1. R.-M. S. Heffner, General Phonetics (Madison: University of Wisconsin Press, 1950), pp. 175-202. The types of such phonetic "fusion" include "dynamic displacement" ("dynamic" implying the sequential pressure of "motor context"), as when a phrase like "all but universal" tends to be pronounced as "all bu(t) tuniversal"; "doubling" (where the "back stroke" of one word becomes the "beat stroke" of the next), as in "ha ve v arious" or "an d d egrees"; "reduction" (an unwritten contraction), as in "extr ao rdinary" when enunciated as a five- rather than six-syllable word, or in "th' eternal"; "omission" (the contractive forms of aphaeresis, syncope, and apocope when graphically indicated by apostrophe); "glides" (related to "dynamic displacement," but involving the introduction of an adventitious transitional sound), as with "don't ch ew" for "don't you"; "linking" (a bridging of hiatus either through retention or insertion of a consonant), as in the strict form of such French liaisons as raconte-t-on des histoires; "adaptive changes" (shifts that alter the enunciation of a speech sound either to match--by anticipation or continuance--or sometimes to avoid reduplicating a proximate sound), a tendency often related to "doubling" in consonant conjunctions like "hi p b oots," "re d t ie," or "bi g c ake." In Heffner's own account of these categories, however, no likelihood emerges of any real morphophonemic--hence lexical, hence semantic--ambiguity, unless one could imagine a context, say, in which "Don't chew?" (the negative imperative cast into an interrogative fragment) could be misheard for "Don't you?" According to Heffner's meticulous phonetic transcriptions, the phenomenon of "fusion" is ultimately prevented between "speech measures, or phrasal groups" (p. 200), inhibited by the nature of speech rhythm itself; and between morphemes by the self-signaling "initial element" (p. 201), pronounced as it could never be in the position of a terminal feature (a phonetic difference, therefore, within the same functional phoneme). The "cessation of the commingling of movement patterns at the continue

end of a speech measure" (p. 200)--the arrest of fusion--thus also obtains at the boundaries of a morpheme; most potential fusions are therefore suspended by "intonational and accentual patterns" (p. 201)--what linguistics elsewhere calls "suprasegmental" features--if not by clear phonetic distinction.

2. Ibid., p. 3.

1. R.-M. S. Heffner, General Phonetics (Madison: University of Wisconsin Press, 1950), pp. 175-202. The types of such phonetic "fusion" include "dynamic displacement" ("dynamic" implying the sequential pressure of "motor context"), as when a phrase like "all but universal" tends to be pronounced as "all bu(t) tuniversal"; "doubling" (where the "back stroke" of one word becomes the "beat stroke" of the next), as in "ha ve v arious" or "an d d egrees"; "reduction" (an unwritten contraction), as in "extr ao rdinary" when enunciated as a five- rather than six-syllable word, or in "th' eternal"; "omission" (the contractive forms of aphaeresis, syncope, and apocope when graphically indicated by apostrophe); "glides" (related to "dynamic displacement," but involving the introduction of an adventitious transitional sound), as with "don't ch ew" for "don't you"; "linking" (a bridging of hiatus either through retention or insertion of a consonant), as in the strict form of such French liaisons as raconte-t-on des histoires; "adaptive changes" (shifts that alter the enunciation of a speech sound either to match--by anticipation or continuance--or sometimes to avoid reduplicating a proximate sound), a tendency often related to "doubling" in consonant conjunctions like "hi p b oots," "re d t ie," or "bi g c ake." In Heffner's own account of these categories, however, no likelihood emerges of any real morphophonemic--hence lexical, hence semantic--ambiguity, unless one could imagine a context, say, in which "Don't chew?" (the negative imperative cast into an interrogative fragment) could be misheard for "Don't you?" According to Heffner's meticulous phonetic transcriptions, the phenomenon of "fusion" is ultimately prevented between "speech measures, or phrasal groups" (p. 200), inhibited by the nature of speech rhythm itself; and between morphemes by the self-signaling "initial element" (p. 201), pronounced as it could never be in the position of a terminal feature (a phonetic difference, therefore, within the same functional phoneme). The "cessation of the commingling of movement patterns at the continue

end of a speech measure" (p. 200)--the arrest of fusion--thus also obtains at the boundaries of a morpheme; most potential fusions are therefore suspended by "intonational and accentual patterns" (p. 201)--what linguistics elsewhere calls "suprasegmental" features--if not by clear phonetic distinction.

2. Ibid., p. 3.

3. Jacques Derrida, Of Grammatology, trans. Gayatri Chakravorty Spivak (Baltimore: Johns Hopkins University Press, 1974), p. 63, quoting Saussure (see Course, p. 66, cited in n. 6 below).

4. In her Translator's Preface to Of Grammatology, Gayatri Spivak stresses this as the very principle of the written trace, or gramme, "the name of the sign 'sous rature,'" calling "grammatology" the "science of the 'sous rature'" (1).

5. Jacques Derrida, "Différance," in Speech and Phenomena, trans. David B. Allison (Evanston, Ill.: Northwestern University Press, 1973), p. 132. In engaging with the Derridean sense of such issues, I am attempting a very different level of textual reception than the more traditional approach, for instance, of Raymond Chapman in The Treatment of Sounds in Language and Literature (Oxford: Basil Blackwell, in association with Andre Deutsch, 1984). With a dozen chapters devoted to the linguistic representation of sounds, he only then turns to the sound of representation itself, the phonological aspect of scriptive signification. In his chapter "The Sound of Literature," he admits that the "implications" of the question "how far is sound inherent in written language" do "lead towards areas of literary criticism which would go far beyond the present study" (p. 210). In regard to the "counterweighting of possibilities" (p. 217) between eye and ear, he offers a brief section on how the "study of rhyme is illuminating for the relationship between sight and sound in language" (p. 215), with subsequent commentary on onomatopoetic devices and closing remarks on poetry in oral performance.

6. See Ferdinand de Saussure, Course in General Linguistics, trans. Roy Harris (La Salle, Ill.: Open Court Press, 1983).

7. See Jean Starobinski, Words upon Words: The Anagrams of Ferdinand de Saussure, trans. Olivia Emmet (New Haven, Conn.: Yale University Press, 1979), and Sylvère Lotringer, "The Game of the Name," Diacritics (Summer 1973): 2-9, reviewing the French edition of Starobinski on Saussure.

8. The generative model of this eccentric theory anticipates the tenets of a more sustained later methodology. Whereas Saussure saw an incantatory or ritual motive behind the encrypted name of the god or hero, Michael Riffaterre substitutes an entirely secularized logic of explanation in his encoded semiotics of poetry. As we have seen, the indirectly voiced core of the text is not, for Riffaterre, a "theme word" necessarily but often a fuller sense of a thematic idea. Literary language is in this way found to develop motifs that cannot speak their names--because to do so directly would be to fall from originality, to lapse into cliché. The unsaid matrix of a poetic text surfaces in continue

paragrammatic variants that are the virtual equivalent of Saussure's "hypogram" (a term Riffaterre also uses at certain points). Further, the "model" in Riffaterre--first and guiding form of the variant, which sets a pattern for ensuing conversions and expansions of the theme--would seem to have derived its own model from Saussure's "mannequin" or locus princeps: "a tightly drawn sequence of words which one can designate as especially consecrated to that name" ( Words upon Words, p. 33)--consecrated, that is, to the theme word. The "mannequin" closely resembles what a phonologist like Masson calls the "bracket," with the Saussurean example " Prima qui es " yielding up, for instance, the first and densest suggestion of the name "Priamades." Framed or bracketed by the mannequin, in this case, are the structuring pieces of the name, modules known as "syllabograms." Paul de Man has also mentioned, more briefly, the derivation of Riffaterrean semiotics from the Saussurean anagrams in his omnibus review of Riffaterre's work, "Hypogram and Inscription: Michael Riffaterre's Poetics of Reading," Diacritics 11 (Winter 1981): 24.

9. To borrow Saussure's own example of a diphone as base unit (with annexed monophones), the coded cluster TAE can be given by TA + E, T + AE, or TA + AE; but since it cannot be given TA + TE--which would require the letters "to be amalgamated outside time, as could be done with two simultaneous colors" ( Words upon Words, p. 30)--neither can the diphone be overthrown altogether, as would be the case with T + A + E.

10. Gilles Deleuze, "The Schizophrenic and Language: Surface and Depth in Lewis Carroll and Antonin Artaud," in Literature and Psychoanalysis, ed. Edith Kurzweil and William Phillips (New York: Columbia University Press, 1983), p. 329.

11. Otto Jespersen, Language: Its Nature, Development, and Origin (New York: Macmillan, 1922), p. 8. In line with childhood's illiterate transformations of the language as studied by Jespersen, one must confess that solecisms based on junctural ambiguity can creep into almost any written utterance; indeed, it was only in a late redrafting of one of the present chapters that I finally noticed how I had taken down in phonemic dictation from myself--and many times reread without correction--"making could on."

12. This seems to me what Barthes is suggesting as well when he asks rhetorically, "Is not the entire space of the voice an infinite space?"--adding, "No doubt this was the meaning of Saussure's work on anagrams" ("The Grain of the Voice," p. 272; see my Prologue n. 8).

13. Jonathan Culler, Ferdinand de Saussure, rev. ed. (Ithaca, N.Y.: Cornell University Press, 1987), p. 134.

14. Jacques Derrida, Signéponge/Signsponge, trans. Richard Rand (New York: Columbia Univ. Press, 1984).

15. Christian Metz, The Imaginary Signifier: Psychoanalysis and the Cinema, trans. Celia Britton, Annwyl Williams, Ben Brewster, and Alfred Guzzetti (Bloomington: Indiana University Press, 1977), p. 240. break

16. Walter Redfern, Puns (Oxford: Basil Blackwell, 1984), pp. 78-79, quoting Leiris's collection, Mots sans mémoire (Paris: Gallimard, 1969).

17. Paul de Man, "Shelley Disfigured," in Deconstruction and Criticism, ed. Harold Bloom (New York: Seabury, 1979), p. 60.

18. Michel Foucault, Death and the Labyrinth: The World of Raymond Roussel, trans. Charles Ruas (Garden City, N.Y.: Doubleday, 1986), pp. 38, 31. Foucault explores another example from Chiquenaude (1990), where the multiple meaning of doublure (understudy, lining, rehearsal) underwrites a more complex homophonic play (p. 27). After the early mention of a theatrical piece called Red Claw the Pirate (Forban talon rouge), the play's title returns, through metamorphic recirculation, as the phonemic doublet fort pantalon rouge . As with billard/pillard, the "metagram" turns once again on the alternative plosives /b/ and /p/ from the phonemic paradigm of secondary articulation, but it projects them onto the syntagmatic axis as latently copresent alternatives this time, the trace of alternation itself. Quite by coincidence, the most sustained Rousselian experiment I know since his writing, in a relentlessly farcical vein, is the Nabokov-like parody of a found manuscript dutifully transcribed and meticulously glossed without the fictionalized pedantic editor recognizing that the recovered verse fragments under examination are nothing more than Mother Goose rhymes in ludicrous French phonetic transcriptions. The absurd contortions of glides, elisions, and liaisons in Courtlandt H. K. van Rooten's Mots d'Heures: Gousses, Rames (New York: Grossman, 1967) amounts to a checklist of effects by which the Rousselian "metagram" was negotiated. Here Humpty-Dumpty is rechristened "Un petit d'un petit" (p. 1), Peter, Peter, Pumpkin-Eater, emerges as "Pis-terre, pisterre / Pomme qui n'y terre" (p. 8), Jack and Jill are fused in "Chacun Gilles" (p. 11), the Baker's Man buried under "Pas de caïque, pas de caïque, bécasse, mâne" (p. 17), and the whole rounded off with "TIENS, DE" in place of "The End."

19. Raymond Roussel, La doublure (Paris: Jean-Jacques Pauvert, 1963), pp. 45, 46.

20. See frame enlargements from Duchamp's Anemic Cinema in Arturo Schwartz, The Complete Works of Marcel Duchamp (London: Thames and Hudson, 1969), pp. 319-28.

21. Marcel Duchamp, Rrose Sélavy (Paris, 1939).

22. Louis Aragon, Paris Peasant, trans. Simon Watson Taylor (London: Jonathan Cape, 1971), p. 102.

23. It is a muted form of this same aural preoccupation that appears implicitly in Virginia Woolf's essay, "On Being Ill," her approach to the alogical force of poetic sonority at which we glanced briefly in the Prologue (see n. 20). When Woolf quotes two lines of Rimbaud without further specification of effect--" O saisons ô châteaux / Quelle âme est sans défauts? "--who, she asks rhetorically, "shall rationalize the charm?" (p. 200). Not Leiris, certainly not Roussel. But their work can nonetheless be taken to continue

analyze, even to systematize, the mysterious seduction of such phrasing. In doing so, such work, as if according to the very definition of modernism, places under intensive scrutiny the irrational machinations of language itself out of which poetry from any period has always drawn a part of its force. In Woolf's example from Rimbaud, we find the Leiris-like internal rhymes of " saisons " with " est sans, " the latter, when reversed, a sly chiastic belling of the former (given the liaison of "sans est"). Exaggerating the euphonious recurrence of poetry in general, it would be more like the radical modernism of Roussel to tease the double liaison of " Quelle âme est sans " into an irrelevant semantic echo of la maison --in contrast, somehow, to " châteaux ." In this line of willful homophonic effects, there also appears in Leiris's glossary just the sort of encompassing lexical slippages that--if separated by other textual material inserted to "process" the transition from one to the other--might well have been found in one of Roussel's experimental exercises: "AMEN--âme mène" (p. 74); "OCEAN--eux et ans" (p. 102). In the closed system of langue, with its fluid depths like an opaque "lagune" indeed (as Leiris's earlier transform suggests), one of the most accurate "glosses" on a word--as act of wording rather than as single lexeme--is always, as Leiris joins Roussel in suggesting, the transliteration of its constituent phonemes, the parsing of its material base.

24. These are terms drawn from Peter Brooks's chapter "Freud's Masterplot" in Reading for the Plot: Design and Intention in Narrative (New York: Knopf, 1984), pp. 90-112.

25. De Man, "Hypogram and Inscription," p. 22. The example given of Riffaterre's "phenomenal"--hence, in part phonological--criterion: "The French novelist and occasional theoretician Ricardou is reproved for claiming to discover the anagram 'gold' in Poe's phrase 'right h old ing,' despite the fact that the 'g' in 'right' is not sounded in English."

26. Michael Riffaterre, Semiotics of Poetry (Bloomington: Indiana University Press, 1978), p. 16.

27. Julia Kristeva, "The Semiotic and the Symbolic," in Revolution in Poetic Language, trans. Margaret Waller (New York: Columbia University Press, 1984). Without relying on the Kristevan notion of the semiotic "chora," Geoffrey H. Hartman, in Saving the Text: Literature/Derrida/Philosophy (Baltimore: Johns Hopkins Univ. Press, 1981), muses on the "intransitive intimacy" of certain verbal formations that bespeak "the wish to encrypt in oneself the womb--the maternal (paternal) source--of verbalization[,] . . . of chatter and babble developing into verbal thought" (p. 144).

28. From a series of lectures at the University of California, Santa Barbara, in Spring 1987, based on Kittler's Grammophon, Film, Typewriter (Berlin: Brinkmann and Bose, 1986), soon to be published in translation by the University of Minnesota Press. The dangers of transegmental slippage--for "practical" rather than poetic communication--are confirmed by an experience I once had in a noisy restaurant. I was trying to explain Kittler's thesis to continue

a friend who hadn't encountered his work, but over the din, I sounded, it seems, as though I was discussing a critic with the unfortunate name of Friedrich Hitler.

29. See Denis Donoghue, Ferocious Alphabets (New York: Columbia University Press, 1984).

30. John Vernon, Poetry and the Body (Urbana: University of Illinois Press, 1979).

31. Ake W. Edfeldt, Silent Speech and Silent Reading (Chicago: University of Chicago Press, 1960), p. 13.

32. M. M. Bakhtin, The Dialogic Imagination, ed. Michael Holquist, trans. Caryl Emerson and Michael Holquist (Austin: University of Texas Press, 1981).

33. In chapters focused on Stendhal's autobiography, Louis Marin posits the nature of discourse as "necessarily auto-bio-phonic " and speculates on the "ruses" and "machinations" by which space is made therein--the space of the "silence of the saying within the said"--for that unique narrational condition of discourse "called auto-bio- graphic "; see "The Autobiographical Interruption: About Stendhal's Life of Henry Brulard, " MLN 93 (1978): 599-600. This essay appears as part of the untranslated book La voix excommuniée: essais de mémoire (Paris: Galilée, 1981). By "interruption" in the English title of this material, Marin intends two related meanings of "syncopation" at the lexical level: "excision of a syllable from a word" and "liaison" in the musical sense, "the end of a note in one section . . . heard at the same time as the beginning of a note belonging to the opposite section" (p. 614). Exploring the insistent refrain of certain dates in the Stendhal text, for instance, Marin reads the slippage between " deux cent cinquante " and " cinquante " as one of the revealing "rhythmic movements woven with phonic signifiers in the written text," movements by which, presumably, the excommunicated "saying" returns to haunt the autobiographical memorial of the "said." The interplay between phonic and graphic signifiers is even more apparent in another section of La voix excommuniée previously appearing in English translation, "The 'I' as Autobiographical Eye: Reading Notes on a Few Pages of Stendhal's Life of Henry Brulard " ( October 9 [Summer 1979]: 65-79), where a vertical list of women's names by which the memoirist charts the arc of his desire, when later rewritten horizontally and by initials only, presents the reader with a vexing but fertile instance of the unreadable. What we encounter are "pure graphemes, mere signifiers without signifieds" (p. 75) that can only be spelled, not apprehended as words. The "two directions" of attention posed by the text at this point, in its contradictory pull between the strictly visual and the actively verbal, tend to emphasize by default the inherence of the phonic within the graphic manifestation of the latter. According to Marin, the reader resists the sheerly pictographic response, even in the vicinity of an actual drawing inserted into the text, and attempts rather to "stammer" (p. 75) out something like a continuous verbal processing: "To take now the other direction, that of continue

reading, what we read, or better what we hear when trying to read the letters as a word or as a name, what is uttered in the mumbling in which, suddenly and for a moment, our reading is deconstructed, is the mother's name stuttered by a child's voice" (p. 76). In the letters "v A A M A A A M c g A" (as capitalized by Marin), that is, an enunciation of a 's and m 's tends to produce a phonetic anagram as well as a babbling pronunciation of that "auto-biophonic" sign at the origin of all life-writing. Though Marin does not mention Kristeva, we may find in the Stendhal text, under Marin's voicing eye, a genuinely strategic emergence of the maternal semiotic in syncopation with the symbolic order of script.

Marin's approach comes even closer to the procedures of this book when he pursues moments of lexical fissure and suture within ordinary vocabular, as well as alphabetical and numerical, sequencing. Working from a cryptic and abbreviated inscription that Stendhal recalls writing on the inside of his belt--"J. vais voirla 5." for "je vais avoir la cinquantaine" ( La voix, p. 119)--Marin submits it to a detailed linguistic analysis concerned with the syllabic "amputation" and "disappearance" (p. 124) it involves, as well as to a generalizing analysis about the "secret" recoding of script by voice in the reading of autobiography. Here the Stendhal text itself lends occasion to the most minute lexical and syntactic instances of that "syncopation" (the technical syncope, or dropped letter, as well as "liaison") and "reprise" which provide the master-tropes of Marin's theoretical introduction (pp. 21-22). The latter term is cited from Littré's dictionary as, among other more common acceptations, the "action of mending or darning torn or cut material" in which "the needle uses the thread not yet sewn by the first stitch . . . so that the eye does not perceive the joining of the thread " ("The 'I' as Autobiographical Eye," p. 66; Marin's emphasis). This invisible seaming of invisible rendings, this tear and repair in the weave of the graphic text, is the work of the voice otherwise excommuniée . This approach through "what we hear when trying to read" is to be even more intensively deployed later in La voix excommuniée for a listening "deconstructed" to the point of eliding the differentials of gender itself, in ways which resemble the logic of "liaison" in the verbal experiments of Virginia Woolf (see Chapter 7, n.8).

34. Lucette Finas, Le bruit d'Iris (Paris: Flammarion, 1978), p. 190. As does Donoghue, I will give my own brief selections from Finas in translation.

35. See below, Chapter 7, under the subheading "Catching That D/rift."

36. Christian Prigent, "Reading in Every State," Enclitic 6, no. 2 (Fall 1982): 44-49.

37. Roland Barthes, "Question de tempo," Preface to Le bruit d'Iris, pp. 7-11. English excerpts are from the translation by Annwyl Williams, "A Question of Tempo," Oxford Literary Review 5, nos. 1-2 (1982): 150-53.

38. Christopher Norris, in Derrida (Cambridge, Mass.: Harvard University Press, 1987), understands supplementation in the readily available sense, for instance, of a so-called supplement to an encyclopedia or dictionary continue

(Norris's example is the OED, p. 110). Named thereby is that extra yet necessary extension of the thing without which it would have no claim on the encyclopedic in the first place: the additive that definitively completes.

39. Roland Barthes, The Pleasure of the Text, trans. Richard Miller (New York: Hill and Wang, 1975).

40. William Empson, Seven Types of Ambiguity (New York: New Directions, 1947), p. 45.

41. See below, Chapter 6, under the subheading "Phonemanology."