Author: Scott, Clive
Date published: January 1, 2009
Voice and metrical analysis
If one takes a sample of performances of Henri Duparc's setting of Baudelaire's "L'Invitation au voyage", one may be surprised by the variety of renderings of the second line of the refrain:
Luxe, calme et volupté.
This heptasyllable allows "standard" readings of (i) 1+2+4, (ii) 2'+l+4 (where the apostrophe indicates a coupe lyrique), or possibly (iii) 3+4, if one feels that 'Volupté" is not the third element in an expanding itemisation, as in (i), but a consequence of the fusion of the first two nouns. Bearing in mind that Duparc's score does not allow for version (iii), we note that version (ii) is opted for by ElIy Ameling (San Francisco Symphony Orchestra, Edo de Waart, 1981) and by Felicity Lott (with Graham Johnson, 1987). More surprisingly, the coupe lyrique at "Luxe/" is repeated at "calme/" by Charles Panzéra ("an orchestra", Piero Coppola, 1933), by Victoria de los Angeles (with Gerald Moore, 1 957) and by Janet Baker (London Symphony Orchestra, André Previn, 1977), producing the octosyllable 2'+2'+4, the hiatus at "calme/ et" buffered by a pause. And just as surprisingly, in his 1934 version, with Madeleine Panzéra-Baillot at the piano, Panzéra sings the hexasyllable 1+1+4. These findings take us to the heart of a debate between text and voice, the linguistic and the paralinguistic, which has never been brought to a satisfactory conclusion, since however piously the necessity of their cohabitation is canvassed, they seem mutually to exclude each other.
In his survey of metrical analysis, Benoît de Cornulier identifies some twenty possible scansional methods, all of which fall within "conventional" approaches to the line of verse.1 The crucial questions which attach to the practice of metrical analysis are these: How does one decide what is metrical (metrically pertinent) and what is not? How determining is the metre identified in any particular poem - for example, an alexandrine metrically has 12 syllables, but might the reader, in realising any particular line, exercise some freedom in the treatment of mute e\ synaeresis and diaeresis, and finish up with a line somewhere between, say, 10 and 14 syllables?2 Is metrical analysis concerned with the pre-interpretative or the post-interpretative, that is, is metre inscribed in writing in a latent or an actualised state? At what point does metrical possibility become so multiple that metre can no longer be meaningfully said to exist?
From a metrical point of view, the voice is a distraction, because the paralinguistic merely masks the linguistic, and metre properly belongs to the linguistic being of the text. The only admissible voice, therefore, is the ideal voice, the voice which does maximal justice to the linguistic givens of the text. This view is justified by the proposition that metre, as a deep structure, predates rhythm, which belongs to voice, to sequence and to paralinguistic ornamentation, to surface structure.
These two arguments are clearly inadequate. The notion of the ideal voice is in fact a process of devocalisation. The ideal voice is the voice of the International Phonetic Alphabet, of isolated purified sounds existing in a milieu without an acoustic, without resonance and without a body. Analysts convert the absence of voice in the text into "verbal music", as if language without a voice had a describable acousticity; but only the voice can make language sound. The second argument might be countered with the suggestion that metre is a rationalised spatialisation of something more primitive, that metre 's function is not to permit rhythmic variation as a useful supplement, but to preserve and synthesise the multiple rhythmic pulsions which underlie it. Metre is then a matrix of the readings it makes possible. The real question is not how to read off the written, to recite the written, but how to re-introduce the voice into the written, how to re-oralise the written. If it is the business of verse to maximise both the materiality of language and the capacity of words to act as shifters, to lend themselves to appropriation, then what matters is the psycho-physiological imprint one finds in language and the way in which language carries not a particular voice, but vocality, vocal potentiality.
The two voices
But if we manage to reintroduce voice into the assessment of poetry's metricorhythmic being, we still need to ask "which voice?" In the face of our inability to incorporate into critical discourse the physiology of particular voices, the organic accidents of an individual's vocal apparatus,3 we have focussed our attention on shared systems of voice-use (for communicative and expressive purposes), subsumed within paralinguistics. But this ought to be expressed more strongly: paralinguistic features of the voice (speed, loudness, tone, intonation, enunciation, etc.) have two feces, one turned towards the innate physiology of the voice, its timbre, the other towards expressive resourcefulness, towards strategic uses of voice. In criticism, we have ceased to acknowledge the former face (or nearly); put another way, the latter fece exists at the expense of the former; put another way, voice is superseded by its own powers of expression.
But there are writers in whose work the physiological voice, and more particularly the psycho-physiological voice, has been given its critical credit. The two faces of voice might, for example, be related to Julia Kristeva's distinction between the geno-text, that pre-systemic process which includes psychic drives finding their way to utterance, and the pheno-text, the text which has achieved system, which obeys the rules of communication, which presupposes a subject of enunciation and an addressee.4 And if the geno-text is the semiotic in the symbolic, then we might ascribe to the voice the power to activate, or dredge up, those semiotic presences: "Si on peut l'imaginer dans le cri, les vocalises ou les gestes de l'enfant, le sémiotique fonctionne en fait dans le discours adulte comme rythme, prosodie, jeu de mots, non-sens du sens, rire".5
Kristeva's geno/pheno distinction is borrowed by Roland Barthes as "génochant" and "phéno-chant" in his differentiation of the voices of Charles Panzéra and Dietrich Fischer-Dieskau, in his 1972 essay "Le Grain de la voix", with its 1977 sequel "La Musique, la voix, la langue". Panzéra 's cultivation of the grain of the voice means that his voice emerges directly from his body, "du fond des cavernes, des muscles, des muqueuses, des cartilages",6 is principally located in the throat and is concerned with the diction, the pronunciation of the language. Pronunciation is about the phonetic production of language, about acoustic quality in the voice. Barthes sums up: "Le 'grain', ce serait cela: la matérialité du corps parlant sa langue maternelle: peut-être la lettre; presque sûrement la signifiance".7 FischerDieskau's singing, on the other hand, concentrates on expressivity, subjectivity; it privileges the lungs, breathing, and is concerned with articulation, that is to say with clarity of projected meaning, with the coded ways of communicating feeling, with the rhetoric of the voice. The grain of the voice, the geno-text, is a play of signifiers where meaning is always in the process of being made, unmade, re-made; it is "un jeu signifiant" - of ever-emergent, but no more than emergent, articulacy - "étranger à la communication, à la représentation (des sentiments), à l'expression".8 The expressive voice, on the other hand, is phenotextual inasmuch as it is achieved signification, can represent a signified, and inasmuch as its resources and effects have been systemized, in the service of communicative and expressive efficiency (Fischer-Dieskau's singing promotes "une culture moyenne").
In the paragraphs which follow, I do not wish to plunge very far into the psycho-phonetic concerns of analysts such as Kristeva and Ivan Fónagy.9 I want to indicate the way in which Baudelaire's "Causerie" invites the voice into his poetry and generates an even-handed negotiation between the physiological voice and the communicative voice, between pronunciation and articulation, between timbre and delivery, and to consider how this particular negotiation might be captured and expanded upon in a translation.
"Causerie" presents itself as an apt candidate for translation, partly because of its relative neglect, critically and translationally, but more especially because it promises relaxed and intimate discourse within a structure - the sonnet - which implies a high degree of management and lyrical rhetoricity. The facts are quickly rehearsed: this sonnet, on seven rhymes (a maximum for the form) is part of the Marie Daubrun cycle; it brings a curious mixture of tutoiement and vouvoiement, which most commentators agree, not surprisingly, is to do with changes of emotional and psychological proximity:10
Vous êtes un beau ciel d'automne, clair et rosé! 2+4+2+4/2+4+3 '+3
Mais la tristesse en moi monte comme la mer, 4+2+1+5
Et laisse, en refluant, sur ma lèvre morose 2+4+3+3/3 '+4+3+3!
Le souvenir cuisant de son limon amer. 4+2+4+2
- Ta main se glisse en vain sur mon sein qui se pâme; 2+4+3+3/4+2+3+3
Ce qu'elle cherche, amie, est un lieu saccagé 4+2+3+3/5'+2+3+3!
Par la griffe et la dent féroce de la femme. 3+3+2+4
Ne cherchez plus mon coeur; les bêtes l'ont mangé. 4+2+2+4
Mon coeur est un palais flétri par la cohue; 2+4+2+4
On s'y soûle, on s'y tue, on s'y prend aux cheveux! 3+3+3+3/4'+3+3+3!
- Un parfum nage autour de votre gorge nue! . . . . 3+3+4+2/4+2+4+2
Ô Beauté, dur fléau des âmes, tu le veux! 3+3+2+4/3+3+3 '+3
Avec tes yeux de feu, brillants comme des fêtes, 4+2+2+4
Calcine ces lambeaux qu'ont épargnés les bêtes! 2+4+4+2
The vocal prosody of "Causerie "
What the voice does as it finds its way through this text is at once to shape respiration, pausing, pitch, and so on, but without stability, as vocal experimentation, as a testing of parameters, as a path not towards optimal expressiveness, but towards one's own expression, tuned to the sound of one's voice. At the same time, as the voice gains articulation, takes possession of the text in a variety of guises, so it also yields articulation to pronunciation. By that I mean quite simply that the voice releases and scatters the acoustics of the text, or, contrarily, concentrates it in moments of repetitious insistence; and this is a process which, by the very saying of the text, cannot be prevented: the voice of the reader accedes, as it were, to the extraction by the text of his/her own orality. I do not lend my voice to this text, as a polite visitor, as might anyone else, in order that its textuality may be actualised. I use this text as an instrument to perform my own voice, in a way that releases in it not only its known physiological features, but also its unknown psycho-physiological capabilities.
My analysis of this poem will be unsystematic and selective, but it will give some sense, I hope, of the concerns that I feel a translation should address. The first line identifies a recurrent problem: the case of the coupe lyrique. Does one say 2+4+2+4 and give voice to two matching movements of syllabic expansion, the one confirming the other, which shared phonemes only serve to endorse? Such a move helps me pick out, too, the worm already in the rose: the line's central "hemistich" - "-/es un beau ciel d'auto/mne" - establishes the imminence of the other 4+2 's, in which an adjective, or noun complement, has a peremptory and damning effect, or in which the paired measures dissociate themselves from each other:
Le souvenir cuisant de son limon amer
Ce qu'elle cherche, amie,
Ne cherchez plus mon coeur
de votre gorge nue (! !)
Avec tes yeux de feu
There are other senses in which this line is the beginning of the poem's undoing, /e / will become the note of despair and aggression that undermines "êtes" and "clair" ("tristesse", "mer", "lèvre", "amer", "chercher", "bêtes"). And thé /o/ of "beau" and "rose" finds very different values in the final stanza ("Beauté", "fléau", "lambeaux"). These diverted trajectories of phonemes look like part of a rhetorical strategy - "fléau" and "lambeaux", for instance, appear at the caesura - an ironisation of the text, a campaign for semantic unreliability. But we shall have reason to revise this judgement.
Similarly, the 2+4+2+4 reading of the first line places it squarely within a certain scansional tradition designed to bring out its resonant orderedness, to bend the voice to its imperious repetition, to produce the pleasure of aesthetic fulfilment. But if we read it with a coupe lyrique (2+4+3 '+3), we introduce unstructuring consequences of a purely vocal decision, where the e serves not syllabicity or rhythmic fluency, but the psychic in the phonetic: the "choc funèbre" of "automne", which "clair et rose" cannot mask, because they themselves are dissociated by it, isolated in a purely decorative self-sufficiency.
This initial possibility of a coupe lyrique triggers others, even more disruptive, because not only do they likewise grow from a punctuational pretext (a moment of the voice recorded in real time, as it were), but they also and at the same time entail a hiatus, which breaks the bounds of the alexandrine syllabically (if this still matters) and produce the laryngal constriction, the frustration of vocal impulse, associated with glottalisation:
2. Et laisse, en refluant, sur ma lèvre morose 2+4+3+3 BUT 3 '+4+3+3 !
6. Ce qu'elle cherche, amie, est un lieu saccagé 4+2+3+3 BUT5'+2+3+3!
10. On s'y soûle, on s'y tue, on s'y prend aux cheveux 3+3+3+3 BUT4'+3+3+3!
To these we might add the possible coupe lyrique in line 12, more muted, of the same kind as that in line 1 :
12. Ô Beauté, dur fléau des âmes, tu le veux 3+3+2+4 BUT 3+3+3 '+3
Thèse instances are made differently significant by the poem's pattern of articulated and elided word-terminal, non-clitic e\ traced in the following tabulation:
2. e //e e
6. e vil
7. e// e
8. // e
12. // e
[first hemistich = 10 (3 + ?); second hemistich =10(10)]
Remarkably, none of the second-hemistich e's is elided. I have argued elsewhere that the alexandrine, as a "vers composé" (created by two hexasyllables) not only encourages us to read rhythm in terms of paired measures, but to read each hemistich as a particular kind of vocal/expressive space.11 One might feel, looking through the unelided e's of the second hemistichs that Baudelaire was encouraging the voice to indulge itself- in the wistful, the plaintive, the melodramatic - and to "platform" simile. These kinds of tonal lyricism are, on the whole, suppressed in the first hemistichs, where a more matter-of-fact voice delivers unadorned statements. The poem is, in fact, a dialogue between a "vous" and a "tu" on another axis: the "vous" perspective tends to occupy the first hemistich, the "tu" perspective the second. The effect of the re-installation of e at lines 3, 6 and 10 acts not as the reintroduction of lyrical protraction, but as the notation of an emotional/experiential impasse, an exhaustion of possibility.
If the coupe lyrique tends to undermine the rhetorical integrity of the alexandrine, and its confidence as representative utterance, if it pushes the voice away from impersonation and towards self-intrusion, so equally phonemic patterns can push one away from the tracing of patterns in carpets towards unpredictable and irregular acts of pronunciation, confrontations with nagging psychosomatic pulsions and self-generating associative networks. We have argued that /e / is a sound which, in the design of the poem, goes sour: the "êtes'Of "vous êtes" becomes the "-êtes" of "bêtes". We might argue that the same is true of /m/, the /m/ of self and mother, but also the /m/ of "femme" and "amie" and "limon amer" and "mange" and so on. At the same time, like /e/, the sound flits about the text as capriciously and erratically as a butterfly, so that its sound-symbolic value is never definitively established and can transfer itself to the reader-speaker's own existential and psychic preoccupations. Fónagy describes /m/ as the "normalisation linguistique du mouvement de succion des lèvres"12 (new light on "ma lèvre morose"?) and quotes Ostwald, who notes among his schizophrenic patients "une corrélation entre l'émission répétée de MM (humming) et un repliement sur soi-même, une régression narcissique profonde (qui les ramènent à l'unité duelle primitive de mère/enfant [...]".13
One might justifiably propose that there is nothing Baudelaire-induced about this shift from phoneme as element of textual design to phoneme as vocal eruption, as pronunciation, as timbre of reader/speaker. My argument is that there are signs of such textual engineering. What is curious about line 5, for example, is the sudden short burst of /e/: "main" > "vain" > "sein". This phoneme occurs nowhere else in the text. Of course, it gives the line a peculiar acoustic density (along with /m/, /s/ and /i/), but this very density, so arbitrary in its imperious eruption, takes us to other places of fixation and psychic paralysis. Similarly, lines 12-13 play with /0/ ("veux", "yeux", "feu"), adumbrated in the "cheveux" of line 10, and the "lieu" of line 6. But why? Again, despite the adumbrations, this relative concentration of the phoneme in a small space seems like a response to some pulsional imperative. Of course, we can make sense of it, by casting it as the antagonist, and destroyer, of /ce/ ("coeur"): both these vowels are front and rounded, but one is half-closed/high-mid and the other is half-open/low-mid. But we are still left with a significant margin of free-associative gratuitousness, of pure, unassigned vocality, with a sound which is expressive (of self), but still not quite the servant of articulacy. One might want to make something of the same case about the punctuation, and particularly about the functions of the dash and suspension points.
Any idle reading of Les Fleurs du Mal might identify the dash with three functions: (i) to change the direction of address (here the dash may indeed be accompanied by inverted commas); (ii) to set up a closing apophthegm or longer endgame; (iii) more rarely, to introduce a brief, interpolated response, which the poet is as if powerless to suppress. In "Causerie", the two dashes look designed to alert us to a change of address. The difficulty is that we cannot see the extent of their applicability. Do we divide the poem, say:
11. \-4 Interior monologue ("vous" perspective)
11. 5-7 Direct address ("tu" perspective)
11. 8-10 Interior monologue ("vous")
I. 11 Direct address ("vous")
II. 12-14 Interior monologue ("vous")
I. 1 Voice I (direct address)
II. 2-4 Voice II (soliloquy, aside)
11. 5-7 Voice III (direct address or solil.)
I. 8 Voice I (direct address or solil.)
II. 9-10 Voice II (solil., aside)
I. 11 Voice I (direct address)
II. 12-14 Voice III (direct address or solil.)
or perhaps... All these attributions are highly uncertain, and it is impossible to tell consistently where the boundaries fall, whether at the next full-stop, the next stanza-end, or the next dash (which seems to be the case, for example in "Les Petites Vieilles" and "Un voyage à Cythère"). Baudelaire's own lack of clarity is perhaps testified to by the 1857 edition which, like its proofs, has a dash at the end of line 7. This makes more sense of the shift to the second person plural in line 8, but leaves the dash at the beginning of line 1 1 a puzzle. In similar fashion, Baudelaire requires a dash after "nue" (line 11) in the first proof of the 1857 version. This further indication of line 1 Ts isolation is hardly necessary and is ultimately replaced by suspension points (see below). But the exact "distribution" of discourse may not matter. What matters is the way in which dash signals the labyrinth of voice, the shifts in vocal proximity, modality, and the unpredictability of the adjustments of level and threshold (genotext and phenotext, inner and outer voices, degrees of depth in the psyche) - the dash is the sign of the abrupt, the reflex, the peremptory, the disruptive, the impatient.
As far as the introduction of suspension points at the end of line 1 1 is concerned, we should first note Russell Goulbourne's general observations: "They [suspension points] constitute a space, both visually and audibly, both literally and metaphorically, in which the reader is brought into play, invited to think, to use his/her imagination, to hear the unsaid, to read between the lines, or rather, to read between the points."14 Goulbourne's words seem strikingly just, as long as the reader does not merely look to recuperate a suppressed message. Suspension points like these are the point at which the text falls into the hands of the reader; this space is explicitly the textual imaginary, the poem's blind field, endowed with greater or lesser intensity (Baudelaire uses, variably, between three and five suspension points). I say "suspension points like these" because a check against variants tells us that the suspension points in "Harmonie du soir" (5) and "Ciel brouillé" (5) are similarly the replacement of a dash. In other instances - for example, "Les Petites Vieilles", "Le Vin de l'assassin", "Un voyage à Cythère" - suspension points make their first appearance in the 1861 edition, as additions to text. This albeit undramatic gravitation towards suspension points seems to me to be motivated by the realisation that, in regular verse, suspension points can go beyond the linguistic (sign of elision) and the paralinguistic (suspended pitch, voice fade, drawl, pause, lightening of accentuation) towards vocal evaporation.15 Suspension points return the voice to the pre-verbal, to a state of availability to the self's vocal pulsions, and perhaps indeed to the pre-acoustic source of all being ("le langage des fleurs et des choses muettes" - "Élévation").
The voices of "Causerie" translated
The overall objective of this translation, therefore, is to transform the delivery-centred into something which invites timbre, to let the practised infiltrate the represented. This entails the sonnet's transformation into a free-verse form, which preserves the sonnet's stanzaic construction but which encourages the paralinguistic to invade the linguistic in the form of pausing, phrasing, etc. We also need to confront the task of turning the rhetorical - a code of figured speech, where figure supplants prosodie features - into the spoken, without having to descend to the colloquial.
My heart's a
sacked by the
mob, Drink sod
den, they're at
throats, and hair
scourge of souls" -
that's how you want it!
fiery eyes, brightly lit
as if for
what's left of me
after the dogs
The first stanza attempts to establish the bearing principle of the whole translation: to free the voice as a projector of the text, setting the text in motion and at the same time being revealed by the text's movement. The juxtapositional strategies of the spatial text are thus supplanted by the metamorphic ones of the temporal text. But this textual temporality is not the quantitative, homogeneous and discontinuous temporality of metre, but, in the fashion of Bergsonian durée, qualitative, heterogeneous and continuous. Accordingly, I have created a line which, by filling the page's width and using uneven spacing and a promiscuous typography, frustrates any attempt to encompass it as a unit and requires the reader to inhabit its heterogeneous temporalities, loudnesses, acoustic associations. Instead of the time of the line being determined by the pressures of "external" time (the surrounding space), it draws time into itself as the necessary medium of its own experience. Each reader/translator is thus at liberty to re-arrange layout and typography in order to register the durée and peculiar eventfulness of their own reading.
Part of the strategy for decomposing the line as written, as rhetorically purposive, and returning it to the idiosyncrasies and pulsional turbulence of voice, is, as we have seen, the institution of a strong translinear, or intralinear, acousticity, which acts as dispersively, or disruptively, as it does integratively. While we may allow that acousticity orchestrates our hearing of the text, we must equally allow that we hear it, not as part of a formal design, but as the uncontrollable spray of the text, which the voice is attracted to/distracted by, in its passage through the text. This vision of acousticity is the compromise between the written and the spoken, the text read off (voice as instrument, which validates, endorses the text) and the text as stimulus or liberation of the oral (voice as that organ which explores its psycho-physiology, through the text).
In my version of this stanza, therefore, I have used the closing noun phrase "bitter silt" as a rough target for sounds within the stanza, picked out by capitalisation. But these are competed with by other sounds whose purposes are not so clear, whose sudden projections are more darkly motivated. I have used italics and bold to rather the same effect: to create a text which is, above all, temperamental, or is made temperamental by orality itself. The task of the translator is not to promote in the target text a recognition of what is already there in the source text, but to use the target text to create the possibility of discoveries not yet registered.
And I have resorted to an alternative system of punctuation - slashes, brackets, "misplaced" hyphens, preposed exclamation marks, suspension points within the word - in order to suggest that, in the "new order", punctuation is not a servant of periodicity, the external classifier and guardian of the conventions of written utterance, but, on the contrary, an asylum seeker from the written, withdrawing to the heart of vocal continuity, with which it has no wish to interfere, and where it can reinvent itself as the agent of vocal perception.
Finally, another, related objective of this translation is to dynamise and heterogenise the metrico-rhythmic current. IfI describe the final line of this quatrain - "The biting after- taste of bitter silt" - as iambic pentameter, I am implying that it is made up of five rhythmically identical, isochronous units. This quantitative vision of iambic pentameter prejudices the reading voice and prevents our seeing the line as an accumulation of the heterogeneous, of elements all with different qualitative values, to wit:
amphibrach (x / x) + amphimacer (/ ? /) + weak monosyllable + trochee (/ x) + strong monosyllable
The voice melts down iambic pentameter in order to find in it its primary elements, these fragments of a moi profond which does not yet know what shapes it wishes to take, because it is a set of vocal/verbal reflexes.
The second stanza again maintains its "quatrainness", but it is designed principally to give a sense of the tetrametric movement characteristic of the alexandrine. There is no attempt here to limit the line to twelve syllables, nor to rhyme (what would be the point?). What is important is to provide a sense of a particular kind of rhythmic movement, in which each line has four measures whose lengths vary and solicit the voice in different ways: the slackening of tension, the diminution of concern in the longer measures (of five and six syllables); the moments of attentive poise in the amphimacers ("smooth across", "laid to waste", "red in tooth"); the suppressed bitterness or despair in the single iambs (x /) ("amie", "and claw", "my heart's", "long since"). In other words, my version does not rely on rhetoricized patterns of paired measures, but looks to allow each measure to create its own intra- and interlinear affinities and conflicts.
One other feature deserves comment. I have introduced further punctuation of a vocal kind: my colons in the first line are intended to express sensory breaks, those mixtures of the withheld, or momentarily suspended, and the about-to-begiven of erotic experience; these are impatiently interrupted by the dash which throws it all away. This movement is then imitated at a lower pitch in the line following. The suspension points in the third line mark the voice's resuscitation of a dead metaphor, an emotional reinvestment of the verbally automatic with the verbally intended, the recovery of a relish in uttering it. The colon of the final line is a last pale echo of the earlier ones: the note of stilled anticipation, arrested urgency, is now as if a memory, or something so abstract that the physicality of voice mirrors no other physicality.
The columnar presentation of the first tercet presents a moment of vocal free fall. This is the point in the poem where the addressee is particularly difficult to identify: there are no indications of a second person, until, of course, the final, separated line. Are lines 9-10 spokenà la cantonade, to the reader, to himself...? The layout is designed to convey a no man's land of utterance where the reader has no purchase on phrasing and finds it difficult to anticipate text with the eye. Reading becomes a series of sudden encounters with words or word-fragments. Significantly, this verse-column is metrically strict, at least insofar as each line, bar the last, is trisyllabic; and it is this metrical strictness which enforces the enjambements, so difficult to negotiate: metre and syntax are no longer mutually endorsing, but operate at cross purposes, at each other's cost. When the final line of the source-tercet arrives, it does not entirely escape the trisyllabic Procrustean bed, but it lies more easily than the previous lines, with less disruption of its phrasality, and its final word breaks into a space of its own: the metrical spell is broken.
The final tercet brings the translation further into English territory,16 after the gallicized performance of the second quatrain. At the same time, it leaves the terrain vague of the mid-page and returns to a left-hand margin, where speech can once again improvise itself, develop its junctures, its tonal innuendoes, its variations of attack. True, each line of the source text is again broken down into four segments, but there is no longer the "lateral" music of a "vers composé", with medial caesura. Instead we have English rhythm developed as a modulating vocal trace. By this latter I mean: free verse allows English rhythms to re-establish contact between stress intensity, syllabic duration and pitch (metre has no interest in either the nature of stress, or in degree of stress). It also makes audible the unfolding of rhythmic variation as Heraclitean, metamorphic experience rather than as a sequence of adjacent similarities and differences. This series of lines begins with a trochee, but with the stressed vowel not insisted upon. "Unyielding" provides the trochee with a "prefix" and an extended stressed syllable. Thus, paradoxically, the negative "un-" is swallowed and "yielding" begins to acquire an optative flavour. "Scourge of souls" reverses the previous line's amphibrach (x / x) into an amphimacer (/ ? /), with more length in "scourge" than in "souls". The fourth line might be read with a flattening iambic rhythm, rather matter-of-fact: x / x / x. I prefer the adonic configuration - / ? ? / x - because it makes better sense of the exclamation mark: of the poet's simultaneous realization and acceptance of the pointlessness of further parleying. After the trochee and amphimacer of lines 5-6, the rhythm loses its focus and the weak syllables become more sapping. The second paeons (x / x x) of lines 9-10 (with a one-syllable upbeat-"To" - in line 9) eloquently capture the difficulty of generating stress and the ease with which the impetus falls off. The choriamb (/xx/) of line 1 1 picks up the choriamb in the adonic of line 4, and the final iamb is like a reversal, gradually and variously engineered, of this tercet's initial trochee.
Suffice it to say, by way of conclusion, that translation has concerned itself remarkably little with voice quality, as if poetic texts were not made peculiarly complex by paralinguistic potentialities, by the voice's construction of the text. It is assumed that the translator translates his/her reading of the text, where "reading" means "interpretation" and where the text is left in its writtenness, however much its "music" may be referred to. But the music of the unspoken written is a far cry from the vocality of the spoken written; and reading, the psycho -physiological inhabitation of the text, is a far cry from any interpretation which might be read out of a text. Translation, it might be claimed, has for too long had its eye on the wrong target.
University of East Anglia
1 Benoît de Cornulier, Théorie du vers: Rimbaud, Verlaine, Mallarmé (Paris: Seuil, 1982), pp. 121-131.
2 Cornulier himself concedes: "Le comptage métrique des syllables est largement mental ('abstrait', comme aimeraient dire certains). Que m'importe qu'on prononce à peu près G? consonne ou qu'on ne prononce pas Ve dans Ariane, du moment queje reconnais intérieurement la possibilité d'interprétation 6-syllabique de Ariane, ma soeur; [...]" (p. 130, n. 1).
3 No doubt part of the reason for this lack of attention to individual vocal characteristics is the difficulty of describing them: in his The Phonetic Description of Voice Quality (Cambridge: Cambridge University Press, 1980), John Laver confesses that "we know now only a little more about the factors that give rise to different qualities of voice than Quintilian did" (p. 1).
4 Julia Kristeva, La Révolution du langage poétique: L' Avant-garde à la fin du XDP siècle: Lautréamont et Mallarmé (Paris: Seuil, 1974), pp. 83-86.
5 Julia Kristeva, Polylogue (Paris: Seuil, 1977), p. 14.
6 Roland Barthes, L'Obvie et l'obtus: Essais critiques ///(Paris: Seuil, 1982), p. 238.
7 Barthes, p. 238.
9 Ivan Fónagy, La Vive Voix: Essais de psycho-phonétique (Paris: Payot, 1991).
10 See, for example, Claude Pichois (éd.), Baudelaire: OEuvres complètes I (Pans: Gallimard, 1975), p. 933.
11 Clive Scott, The Poetics of French Verse: Studies in Reading (Oxford: Clarendon Press, 1998), pp. 70-72.
12 Fónagy, p. 76.
13 Fónagy, p. 77.
14 Russell Goulbourne, "The Sound of Silence... points de suspension in Baudelaire's Les Fleurs du ?aG, Australian Journal of French Studies, 36 (1999), 200-213 (p. 213).
15 This last phrase owes much to loulia-Marina Sarantou's illuminating "The Punctuational Language of Dandyism in Baudelaire's 'Épigraphe pour un livre condamné'", Forum for Modern Language Studies ,40 (2004), 14-26 and is a vocal "vaporisation [...] du Moi" (Pichois [ed.], p. 676). For the erotics of Baudelairean punctuation, see Sarantou, "The Erodynamics of Punctuation in Baudelaire's 'Les Bijoux'", Australian Journal of French Studies, 38 (2001), 213-227.
16 The removal of the apostrophe ("O Beauté") and the scare-quotes around "unyielding scourge of souls" are both measures of dissociation from rhetorical grandiloquence and from the ceding of voice to ready-made formulation.
CLIVE SCOTT is Professor Emeritus of European Literature at the University of East Anglia. His principal research interests lie in French and comparative poetics (The Poetics of French Verse: Studies in Reading, 1998; Channel Crossings: French and English Poetry in Dialogue 1550-2000, 2002 [awarded the R.H. Capper Book Prize, 2004]); in literary translation, and in particular the translation of poetry (Translating Baudelaire, 2000; Translating Rimbaud's 'Illuminations', 2006); and in photography's relationship with writing (The Spoken Image: Photography and Language, 1999; Street Photography: From Atget to Cartier-Bresson, 2007). He is at present working on a book on the translation of Apollinaire's poetry. He was elected a Fellow of the British Academy in 1994.