does indeed take the wind out of inscription, robbing it of that "marriage of flesh and air" (2) which Vernon sees as the essence of poetry in the Whitmanesque tradition. By contrast, the work of this chapter would find the materiality of signification located in a script wed to the possibility of phonemic articulation as its very definition. It is only right that a conception of the body should return to our notion of literature through such a
back door, should return through at least—if no more than—the possibility of reading aloud. Vernon wants something more, but something less plausible. In championing the poets of song rather than scribble, chant rather than jot, Vernon insists on his position that "language is the shadow of speech" and that, in poetry alone of all writing, such speech is bodied forth again in a fuller utterance, supple and sinuous, a "dance" (58) of form. What is never admitted is that no matter how far these impressionistic figures (and they recur frequently) may carry the field of concern away from poetry as an acknowledged text of writing, the dance of its flexed and corporeal form could never be more or less than choreographic at base.
For the present claim that texts are ultimately produced by the latent mouthing of the reader's body there is, beyond intuition, a considerable weight of scientific evidence. In a study concerned less with psycho- than with physio-linguistics, Silent Speech and Silent Reading , Ake W. Edfeldt investigates the process of unvoiced reading by analogy with that hypothesis in the latter part of the nineteenth century which conceived "thinking to be more or less restrained speaking or acting."[31] So too reading. Summarizing decades of elaborate mechanical and electrographic experiments designed to test the muscular responses (of larynx, pharynx, tongue, palate, lips, and so on) during silent reading, Edfeldt concurs in the evidence that reading is a kind of blocked or inhibited speaking, accompanied by suboral "vocalizations." Such findings, one presumes, are exactly of the "behavioristic" sort we have watched Wellek and Warren dismiss in the Prologue, and yet they are invaluable in any full-scale phenomenology of the reading act. Wellek and Warren, of course, invoke the "literate" reader as a standard for nonphonemic reading. It may be, instead, that we can define literature itself as precisely that textual practice whose impedances to the flow of sheer script recover within reading some of the lapsed powers of illiteracy. According to the experimental findings summarized by Edfeldt, one should more properly speak of silent reading rather than of silent listening, with an emphasis on passive vocality rather than on the figurative mind's ear. With auditory subordinated to motor habits, the reading process becomes more somatic than imaginary. Implications for specifically literary reading emerge only when Edfeldt's findings suggest that the more stumbling blocks a dense or difficult text throws in the path of comprehension, the more silent vocalization is likely to be produced. Such is the conjunction of activated text and passive anatomy, however, that an approach like John Vernon's in Poetry and the Body never takes up.
It does surface implicitly as an issue, only to be at once foreclosed, in one of the most influential theories of fiction to have gained currency in the last several years. We shall return in Chapter 5 to the potential application of M. M. Bakhtin's "dialogism" to the dualisms of transegmental oscillation in
prose. For now, it is Bakhtin's understanding of the temporal and spatial coordinates of the textual site (as well as the narrative situation) that momentarily raises the possibility of a theory of evocalization. In Bakhtin's account of monologic (epic) form dispersed into the heteroglossia of the novel, he stresses an analysis of the "chronotope," the "time-space" of fictional form.[32] He then attempts to extend this bipolar determinant deep into linguistic structure, on the one hand, and out into the realm of reception, on the other. Besides the general temporal and spatial relationships pertinent to representation and plot, there is the level at which, for instance, nouns of person, place, and thing are rendered metonymic, syntactic—hence cast into sequence, into time. Such a metonymic factor inherent in grammar must also be said to meet its extrinsic reduplication in the serial pace of reading. In designating the "chronotopes of . . . the listener or reader" (252), Bakhtin skirts very close to Riffaterre's later position in regard to the reader "creating" (or producing) the text: "We are presented with a text occupying a certain specific place in space; that is, it is localized; our creation of it, our acquaintance with it occurs through time" (252). Just a sentence before, he has also suggested that in our response to "the external material being of the work," its status as an inscribed system of signifiers, "we not only see and perceive it but in it we can always hear voices (even while reading silently to ourselves)."
That last remark, however, cannot be taken in the spirit of the present investigations. Bakhtin's figurative sense that we "hear voices" has undergone no deconstruction. Pointing away from the reader at this point, it is openly phonocentric, for Bakhtin adds in clarification that "we always arrive, in the final analysis, at the human voice," always "come up against the human being" (252–53). This confirms his sense that both writer and reader inhabit a chronotopic world (or worlds) different from the represented world of the fiction, less dialogic, more "real, unitary" (253), and that the novel is the point of intersection between the author's time-space and ours. One thing to say about this is certainly that it has overstepped the bounds of textual theory. For a moment, though, if only in passing, Bakhtin seems to have alluded to another chronotope given voice at the receiving (or creating) end of the text, a voice generated not by its origin in the spatiotemporal realm of the author's world but produced instead by the activation of the text's material basis as a signifying system. This is the chronotope of the reading rather than dictating voice, of the reading that voices "even while reading silently to ourselves." It is the "time-space" inscribed by both the duration and tempo of reception in the dialogized locus somewhere between scripted text (in a phonetic alphabet) and the enunciating inner voice of a reading subject. Marked out in this way, again, is literature's double construction at the reading site.
Earsay
In contrast to such mystified "epireading" as we find at times in Bakhtin or Vernon, Donoghue closes his survey of "graphireading" with an extreme case of avowedly Derridean commentary by Lucette Finas, including the fullest quotations from her work yet translated into English. The way she replays a text involves the modes of transegmental reading more often than anything else I have seen in criticism—outside, that is, of a few phonemic flash points in untranslated sections of Louis Marin's book on autobiographical "voice."[33] Finas identifies herself (in Donoghue's translation) as interested in "the repercussion of the smallest shift, whether graphic or phonic" (192), while in fact her emphasis falls on the phonic "discomfiture" (her word) of graphic expectations. She calls it "cacophony," and is ready with illustration. Mallarmé's "Yeux, lacs avec ma simple" is heard to "trip up" eyes other than those it names, one phoneme following too closely on another for lexical discretion: "lac(s) avec, Lacavec, where, as a further chasm can be heard les yeux là, caves , those abysmal and cavernous eyes." The slide is not just in Mallarmé's verse but in Finas's own prose. She edges from chasm to abyss, from phonological figure to theme—as if the text knows no distinction between a verbal lacuna and a described lake. Finas is rarely so open (though, even here, not explicit) about the linguistically reflexive bent in her readings, a procedure by which the contour of sound remodels the semantic content. Indeed, she is elsewhere less insistent even about the subversive malleability of phonemes; in her chapter on Paul Claudel's prose poem, "Décembre," the prosodic term "hiatus" (rather than the figurative "chasm" earlier) is introduced to characterize "contrée et ce"—without mention of an elision blurring over the internal caesura.[34] Just as often, though, Finas will actively pursue the phonic shuffling beneath the graphic mark, as in a passage on Mallarmé's "Salut"—not translated by Donoghue—where a line like "Solitude, récit fait toile" is traversed by the bidirectional "liaison" that dissolves it into such a variant as "Solitude, récif et toile." It is a liaison both "indicated and contradicted at once" (122)—in other words, written off by script even while sounded out. Such a textual listening is often, as here, unattached to any thematic logic. Following the title of Finas's collection of essays, Le bruit d'iris , it is the unprogrammed and cacophonous "noise of the eye" that is her most original subject, a poststructuralist equivalent, in effect, for Gertrude Stein's "sound heard by the eyes."[35]
In a methodological note (not discussed by Donoghue) to her reading of Mallarmé's "Salut," called "Interdicted Plays, Forced Rapprochements," Finas portrays her readings as "a process of integration by disintegration at
the boundaries of the supportable," drawing on the manipulated sonorities of a "broken text," the "tremors" and "rebounds" of the "phonic zigzag" (126). She specifically rejects any suspicion, however, that her commentary is "playing according to Saussure," in any way "searching for some hypogram(s)." There is certainly no dispersed Orphic theme word here to be reconstituted, and yet no reading could be more reliant on the diphonic principle of Saussure's within the domino play of collapsing syllabification. Every bit as much as Saussure, Finas is committed to following out the text's "dilapidating of syllables and sounds" (125), her own chosen verb form serving in this case to negate the text as lapidary inscription, to chip away at any presumption of its marmoreal stability. Finas has thus brought to fruition the libertine threat, the diphonic free association, that passed under unspoken censure by Saussure as he left the anagrammatic studies behind. While denying all comparison to these studies, Finas nevertheless makes good on Saussure's checked promise of the open text. At the same time, she distances herself from the Freud of the lapsus , the loaded slip, and hence liberates the quasi-oneiric iridescence of language from any secretly prescribed agenda in the unconscious, as straitening as the Saussurean anagram in its way. In recapitulating the effects she proposes to disclose in a text (or generate from it)—"to retrieve (or produce? that's a question)" (124)—her stress is placed, as in no other literary commentary, not even Derrida's, on such transegmental shuntings and "generalized rhyme" as "hormis l'y taire" turning into "militaire" (123), "salarie" into "qui sale a ri" (127), or the slide from "s'honore" into "sonore" (126) that recalls Riffaterre's one unpursued foray into the possibility of a phonemic as well as a lexical matrix.
In the single essay on Finas's work that has been translated into English, Christian Prigent aptly compares her "echolalic throbbings"[36] with the "mechanism" of the Roussel text (45). Finas finds, in Mallarmé particularly, a kind of perplexed lexicon given to the "stroboscopic ebb or flow" (47) of syllables and sounds. Explicitly alluding to Kristevan semiotics, Prigent locates in Finas's readings "an anamorphic torsion of the pheno-texte" striving for "an access to the prodigious or monstrous theatricality of the geno-texte" (47). Though unexplored in his commentary, this approach to literary textuality inflects her own discursive emplacement as analytic agent. Throughout her undoing of Mallarmé's "Salut," Finas also deliberately undoes her own expository position. She speaks of her labors with an italicized first person pronoun (Je ), as if it were a foreign word even in French, and always in the third-person singular, as in the glossed elision of "I detaches 'sept' from 'cette écume'" (my emphasis in translation). Only in view of such pronominal foregrounding can one appreciate the full force of Roland Barthes's vocative
gambit in a revealing preface to her work, "Question de tempo," subtitled "à Lucette Finas ."[37] Such a preface to her—couched in the form of second-person address, a little epistle—might seem quaint, curiously avuncular, even patronizing, were it not heard to utter its commendations from a station (in respect to Finas's Je ) equidistant from its own textual surface.
In this late essay, Barthes thus seizes the occasion to review and condense many of his own most challenging textual postulates. As in Finas's writing, Barthes's I-Thou pattern speaks out of nowhere and to no one, a dialogism without voice. he shares her interest in the tempo of reading as it guides the production of the text. He is especially intrigued, further, with the notion of "excess" (151) in Finas, an excess that various speeds of reading struggle with or shake loose, according to "the battue (the beat) of the text" (152)—the text, in other words, as performed. This "excess" is what Barthes sees as the chief contribution of Finas's readings, her instinct for those "oppositions of intensity" (152) left out of the binary calculations of structuralism. Finas finds an impertinent overplus haunting the pertinent opposition, and since this finding is in the reading, not strictly in the text, she, according to Barthes, "turns reading itself into a text" (153). When such a text of reading is in turn read, Barthes goes on to imply, the question of the reading subject comes into focus. Pushing beyond structuralism, the reader reads herself as structured by the reading, polysemously. This is what Finas no doubt means when, in the last paragraph of her chapter on Mallarmé, having rejected both Saussurean and Freudian premises, she returns to the lure of the "psychoanalytic siren"—only to turn it on its head by offering that "the text of Mallarmé, instead of being, itself, analyzed, analyzes" (131). This is where the reader comes in—comes in for deconstruction, as well. Reading the modernist text, reading the writerly, is a participatory decentering; by letting the tempo of reading remain undetermined in advance, contingent, variable, Finas puts herself, her Je , under the textual sway. As Barthes's own text summarizes it in grammatical address to her(s): "In opening up the tempo of reading, you are therefore opening up the subject. Your reading is therefore, to my mind, strictly materialist, more materialist than many materialisms which announce themselves as such" (153).
Its material base is, of course, that of the signifier, in perpetual counterpull between phonic and graphic demarcation. Read by her, that signifying insurgency by which words are spent and replenished on the run operates through the disruption of the linguistic continuum with wellings from the semiotic pulse below, from lalangue, from an acoustic latency of pure excess, unimpeded, unshaped. Having no access to the regimen of script, this is a super-fluency before language, in part genotextual, in part expendable. In this sense we might want to say that textuality enacts as well as lays bare a contradiction.
In the rhythmic tempo of the modernist text, as read by Finas, the bleeding between lexemes accomplished by evocalized script—by predication and (let us say) verberation—highlights a phonotexture that seems at once origin and "supplement," the trace as both basis and dividend. Such an effect—so often a homophonic phantom in her reading, its very "materialism" half disembodied—emerges as just what is left over (and, in most readings, left out) after phonic impulses are channeled through script along the ruts of words, the route of syntax, the routines of the sign. It is the remainder after lalangue is pieced out to la langue in order to be activated as parole . One thus confronts head-on the variance of the genetic from the heterogeneous premise, the primalist from the supplementalist doctrine, by understanding the phonemic trace as an excess at the source . "Graphonic" reading would measure, then, a surplus inherent in the fund of semiosis—as the very definition of its emergence into symbolic script. This emergence not only leaves traces, it is the trace in its first instance, the first nudge of mumble into utterance. For all the interknit distinctions upon which language is mounted, the originary (though never fixed) difference—that of language from murmur—is what marks the path of prelingual phonation on its way to phonemic articulation. Difference is first measured, therefore, from that primal voicing against which any functioning vocable, or its constituent phonemes, must be thrown into relief. That throwing, that veering, that verging is therefore the trace in the making.
I have used here the Derridean term "supplement" not lightly but with the same idiomatic sense to which Derrida appeals in order to vex it into reconception.[38] In the Western tradition at least since Plato, and most overtly in Rousseau's Confessions , writing has been considered the "supplement" to speech, subsequent, ancillary, denaturing. With Saussure, at last, emerged the definitive structuralist science of that subordination, under the founding premise, as we have seen, of "acoustic pattern." Derridean grammatology does not simply avert this elevation of speech over writing, it tends to reverse it. Language is not possible without spacing. Spacing is not acoustic in speech production. On this score alone, language has at base a spatial or graphic, rather than just temporal and phonic, dimension. Defying the traditional assumption, therefore, Derrida finds in spoken utterance something more like the supplement of the gramme. My whole project in aural reading begins, in one sense, by accepting the Derridean reversal (itself unfixed) of origin and supplement—as a quintessential case of the deconstructive agenda in action: a reciprocal overturning of priority rather than the rigidifying of a new hierarchy. In this spirit, the "graphonic" relativism of phonemic reading finds in the process of subvocal text production a uniquely perceived instance of the mutual supplementation of gramme and phone. In the full paradox of the
Derridean supplement, the inner is constituted by the other, the outer, and identity defined by difference. An important parallel follows from this. Let us accept that speech—as language rather than as sound—is differentially defined not from without, by writing (as its secondary transcription), but from within, by the gramme (as its formal principle). It may then emerge that reading—understood not just as recognition but as cognition—should be defined as a response not merely to graphemes but to their difference as—as well as their difference from—phonemes.
At its transegmental crux, then, the phonotext is produced as follows. A phrase has been written where no succession of sounds has been made. A word is read with no presumption that a voice either preceded that textual mark or can be elicited from it. Yet a voice is brought to it, one that often produces an adventitious, potentially submorphemic "sound" even when the written word is over—a voiced though silent sound, one that sticks, whether borrowed in advance from the successor or lent to it in the aftermath of the lexeme it has already served to complete. The phonemic "supplement" is thereby created, created upon the trace of a grapheme, but no sooner created than expropriated, incorporated, internalized, made constitutive, changing one root or pluralizing another, say, but in any case found momentarily inherent in—rather than adhering to—a lexeme to which it has no graphic attachment. This inherence is the very "production" of the lexeme as difference—but not just different from the word as written; activated here is instead that (both morphophonemic and grammatological) difference within that structures only by opposition. The phonemic supplement is thus actually recirculated through the written word as the revealed dynamic of its generation: its deviance from, and so deferral of, alternatives. In this sense, a given word is distinguished from its near miss—and phantom double—by that very binary logic which the principle of the supplement leads us to rethink at a more paradoxical level. One word's difference from its alternatives in the paradigm is thus kept before us, as a signifying principle, by a given word's shifting difference from its successor in the syntagm—and this by way of that phonemic supplement which co-opts the gramme of the blank in its transit to a new lexical bond. Such transegmental drifts, exposed as they are by phonemic reading, isolate in this fashion—at the level of a single linguistic feature rather than governing principle—the full paradox of the internal supplement. In literary reading, especially, such drifts operate, in short, to sustain the structuring "versus" of binary opposition as a viable register of the final, never-finished text.
But the reading mind doesn't stop with this partly somatic turbulence of wording within meaning. In Finas's phrase, the receiving mind doesn't just listen to the noise of the eye. It processes. It concerns itself with the pro-
nounced effect . It reads . Which brings us round again to how, and why. By pressing the act of reading to the outer edge of coherence, Finas serves to demonstrate a textual conservation, if not conservatism, at the base of even the most radical reading projects. She only goes so far, so far out. The distance she gains on more subservient readings, ones more deferential to the supposedly given text, only marks a far perimeter within which the text continues to function . Her maneuvers are curtailed only by the value she implicitly places on such continuance. The hold of language upon Finas's readings, even as it relaxes back into lalangue , is still strong. Syntax undergoes a rending—but not a complete surrender. Why it does not, could not, is a question whose answers spread wide across interpretive practice.
It would appear that there are three phases of coherence which Finas sporadically reimposes on the textual decomposition she performs: the syntagmatic, the mimetic, and the reflexively thematized, the last two often accompanying each other. These three phases of reading are neither defended nor even confessed. yet they bind her, if loosely, to an inescapable logic of interpretation nonetheless, however much she protests against it in the name of a freer reading. Though langue , or momentarily even lalangue, irrupts into parole , the law of "speech" is not abolished. The message regroups. New words slip in on the coattails of the vanished. Language unveils itself as mere traces, but it remains language nonetheless. Lexical segmentation may shiver to bits, but the phonetic shards are reheard as a new syllable; the lexicon is consulted for a new morphological opportunity, the syntagmatic axis kept intact—at least for a moment, even if no stable double grammar emerges. That is the first kind of normalization to which the ear's heresy submits: grammar, in short, is maintained. Not only mitigating disruption, the second kind of normalization renders it imitative—if only (and here we move to the third type) imitative of itself as text. When "lacs avec," rather than allowing its first c to be silenced, asserts it in a segmental ligature with the adjacent word to form (evanescently enough) the word "cave," its play with the lexical abyss is called "cavernous." All three persistent normalizations are here on view: a sequence minimally cued by grammar that is also a mimesis of the phantom referent, hollowed and resounding, and a thematization of the verbal blank.
Finas's work is too semiotically sophisticated to have much converse with mimed signifieds per se, so the opportunities for style-as-enactment are usually rooted reflexively in the signifier alone. Wordplay is ordinarily able to reproduce only the play of its own words. Nonetheless, the three recuperative moves (syntagmatic, mimetic, and reflexive) do find themselves operating at times together. Shattered, gaping, or set loose, syllables reconvene as words, while the loosening, gaping, or shattering—figure it how you will—is recovered by the textual mechanism as part of the reconstituted message. It is,
further, a message that is regularly folded back upon the text in the form of a metaphor for its very textuality. In other hands than Finas's, or upon other ears, what is as much dutiful as willful about this procedure, what continues to respect the contract of interpretation between text and reader, would come clearer sooner. And could too easily be dismissed. Finas holds the line partly by obscuring it. Which is why I labor into the open the assumption of her procedures, or, more to the point, their inevitability. There is no other way to read. This is not because we are autonomous subjects reading our confirmation in the order of a totalized text. On the contrary, it is because, as constituted subjects, as never better than italicized I 's, we are defined precisely as readers , processors of signs. Texts, deconstructed, ask of us, in reading, the very work that in every sense makes us up.
An analogy dawns between Derrida's critique of phonocentrism and Finas's critique of the psychological subject, as elucidated by Barthes. Both attacks take too little account of the adversarial provocation: the respective appeals, physical rather than metaphysical, of voice and person. Those appeals happen to converge. They are two different formulations of the body as site—what we can call writing's productive supplement in becoming text. Neither authorial source, on the one hand, nor envelope of identity, on the other, this is the bodied site as heterogeneous locus: the body before the book, the body in place. It is because we hear overtones of our voice when reading to, and out of, ourselves—I have argued—that we are tempted to posit a sense, or sanctify a relic, of voice in text. It is because we have an "articulated" anatomy, upon which are strung vocal chords, that we read in this way, read in our own voice. Though our body is not a subject, nor even the outer form of one, it is often most of what we think we can demonstrate of one, as proof of one. This assumption, too, must no doubt be resisted. But it must also be recognized as a particular chimera of reading. The body is our guarantor, especially in "going out of ourselves" (never quite) in reading the words of another, the Other. Situated before a text, whatever its "voice," I have mine, whether I activate it or not; and if what I think I am reading is sayable (that is, meaningful), out loud or not, then it easily confirms my powers of reception as those belonging to me, to an I. The readerly leaves me be, comfortable in my mastery of it. The writable (which is certainly not to say, the exclusively scriptible in Barthes's postclassical terms)—the text in whose production I participate—instead exposes me to incoherences, exposes my own, or that incoherence which is no longer to be called "me" at all. The body's vocal apparatus thus colludes in a reception of texts in a manner that keeps that body in its place, no longer coincident with person, no longer identical . Just as the phoneme, properly recognized, can assist in the deconstruction of the Voice,
so can the body (as the site of a voicing), properly occupied in reading, attend at the decentering of the very subject whose warranty it ordinarily provides.
Epicentrism
Midway between Donoghue's polarized concepts of "epireading," named for the epos of voiced presence, and "graphireading," named for the scripted signifier of all textuality, falls the spectrum of response we have been reviewing: a semiotic reading in Kristeva's sense; a Lacanian reading sensitized to lalangue ; and, by way of Finas, an exploitation of the text in its full phonemic and syllabic fluidity (even though she is consigned exclusively to the "graphic" dimension by Donoghue). Into this space between polarized camps falls also a paradoxical concern for the "grain of the voice" in "vocal writing." I allude there to Roland Barthes's remarkable last two pages in The Pleasure of the Text ,[39] as well as to the essay specifically titled "The Grain of the Voice" which I considered in the Prologue—as both would seem to anticipate the signal phonemic interferences in Louis Marin's restoration of the "excommunicated voice" of textuality. Though Barthes outlines in The Pleasure of the Text an aesthetic of vocalized textuality, his remarks never specify their implicit assumptions about the scene of reading. "Writing aloud ," Barthes suggests, "is not expressive; it leaves expression to the pheno-text, to the regular code of communication; it belongs to the geno-text, to significance; it is carried not by dramatic inflections, subtle stresses, sympathetic accents, but by the grain of the voice, which is an erotic mixture of timbre and language . . . the art of guiding one's body" (66). By the end of the paragraph, one guesses, Barthes himself is attempting something very much like writing aloud, illustrating the "whole carnal stereophony" (66) of such textual production: "It granulates, it crackles, it caresses, it grates, it cuts, it comes: that is bliss" (67). Barthes here has moved well beyond even Empson's feel for literary language as texture, as skein, as skin, "whose flesh has the character of the flesh of an organism."[40] When Barthes stresses "the pulsional incidents, the language lined with flesh" (66), he is speaking, if we may say so, less metaphorically, for he has in mind the very "muzzle" of enunciation.
For Barthes, "writing aloud is not phonological but phonetic" (66)—in other words, more phonic than phonemic, its energy not yet bound up in units of signification. Not entirely , no, but at the same time incapable—in any given textual manifestation—of entirely avoiding the symbolic in the flow of its vibratory semiosis. So, after all: phonemic reading. Silent but material, such reading is disciplined by the laws of combination while at the same time freed by those of equivalence. So where does this leave us as readers? Donoghue would give us too little room. Between epos , as voice, and graph , as mark, we need a third term, a third position—a site, all but a breathing space, for the
reader's silent voice. A name for such a position seems implicit, as a matter of fact, in an alternate etymology of the term "epireading"—indeed, in the etymology we are likely to expect from the term at first glance. We ordinarily encounter that prefix epi in its derivation not from the Greek noun but from a flexible Greek preposition. We find it, for example, affixed to two English technical terms, one philosophic, one meteorological, that contribute to modeling a revised notion of the epitext in the ear. I have in mind "epiphenomenon," a secondary, derivative phenomenon, and "epicenter," a point somewhere above a decisive locus. In literary-critical terms, a phenomenological reading is of course bipolar, the text at the intersubjective center of an alignment between the consciousness of the author and of the reader. A decentering of the text, not to mention of subjectivity itself, breaks the axis, cancels the alignment, spins the text off into an eccentric orbit of its own. No phenomenological access remains available, except through the processed materiality of the signifiers. But an epiphenomenal approach—to the text as evocalized—is still possible once the text is recognized to transmit through the channels of its enunciation, and no more than mutely at that, only the voice of the reader. It is a voicing generated from his or her (if there is gender left at all in the reader as voice rather than person)—we had better say from its , from my —shifting stance toward the writing it activates. It is a stance taken up in part by my upper body's perch over the page—in other words, in the readerly epicenter of the decentered text.
From such a site, as well, the epiphenomena of textual generation may be said to jump off the page at me. It is this notion of reading as an open apprehension, rather than as a suppression, of perceived verbal ephemera that would connect with Hartman's allusion to "saving the phenomena" (or "the appearances" [xv]) in his book-length encounter with Derrida, Saving the Text , a "saving" that is not rescue so much as recognition and account. Part of this very "apparitional" dimension of reading is what he calls in the title of his second chapter "epiphony" (33)–and calls for in the imperative subtitle of a later section, "Look with thine ears" (128; from King Lear ). Throughout, Hartman posits as a central agenda of literary reading the effort to reverse print's tendency "to blind the ear" (142). For Hartman, the "words of a text, in their silence, are but divining rods to disclose other words, perhaps words of the other . . . and critical reading allows us to describe that interiority, to estimate words as words, to see them as living in and off us" (142), subjective and parasitic at once, waiting for us to effect them. This is their phenomenality, and only reading of a certain intensity and give can retrieve it from neutralization: "And that is why poetry makes its curious alliance with critical reading, in order to reactivate the ear" (142). Early on, and despite his own later suspicions about a Derridean resistance to the privilege of the oral in
Joyce,[41] Hartman tries to dispel the assumption that his allegiance to the ear is in any sense an anti-Derridean position, insisting that what may appear as the derogation of the phonic in Derrida is instead a more radical insistence on it—but on a phonic energy uncontainable by script. The "cadence" of the phonic element, that is, "cannot be encased in grammar or meaning. It falls through, into, both" (14).
Very much in the spirit of the present study, then, Hartman acknowledges that the phoneme can drop out of syntactic regimentation only to resurface by fractional displacement. Such material "fallout" constitutes exactly that aspect of textual phenomena most in need of "saving"—not so that writing should thereby be revalorized as "stored speech," that old myth of transcribed voice, nor because such a process in any way confirms the timeworn assumption that "written speech puts sound on ice" (42)—indeed, as Hartman has it, "the phonè is more sphinx than phoenix" (44)—but because writing itself, broadly understood, can be generated phonematically only in its manifestation as reading. Hartman is thus the least likely of those Anglo-American critics engaged with Derridean deconstruction to subsume textual productivity to sheer inscription within a logic of the trace. In the terms of Hartman's closing chapter, "Words and Wounds," it is just the impossibility of doing so—the impossibility of fixing meaning, even the locus of meaning, in script—that makes for the "wound" of "equivocation" (129), the double violence both caused and supposedly cured by literalism, by letteralism, by the stable and bounded sign. Further along, Hartman writes that "closure is a sealing with healing effect" (150), a boundedness that lulls anxiety. Yet literary writing is there to refuse this calming, this balm; art "creates limits that prove to be liminal. In poetry this 'sense of an ending' embraces rhythm, rhyme, and caesura: what happens toward the end of a line, or rhythmically within the line, until lineation itself is affected"—or, as I would add, what happens even toward the ends of words, until both lexicality and syntax are themselves affected, one might even say infected, by the wound of ambivalence. This seems much the sense of Hartman's own appended remark: "The compounding or tmetic disjoining of words is also involved" (150–51).
The problematizing of linguistic juncture is thus at the microlevel a rebuke to closure, whether of lexemes or even of single morphemes. Such resistance to closure is a lateral friction that appears in the phenomenality of the text as performed, an interference that can be preserved only in reading at a certain pitch of responsiveness. To draw the notion of the "wounded" word more closely into the orbit of this study in connection with such transgressed closure, one may contemplate a Derridean parody of the Logos in its descent into time: a miniature mythography turning on a single model of the universal lapse, or cadence, of the Word into the mutable world, of speech into writing.
The trace of the morphophonemic Fall I have in mind is manifest, diachronically and synchronically at once, in that etymological trajectory which has transformed the sacrilegious exclamation "God's wounds!" (subjective as well as objective genitive, let us say) by euphemism into "His wounds," by elision into "'S wounds," by liaison and exaggerated voicing into "Zounds," and eventually into a rhyme with—and virtual echo of—the sound of the word "sounds" itself. By such contraction and displacement, such coupure , the archetypal wound is made flesh in read writing. What the genealogy of an idiom has done in this case over etymological time, at another level intensive silent reading—from the epicenter of the subvocal body—cannot help but do over the space of a line: in every sense such reading lets things slip. In Hartman's terms, reading in this degree opens itself to the literariness of textuality by "saving" any number of such epiphenomenal slips, recuperating them syntactically even in their full ambivalence, making them tell.
With the receptive epicenter of textual production thus hovering above the page, reading becomes a demystified but still palpable, still material, speech act—silent, reclusive, unmastered, but placed . To such a place we will next, yet again, let textual evidence direct us. If, in doing so, we extrapolate a major Derridean paradox to the level of text production, such embodied reading will stand disclosed as the very "supplement" of that textuality for which it is also the necessary condition. Prepared as we are, then, by a number of linguistically oriented theories indirectly related to the evocalizing of one word over the edges of another, to the lap and palpitation of the acoustic stream, to the reading of phonemic overlay within lexical sequence, we are ready now for a return to the phonotextual fray—and to that graphonic interference, that silent vocal feedback, which it everywhere entails.