Jump to ContentJump to Main Navigation
Vers LibreThe Emergence of Free Verse in France 1886-1914$

Clive Scott

Print publication date: 1990

Print ISBN-13: 9780198151593

Published to Oxford Scholarship Online: October 2011

DOI: 10.1093/acprof:oso/9780198151593.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 04 July 2020

Theories and Enquiries

Theories and Enquiries

Chapter:
(p.120) 3 Theories and Enquiries
Source:
Vers Libre
Author(s):

Clive Scott

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780198151593.003.0004

Abstract and Keywords

This chapter discusses the theoretical foundations of free verse. A discussion of the theoretical works of Gustave Kahn's ‘Preface sur le vers libre’, Albert Mockel's Propos de literature and Edouard Dijardin's Les Premiers Poetes du vers libre is included in this chapter. The chapter also includes a presentation of the findings of two literary enquiries made by Huret's Enquete sur l'evolution litteraire and F. T. Marinetti's Enquete international sur le vers libre. Included as well in this chapter are quotations of the text examined and analysis of the related poems to give the chapter a documentary value and to allow measurement of the difference between theory and practice.

Keywords:   theoretical foundations, free verse, Gustave Kahn, Albert Mockel, Edouard Dijardin, Huret, Marinetti

During the course of the foregoing chapters we have encountered three principal versions of free verse: (i) the free verse in which the stanza is a sequence of rhythmic measures combined (or not) in various ways to create a lineation or set of verse-spans, and in which the measures are still defined by their syllabicity; (2) the free verse which we have called ‘cadence accentual’ and in which the line continues to be the rhythmic unit, détermined by a consistently recurring number of measures and sense-accents; this kind of free verse is particularly associated with longer lines and the verset, and in it syllabicity plays only a role of accompaniment; (3) the free verse, yet to be explored, which is essentially a sequence of deviations from, and approximations to, regular verse lines and in which, consequently, the line again continues to be the basic rhythmic unit, but a line defined by number of syllables rather than by number of measures. (1) is the free verse of, among others, Alain-Fournier; (2) owes something to the Whitman translations and is the verse of Maeterlinck, Claudel, and Cendrars; (3) is the free verse of Laforgue and Henri de Régnier. To (1) might be added the verse of Gustave Kahn, who distinguishes his approach from Laforgue's in the following terms:

Dans un affranchissement du vers, je cherchais une musique plus complexe, et Laforgue sʼinquiétait dʼun mode de donner la sensation même, la vérité plus stricte, plus lacée, sans chevilles aucunes, avec le plus dʼacuité possible et le plus dʼaccent personnel, comme parlé. Quoiquʼil y ait beaucoup de mélodie dans les complaintes, Laforgue, se souciant moins de musique (sauf pour évoquer quelque ancien refrain de la rue) négligeait de partipris lʼunite strophe, ce qui causa que beaucoup de ses poèmes parurent relever, avec des rythmes neufs à foison, et tant de beautes, de lʼécole qui tendait seulement à sensibiliser le vers, soit celle de Verlaine, Rimbaud et quelques poètes épris de questions de césure, doués dans la recherche dʼun vocabulaire rare et renouvelé. (1897: 17)

(p.121) These words are taken from Kahn's ‘Préface sur le vers libre’ which introduced his collected Premiers poèmes of 1897, and whatever we may think of his claims to have invented free verse— ‘je nʼentendais pas fournir, en créant les vers libres, un canon fixe de nouvelles strophes’ (1977/1902: 68)—we must allow him the privilege of having been its first theorist. For though the ‘Préface’ dates from 1897, it retains, as its central section (pp. 22–8), a reply, published in La Revue indépendante, 26 (Dec. 1888), to an article by Brunetière. Kahn was also to summarize his theoretical views in his answer to Jules Huret's Enquête sur lʼêvolution littéraire of 1891.

(i) Gustave Kahn

In tackling Kahn's theory, therefore, I would like to begin at the beginning, with his reply to Brunetière. After some general introductory comments on the inevitability of evolution in verse, Kahn taxes the Romantics with having missed the point’ in their own attempts to liberate verse. Instead of clutching at the arbitrary straw of enjambement to undermine the alexandrine, they should have paid closer attention to the intrinsic construction of the classical line; had they done so, they would have discovered that the formula for free verse was already contained within it. Kahn sets out to demonstrate his argument by providing an analysis and a scansion of the opening couplet of Racine's Athalie:

  • Oui, je viens dans son temple adorer lʼÉternel,
  • Je viens selon lʼusage antique et solennel.
Of these lines Kahn writes:

le premier vers se compose de deux vers de six pieds dont le premier est un vers blanc

  • Oui, je viens dans son temple
et dont lʼautre
  • adorer lʼÉternel
serait également blanc, si, par habitude, on nʼétait sûr de trouver la rime au vers suivant, cʼest-à-dire au quatrième des vers de six pieds groupés en un distique.

Donc à premier examen ce distique se compose de quatre vers de six pieds dont deux seulement riment. Si lʼon pousse plus loin lʼinvestigation on découvre que les vers sont ainsi scandés

(p.122)                       Theories and Enquiries

soit un premier vers composé de quatre éléments de trois pieds ternaires, et un second vers scandé 2,4,2,4.—Il est évident que tout grand poète ayant perçu dʼune façon plus ou moins théorique les conditions élémentaires du vers, Racine a empiriquement ou instinctivement appliqué les règies fondamentales et nécessaires de la poésie et que cʼest selon notre théorie que ses vers doivent se scander. La question de césure, chez les maîtres de la poésie classique, ne se pose même pas. (1897: 24–6)

At a single stroke, then, Kahn sweeps away the caesura, denies it any rhythmic, let alone metrical, significance. The caesura is a creation of the prosodic legislator (Boileau) and plays no part in the way the poet constructs his verse. Racine writes regular verse almost by accident, because his combinations of measures happen to conform to a set of metrical principles which are in essence alien to the driving rhythmic impulse of his verse. And this rhythmic impulse is housed in the individual measure, the ‘cellule organique et indépendante’ (p. 26) whose terminal boundary is marked by the coincidence of a juncture in meaning with a rhythmic juncture. This seems to say little more about the rhythmic measure than traditional scansion would want to say anyway. But it does have important consequences for the constitution of the line and the way in which it is built up. The line is no longer a formal terminus ad quem dictating, by the tyranny of its number, the fractions of itself which are available to the poet, but a precipitate produced by the interaction of measures freely combined.

Rhythm then is produced by the repeated or changing pulse of accent in its coincidence with the boundary of a sense-unit. This will sound less novel to our ears than to those of Kahn's contemporaries, since the thinking of verse-analysts in the late nineteenth century was only gradually shifting from the idea that French versification was fundamentally syllabic and non-accentual to the suspicion that French rhythms might indeed be accentually motivated. Be that as it may, Kahn's insistence on the measure's terminal limit being ‘un arrêt simultané du sens et du rythme’ meant that enjambement was ostracized and made clear how misguided the prosodic innovations of the Romantics were.

But if the overall number of syllables in the line no longer (p.123) provides the cohesive force which binds measures together, what does? This is where Kahn's concern with verse-music begins to justify itself, for it is acoustic kinships which make individual measures cohere together. In a passage already quoted (see above, p. 62, Kahn writes:

Pour assembler ces unités et leur donner la cohésion de façon quʼelles forment un vers il les faut apparenter. Les parentés sʼappellent allitérations, soit union de consonnes parentes ou assonances par des voyelles similaires. On obtient par assonances et alliterations des vers comme celui-ci:

Des mirages | de leur visage | garde | le lac | de mes yeux

(1897: 27)

Kahn says no more than this, but the implications of these remarks are considerable. We must begin to imagine a verse in which acousticity is an integral part of rhythmicity, in which all accents, and not just the accent on the rhyme-word, are phonic rather than tonic, in short, in which the accent's production of a certain sound is more important than a sound's production of an accent. Of all this we shall have more to say in the next section, but we should observe now that while patterns of sound are important in all verse, both in structural and expressive terms, such patterns are rarely rhythmically determining; they are rather reinforcements of, or alternative phonetic and semantic configurations in counterpoint to, rhythms established otherwise, in answer to metrical and syllabic requirements. Kahn's line offers us a verse in which patterns of assonance (supported or not by alliteration) are actually constitutive of rhythmic structure, are what articulates rhythmic measures.

Kahn's first manœuvre against the ascendancy of the line is complete. He has demonstrated to his own satisfaction that the individual, organic measure is the basic rhythmic unit of French verse, rather than the line. The measure is no longer justified by, and answerable to, a fixed rhythmic perspective (the line), but has a life of its own and rhythmic demands of its own. The line is merely the result of the combination of a certain number of measures. Kahn's next step was to create a direct relationship between the measure and the stanza, to show that this relationship was unmediated by the line, that the line had no part to play in the prediction of stanzaic formation. Ironically perhaps, because the (p.124) measure itself has recovered its rhythmic autonomy, so the line, as a nonce combination of measures, must itself enjoy an increased autonomy; but this is an autonomy, it should be re-emphasized, which is owed not to some abstract metrical completeness but to the self-sufficient nature of its measures. But how are lines so formed to be related to other lines?

Tandis que le vers classique ou romantique nʼexiste quʼà la condition dʼétre suivi dʼun second vers, ou dʼy correspondre à brève distance, ce vers pris comme exemple [i.e. ‘Des mirages de leur visage …’] possède son existence propre et interiéure. Comment lʼapparenter à dʼautres vers? par la construction logique de la strophe se constituant dʼaprès les mesures intérieures du vers qui dans cette strophe contient la pensée principale ou le point essentiel de la pensée. (p. 27)

Two presiding notions emerge from Kahn's vision of the free-verse stanza. The first is contained in the phrase ‘la construction logique de la strophe’. If, as Kahn affirms in his reply to Huret's enquiry (Huret 1982/1891: 322), the stanza represents one complete facet of the ‘idea’ explored by the poem as a whole—‘Quʼest-ce quʼune strophe? Cʼest le développement par une phrase en vers dʼun point complet de lʼidée’—then the process of lineation is a process whereby that facet is analysed, is itself faceted, according to the logic of its own internal construction. This tallies closely with what Vielé-Griffin has to say about the structure of the free-verse stanza (see quotation above, p. 46), and indeed Vielé-Griffin puts the creation of the ‘strophe analytique’ at the very centre of free verse's achievement (see Marinetti 1909: 34). Secondly, Kahn argues that the free-verse stanza is rhythmically organized around a principal line, or rather around the ‘mesures intérieures’ of a principal line, which carries the main burden of the thought. This line may be the first line of the stanza or some other: ‘La strophe est engendrée par son premier vers ou son vers le plus important en son évolution verbale’ (Huret 1982/1891: 323).

Kahn closes his reply to Brunetiére with two further observations: the new prosody will not only release new harmonic combinations hitherto neglected, but will allow each poet ‘dʼecrire son rhythme propre et individuel au lieu dʼendosser un uniforme taillé dʼavance et qui le réduit à nʼêtre que lʼélève de tel glorieux prédécesseur’ (1897: 28); the old prosody will still have a role as a ‘cas particulier’ within the new.

(p.125) We need not delay over Kahn's written answer to Huret in 1891 since it merely repeats, more briefly, the points made in the reply to Brunetière; we have already quoted its principal clarifications. Suffice it to add that Kahn rather goes back on what he has said about the continuing viability of traditional forms as ‘cas particuliers’ of the new: ‘il vaut mieux pratiquement éviter les mélanges; enguirlander de vers fibres des strophes dʼalexandrins, cʼest sʼexposer à parler bègue’ (Huret 1982/1891: 324).

Much of Kahn's filling–out of his reply to Brunetièe in the later ‘Prèface sur le vers libre’ (1897) is in fact a filling-out of historical context. He is able to mount a more sustained attack against Boileau and his Draconian rules which are ‘du pur arbitraire … a volonté dʼun critique gâté’(p. 12) and to trace free verse's debt to Baudelaire, Verlaine, Rimbaud, and Mallarmé. But there are three things that should be salvaged from Kahn's historical survey because they serve to underline points made in the previous chapter. Contrasting the new Symbolist poetry's musical orientation with the essentially pictorial inspiration of the Parnassians, Kahn acknowledges the influence of the music of Wagner (and of Beethoven and Schumann) on the discovery of ‘une forme poétique, à la fois plus fluide et prése’ (p. 9). He acknowledges, too, the part played by the Romantics in the undoing of the alexandrine, and particularly of the alexandrine's hemistichiality: ‘Ce vers est devenu le vers de douze syllabes, avec dix céres possibles, avec césure obligée, car il nʼy a pas de mots de douze syllabes: ce qui est presque lʼabolition de la césure’ (p. 10). And in view of what we have said about the importance of Banville in maintaining interest in vers libres classiques (see above, Chapter 2, section (iii)(b)), it is worth noting that Kahn constantly refers to him as an authority. Vers libres classiques, incidentally, provide Kahn with an objection to answer, namely that vers libre is nothing new but has long ago been successfully practised by La Fontaine and Molère. Kahn counters by asserting that vers libres classiques are really an instance of ‘le vers familier’, are non-heroic ‘petits vers’ expressly designed to meet the needs of ‘modes lérs’ (p. 36).

We have already mentioned one of the substantive prosodic innovations which Kahn first introduced in his ‘Préce sur le vers libre’, namely the treatment of the line-internal articulated ‘e’ as a ‘simple intervalle’, with the same extrametrical status as a (p.126) line-terminal ‘e’ (see above, p. 35). Although Kahn does not make it absolutely clear, it seems that this procedure applies only to unelided ‘e's occurring at the ends of measures immediately after the accentuated vowel, that is, to phrase-terminal unelided ‘e's:

Pour nous, qui considérons, non la finale rimée, mais les divers éléments assonancés allitérés qui constituent le vers, nous nʼavons aucune raison de ne pas le considérer comme final de chaque élément et de le scander alors, comme à la fin dʼun vers rélier. (pp. 30–1)

The argument here seems rather Byzantine: because in free verse the rhyme is not the only music that counts, because all accentuated syllables, and not just the rhyme-syllables, are phonic rather than tonic accents, so the unelided ‘e’ which occurs after any (phonic) accent should be treated in the same way as the ‘e’ which occurs after the rhyme in regular verse. The argument is odd because it argues from regular verse to free verse: because every measure-terminal accentuated syllable is potentially a rhyming syllable it should be treated like the rhyme-syllable in regular verse. But the underlying gist of Kahn's argument is clear: syllables themselves (including the e atone) are not important in free verse; only acoustically motivated syllables are. And from this general conviction derives his assault on rhyme proper, which for too long has been ‘le coup de cymbale à la fin du vers, trop prévu’ (p. 33). Free verse is not to be associated with regular verse so much as with poetic prose ‘rythmée et nombrée, avec une sorte de musique’ (p. 32), precisely because its music, its patterns of acoustic kinship, are more floating, more freely distributed. Rhyme, whether it be mere assonance, different modes of halfrhyme, or full rhyme, must become mobile, and as important line-internally as it is line-terminally.

Underpinning these changes in the treatment of rhyme and the e atone is Kahn's belief that verse must recover its essentially oral and aural nature from the written and the eye-read. And it is this belief that leads to his emphasizing the importance of the ‘accent dʼimpulsionʼ*, a term whose meaning is as elusive as it is wide-ranging:

Cet accent tonique quʼon pourrait relever dans les mots, en les laissant immobiles, soit en les citant à la file, en exemples, disparaît à la conversation, à la déclamation, ou mieux, il ne disparaît point, mais se modifie. Il y a done un accent général qui, dans la conversation ou la (p.127) déclamation, dirige toute une période, ou toute une strophe, y fixe la longueur des valeurs auditives, ainsi que les timbres des mots. Cet accent semblable chez tout le monde, en ce sens que chaque passion, chez tous, produit à peu près le même phénomène, accélération ou ralentissement, semblable au moins en son essence, cet accent est communiqué aux mots, par le sentiment qui agite le causeur ou le poéte, uniquement, sans souci dʼaccent tonique ou de nʼimporte quelle valeur fixe quʼils possédaient en eux-mêmes. Cet accent dʼimpulsion dirige lʼharmonie du vers principal de la strophe, ou dʼun vers initial qui donne le mouvement, et les autres vers, à moins quʼon ne recherche un effet de contraste, se doivent modeler sur les valeurs de ce vers telles que les a fixées lʼaccent dʼimpulsion. Cʼest cette loi fondamentale que MM Mockel et de Souza ont discernée à leur tour, en étudiant le rythme poétique et quʼils dénomment lʼaccent oratoire. (pp. 29–30)

I will leave consideration of the reference to Albert Mockel and Robert de Souza to the next section. I have quoted this passage in extenso to give some idea of the shifting and rather intangible nature of Kahn's definition; but three basic properties seem to belong to the ‘accent dʼimpulsion’. First, it is closely related to what we nowadays would call the ‘accent tonique’; Kahn uses this latter term to describe the accent which originally inhered in individual words; but this accent has gradually weakened and been superseded by the word-group or phrasal accent. In Kahn's terminology, then, the ‘accent tonique’ is the vestige of accent which individual words would have were they uttered in isolation; when words are strung together in connected discourse the ‘accent tonique’ (word accent) is superseded by the ‘accent dʼimpulsion’ (word-group accent; what we would now call the ‘accent tonique’). But the ‘accent dʼimpulsion’ and the modern ‘accent tonique’ are not identical though they most frequently coincide; for the ‘accent dʼimpulsion’ not only détermines where the accents are in a line, it also détermines how different degrees of emphasis are distributed among them, whether they are predominantly accents of pitch, intensity, or duration, and what kinds of patterns of intonation they correspond to. In short the ‘accent dʼimpulsion’ is the way voice inhabits and motivates text, is concerned with vocal impulses, with the way in which emotion, idea, sensation give movement (rhythm) to utterance. In this sense, the function of lineation in free-verse poems is to trace the protean life of the ‘accent dʼimpulsion’, to ‘sort out’, make visible, its changes of (p.128) momentum, intonation, and tone. In this sense, the ‘accent dʼimpulsion’ has a great deal in common with the sense-accent we discussed in the previous chapter, and may well highlight words that the tonic accent overlooks. The ‘accent dʼimpulsion’ is what orientates the reading of a whole stanza, centring its propulsiveness in the stanza's principal line (possibly the first line, possibly another). It is the point at which the free verse of the individual measure and cadence accentual free verse overlap. It is a paralinguistic supplementation of the linguistic, for while the tonic accent segments the syntactic chain linguistically, according to concerns of grammatical grouping, the ‘accent dʼimpulsion’ is generated by the enunciatory drives of poet and reader, by the need to invest the text with the varied psychophysiological impulses of the uttering organism. Thirdly, the ‘accent dʼimpulsion’ is a shared phenomenon (‘semblable chez tout le monde … semblable au moins en son essence’) and yet unique, in its particular configurations, to each poet and reader (‘le sentiment qui agite le causeur ou le poète, uniquement’); its activity is thus rather like the activity of Schopenhauer's Will. A little further on in the ‘Préface’, Kahn associates the ‘accent dʼimpulsion’ with ‘la déclamation instinctive du langage’ (p. 31), which also, according to him, justifies his treatment of the e atone.

It would be extremely difficult to translate these remarks about the ‘accent dʼimpulsion’ into scansional data, to tie down the presence of the ‘accent dʼimpulsion’ in reliable, observable findings; the ‘accent dʼimpulsion’ is after all a recitational accent. But it encourages us to try and diversify our account of the differing nature and varied interplay of accent, tempo, and intonation. To say, as Carmody (1969: 44) does, that Kahn obviously sees five ‘accents dʼimpulsion’ in the line

Des mirages | de leur visage | garde | le lac | de mes yeux

is to say no more than that Kahn sees five tonic accents. If tonic accents are to be ‘raised’ to ‘accents dʼimpulsion’ then the five accents must be described with a much more detailed code of intensity, duration, intonation, tempo, thus perhaps:

                      Theories and Enquiries

(p.129) where indicates accent of duration, … fadingness of accent, capital letters force of attack, ‴ and ′ accents of different degrees of intensity,                       Theories and Enquiries upward and downward pitch curves, and                       Theories and Enquiries close liaison of two measures.

But the ‘accent dʼimpulsion’ suggests one further, much larger, but crucial proposition. Kahn writes:‘Le vers libre est essentiellement mobile et ne doit point codifier de strophes. Cʼest lʼaccent dʼimpulsion et son appropriation à lʼimportance, à la durée du sentiment évoqué, ou de la sensation à traduire qui en est la déterminante’ (pp. 33–4). The proposition which seems to me to emerge from this remark is this: vers libre is not so much the product of formal developments which had been taking place over a period of some fifty years, but the direct response to changes in the perception of versification, and in the reading habits these entailed. The ‘accent dʼimpulsion’ is not a prosodic datum which derives from free verse; it is a transformed awareness of the nature of accentuation which necessitates free verse.

How appropriate are Kahn's theories to his own practice? To begin to answer this question let us look at the eighth poem in the ‘Finale’ cycle of Les Palais nomades:

No. of syllables

Scansion

10

1. Tant grande douleur | vint des gestes pâles

5+5

10

2. du timbre du verbe | illusoire | aux soirs.

5+3+2

10

3. Tant cruelle étreint(e) | vint de tes mains pâles

5+5

10

4. lʼâge du mirag(e) | des caress(es) | des soirs.

5+3+2

7

5. Abandonnée | dans ta foule

4+3

10

6. toute fléchissante | en ta dureté,

5+5

13

7. La neig(e) | de lʼimmanent hiver, | à ton cœur | qui croule

2+6+3+2

10

8. émanait | de langueur | des roses-thé.

3+3+4

11

9. Ah si nous savons | se déchirer | demain

5+4+2

11

10. laisse le sommeil | sʼimposer | de tes mains,

5+3+3

11

11. fuyons | la peur de neige | aux pupill(es) | solaires

4+2+3+2

12

12. boucliers | lucescents | de ta fac(e) | nécessaire.

3+3+3+3

preliminary scansional note: Following Kahn's precept, I have not counted the measure-terminal unelided ‘e’s; these are bracketed. Measure-internal ‘e's have been counted according to classical rules.

The meaning of this poem is extremely difficult to come at, both because of the ambiguity of syntactical relationships and because of the indeterminacy of several of the prepositions. For example: because of the lack of punctuation at the end of the first line, it is (p.130) impossible to tell whether the second line introduces a new element—‘vint des gestes pâles, vint du timbre du verbe’—or whether it simply creates a string of genitive modifiers—‘vint des gestes pâles du timbre du verbe’; it is impossible to tell whether the fourth line is in apposition to the third, or to the preceding three; although line 5 is clearly an adjectival phrase describing ‘La neige’, it is not clear whether line 6, also an adjectival phrase, belongs also to ‘La neige’ or to ‘ta foule’. Equally it is difficult to tell whether the ‘aux’ of line 2 is a temporal preposition (‘in’) or part of the construction ‘illusoire à’; the ‘ à ‘ of line 7 may be a positional preposition ‘in’, telling us where the ‘hiver’ is, or a directional preposition connecting the heart with ‘émanait’, or a preposition equivalent to the English ‘for’; and though the ‘de’ of line 8 is grammatically a ‘de’ of reason—‘out of languor’, ‘because of languor’—it slides towards a ‘de’ of source. Prosodically, on the other hand, the poem has the look of a barely masked regularity: the rhyme-scheme changes in the final stanza, it is true, where singular forms also rhyme with plural ones, and repetition replaces rhyme in the first stanza; but the alternation of feminine and masculine rhymes is more or less maintained (one might expect the third stanza to begin with a feminine rhyme); and the apparent regularity of the decasyllable (with a 5 ‖ 5 caesura?) is only deviated from in a relatively marginal way as the poem gradually heads towards a final resolution in an alexandrine of recurrent trisyllables.

So much for first reactions. Let us embark on a closer investigation. The first stanza sets up an atmosphere of slow time (‘lʼâge’), growing darkness (‘soirs’), pallor (‘pâles’), and illusoriness. Within this process of etiolation and moribundity, the poet is subject to the cruelty of a woman. The relative immutability of this state of affairs is driven home by the rhythmic regularity: each line begins with a pentasyllable measure which is either sustained in the second measure (ll. 1 and 3) or disaggregated into smaller measures (ll. 2 and 4). The pattern of maintenance and disaggregation corresponds with the stanza's alternating rhymes. But let us remember that for Kahn the overall decasyllabicity of these lines is not a metrical stipulation, but an accidental product of the combination of certain measures and the way in which the ‘idea’ of the stanza is analysed into its component parts (gestures, voice, embrace, summary)—this latter consideration already helps to (p.131) sort out some of the syntactical difficulties. In other words the poet is not writing decasyllables, but combining measures according to the dictates of the faceting of the underlying idea, only to find that he is constantly being reimprisoned within the same lineal perspective; the recurrent decasyliable is an expression of the pervasive uniformity which absorbs the ventured self-renewal of the individual rhythmic measures. And it is the very regularity that the measures impotently return to that defines the last line as the inescapable summation, the objective, the principal line, of the stanza. But for what prosodic reasons might we identify this last line as the nuclear line of the stanza? To answer this we need to take into account the combined activity of the ‘accent dʼimpulsion’ and acoustic kinships (phonic accentuation). First, though, a fuller scansion of this fourth line:

                      Theories and Enquiries

All the accents in this line seem to me to be predominantly of duration, with some élément of shallow pitch-change. Intensity of any kind is all but absent from this stanza; the relative immutability of duration presides. The intraphrasal accent on ‘âge’, encouraged by its near-assonance with ‘mirage’ (/a/ and /a/) and their common post-vocalic /Ʒ/ picks up the potential intraphrasal accents in the second measures of lines 1 and 3 (‘vint | des gestes pâles’, ‘vint | de tes mains pâles’), where equally the intraphrasal accent would fall on the first syllable of the measure. The latter two measures of this last line are acoustically linked by /R/ and /S/ and, though /S/ plays a fairly subdued part in the remainder of the stanza, /R/ is an obtrusive phoneme. The disaggregative impulse of these last two measures is also important, for while this last line encapsulates the fadingness of the light and the will, in the dying interval of its measure-terminal ‘e's (this is the only line in the poem with two of them) and the lethargic resignation of the poet, and in the durational nature of its accents, it also expresses the half-buried desire to break the pattern, to modulate into other, liberating combinations. The other acoustic élément within the stanza which has implications for accent is the recurrent / ε̃/phoneme (‘vint’, ‘timbre’, ‘étreinte’, ‘mains’), because, by its repetition, it expresses the voice's reaching for intensity, for an intensification of experience within the pervasive mood of atrophy; and indeed in (p.132) the third line a momentary point of intensity is reached in ‘étreinte’:

                      Theories and Enquiries

But this is an isolated occurrence and the ‘accent dʼimpulsion’ of the stanza as a whole is one which reaches for the serenity of its own attrition, while still nursing the half-hope of escape. As a postscript to this stanza one might mention that the repeated rhyme-words themselves act to paralyse activity and invention, and one might further suggest, in line with what was said about line-terminal repetition in Chapter 1, that, as it is repeated, so the rhyme-word undergoes a process of semantic atrophy or entropy.

As we move into the second stanza, the field of polarization shifts, from oppositions between cruelty and caress, reality and illusion, light and dark, vividness and pallor, swift time and slow time, to oppositions between coldness (‘neige’, ‘hiver’) and warmth (‘roses-thé, hardness (‘dureté’) and yieldingness (‘fléchissante’, ‘qui croule’), inwardness (‘immanent’) and outwardness (‘émanait’), activity (‘foule’) and apathy (‘langueur’). ‘Foule’ might mean the train of the woman's admirers by whom her attention is distracted from the ‘immanent hiver’; but it might also mean, I think, the multiplicity of the woman's aspects, many of which are more yielding than her mask of hardness implies—this idea is taken up in the reference to ‘ton cœur qui croule’. At all events, the apparent voluptuous warmth conceals and reveals an imminent and immanent winter of the soul, a winter deriving from the careless unguardedness of the very warmth which seems to oppose it. The unsolved or dialectical nature of the oppositions in this stanza is mirrored in the prosody, which is essentially a prosody of transition. A decasyllabic product appears twice in this stanza, at lines 6 and 8, but each time in a form crucially different from the decasyllables of the first stanza. Although the decasyllable of line 6 derives again from the combination of two pentasyllables, these pentasyllables are rhythmically impelled by accents of intensity rather than duration, accents deriving their thrust from the newly established acoustic dominants of the peremptory /ã/ (‘Abandonnée’, ‘dans’, ‘fléissante’), the high-mid front vowel /e/ (‘Abandonnée, fléchissante’, dureté’), and the voiceless dental /t/ (‘ta’, ‘toute’, ‘fléchissante’, ‘dureté’), thus:

                      Theories and Enquiries

(p.133) And the decasyllable of line 8, though two of its accents are once again durational, is constituted by an unfamiliar combination of measures:

                      Theories and Enquiries

This re-emergence of the durational is largely owed to another acoustic reorientation, around the feminine /R/ (‘hiver’, ‘cœur’, ‘langueur’), the /ε/ of ‘neige’, ‘hiver’, ‘émanait’, and the /œ/ of ‘cœur’ and ‘langueur’; and where one of these vowels is followed by /R/, the durational effect is increased.

Of the other two lines in the stanza, line 5 represents a premonitory break in the pattern of expected combinations, allowing the trisyllable to disengage itself from the shadow of the pentasyllable (and ultimately decasyllable) and establish a new independence for itself. And line 7 has strong claims to being the principal line in the stanza, since this is a stanza of transition and the line is appropriately Janus-faced, its final 3 + 2 combination looking back to the first stanza, while its hexasyllabic measure looks forward not only to the new combination of trisyllables which opens line 8, but also to the poem's close. In the stanza as a whole the ‘accent dʼimpulsion’ is much more mouvementé, vacillating between intensity and duration, testing all kinds of rhythmic combinations, but finding some equilibrium for itself in the tetrametric structure of the seventh line, which nicely balances two measures of cold and inner lifelessness with two measures of yielding inner warmth.

The poem resolves itself, in the last stanza, in reconciliation rather than resignation, in the dialectical cohabitation of opposites rather than in their internecine antipathies. In order that this can be so, time must be suspended in sleep, because it is in temporal development, in their acquiescence to a tomorrow, that oppositions prove mutually destructive, mutually corrosive. The first two lines hark back, in their initial measures, to the rhythmic impulses of the first stanza, the impulse to fade away in a slow time in which everything is increasingly illusory. The measures of 4 and 2 of the first line remind us of their essential negativity, ‘se déirer’ throwing us back to ‘Abandonnée’ and ‘des roses-thé’, ‘demain’ full of the threat of ‘La neig(e)’ and, more distantly, echoing the measures of oncoming darkness (ll. 2 and 4). But if these two measures are inimical as independent entities, when combined (p.134) they promise transcendence in a suprasegmental hexasyllable. But the suprasegmental 6 can only reconcile antinomies in a balance of combined trisyllabic measures, which is essayed in line 10—the woman's hands are no longer cruelly embracing (l. 3), but soothingly ministering, bringing the suspended animation of sleep. But this attempt to establish a new experiential modus is almost immediately thwarted by the resurgence of the apprehensions of the 4+2 combination in line 11. This line, however, reverses the rough 5 | 6 partition of lines 9 and 10, so that the leading élément is no longer the first stanza's plaintive or submissive pentasyllable, but a hexasyllable still in search of an inner concord; and the second élément, though still a suprasegmental 5, is the disaggregated dimetric version. And the trisyllable in the second half of line 11, with its image of the eyes, provides the passage into the tetrametric final line with its pervasive trisyllabic measures. In what sense is this final line, certainly the principal line of the stanza, a reconciliation of opposites? It reconciles hardness (cf. ‘cruelle’, 1. 3, and ‘dureté’, 1. 6) and growing or radiating light in ‘boucliers lucescents’; it reconciles warmth and coldness by the fact that the -aire (/εκ/)suffix of the ‘solaires’/‘nécessaire’ pair rhymes with ‘hiver’ and assonates with the /ε/ of ‘neige’; it reconciles light and darkness, eyes and hands, in the sequence of /s/ phonemes (‘lucescents’, ‘face’, ‘nécessaire’) which echo the same phoneme in the ‘caresses des soirs’ of the first stanza; it reconciles the accent of duration and the accent of intensity by balancing two accents of intensity—on ‘boucliers’ (cf ‘Abandonnée’, ‘dureté’) and ‘lucescents’ (cf. ‘fléchissante’)—against two accents of duration—on ‘fac(e)’ (/a/, cf. ‘mirage’) and ‘nécessaire’ (/εR/ combination, cf. ‘hiver’):

                      Theories and Enquiries

Thus the ‘accent dʼimpulsion’ is a questing accent, an accent impelled towards triumphant declaration, which, though initially frustrated in different ways, finally finds its way to its goal.

I cannot pretend to have done comprehensive justice either to the meaning or the prosody of this poem, but I hope I have begun to show, despite the obstacles thrown up by this poem's very leaning towards regularity, that Kahn's theories are a useful instrument in the analysis of his own poems, that the practice does significantly correspond to the theory.

(p.135) (ii) Albert Mockel

In 1894, Albert Mockel, one of the leading lights of Belgian Symbolism, published his Propos de littérature. He had already published free verse in his Chantefable un peu naïve of 1891 and was to publish another volume, Clartés, in 1902. He made no secret of the inspiration he had derived from Kahn's work.1

Propos de littérature is a comparative study of the poetry of Henri de Régnier and Francis Vielé–Griffin, which allows Mockel not only to outline his own theory of free verse, but to pursue much larger aesthetic issues: the differences between symbol and allegory—an extremely penetrating discussion which has lost none of its relevance—and the equivalences between the various arts, pictorial, scultpural, and musical. I leave these larger issues to one side and turn directly to the fourth chapter, in which Mockel addresses himself to the prosody of free verse; or almost directly: it might be helpful first to quote Mockel's summary of his provisional conclusions about the general differences between the poetry of Régnier and that of Vielé–Griffin; this summary appears at the beginning of Chapter 3:

M. de Régnier, dont le fatalisme répugne à lʼaction, écarte de son art le mouvement; il se manifeste en général par la plastique plutôt que par la musique, et spécialement par des attitudes; il y a du sculpteur en lui. Chez M. Griffin, poète épris de vie, la morale de lʼactivité se traduit dʼellemême: cʼest le geste et le rythme qui caractérisent la strophe en lʼenveloppant dʼun souple et mouvant tissu. (1962/1894: 105)

Chapter 4 opens with some paragraphs on the relationship between harmony and rhythm. Régnier is essentially a harmonist, devoted to the arts of space, stasis, and objectivization; Viélé-Griffin is essentially a rhythmist, concerned with the temporal, the dynamic, and the subjective. Mockel's identification of Régnier as a harmonist is hardly surprising given Régnier's devotion to a vers (p.136) libre which is a series of variations on a present, or absent, alexandrine (see quotation above, p. 18); as Mockel puts it: ‘Le vers libre ne procède pas de soi-même, chez M. de Régnier, mais dérive de lʼalexandrin quʼil allonge ou reduit pour sʼy égaler de nouveau’ (pp. 136–7). But, for Mockel, one ignores the dictates of an intrinsic rhythm, of a rhythm springing directly and ingenuously from self, only at one's poetic peril, and though Régnier often achieves this impulsive suppleness, or manages to mix new with old (‘les mouvements naturels de la voix avec la fixe arcature dʼun metre logique’, p. 137), he too often falls into sequences of lines whose rhythms do not live their own lives but are borrowed from the classical pattern-books, and are therefore non-propulsive and inert. Such verse of course may have compensations in its very impersonality, in its lapidary sculpting of ‘la chose définitive’.

The dangers for Vielé-Griffin lie in the other direction. For one thing his verse can become utterance too directly addressed to the reader, and assume a certain didactic insistence. For another, he sometimes makes his lineation too subservient to movements of voice, that is to say to movements which are elocutionary rather than rhythmic, so that the phrase becomes proserelated and the stanza a compilation of short-winded, unsustained, rhythmically incoherent syntactical fragments. As an example of this weakness, Mockel quotes the first stanza of the ‘Cinquième Chanson’ from Les Cygnes, whose opening lines run:

  • Voici ma pensée:
  • Si la flèche,
  • Que mon arc lance aux étoiles,
  • Retombe et blesse
  • Ma main qui lʼa lancée
  • Vers les étoiles.

Mockel, like Kahn before him, is a proponent of the ‘logically’ constructed stanza: ‘Cʼest l‘analyse logique qui détermine les limites du vers moderne … La proposition grammaticale coïncide avec le vers, ou, plus souvent, les membres de chaque proposition sont présentes séparément’ (pp. 129–30). The classical poets were hamstrung by metrical conventions which compelled them to do all their poetic thinking in hexasyllabic units, a condition which the Romantic poets rebelled against in their cultivation of enjambement; but enjambement was no remedy since it destroyed the desirable (p.137) simultaneity of syntactic and musical impressions. With the advent of free verse

lʼanalyse logique peut dorenavant coïncider avec le vers sans amener lʼuniformité, car elle sʼunit au rythme désinvolte et primesautier, et chaque ligne nouvelle (ou presque!) peut offrir un nombre de syllabes nouveau. Le vers est né de sa propre vie; sa longueur comme sa force rythmique ne dépendent plus que du sens grammatical quʼil contient … et de son importance comme élément musical: il est désormais logiquement conçu. (p. 133)

Accordingly enjambement is no longer necessary as a device of variation and metrical transgression, since variation is now built into the very fabric of the verse and there are no metrical boundaries to transgress. Enjambement may still be resorted to in exceptional circumstances and for some specific effect, but otherwise it has outlived its usefulness.

But, as we have seen, in the observations about Vielé-Griffin's verse above, lineation is not just a question of the syntactic segmentation of the sentence. Rhythmicity is not to be achieved by a mechanical process of découpage; it can only be achieved by the addition of a harmonic dimension; that is to say that although rhythm is essentially bound up with the psychophysiology of the poet, with the spontaneous jaillissement of the self, it can only attain to an aesthetic status if it acquires some objective validity, some pattern perceptible in space, some formal integrity, some informing continuity. The presence of harmony manifests itself in three ways: in accentual consistency, in verbal music, and in the accent oratoire.

By ‘accentual consistency’ I mean that feature we have associated with cadence accentual free verse: each line contains the same number of phrasal sense-accents, though there may be variation in intraphrasal accentuation.

Voici: comme je lʼécrivais tout à lʼheure, la grande question, au point de vue des techniques récentes, est le rapport du rythme à lʼanalyse logique. Quelques littérateurs nʼont pas lʼair, il est vrai, de sʼen préoccuper beaucoup; dʼautres le cherchent dans un nombre plus ou moins constant de syllabes toniques et peuvent réaliser ainsi un art où lʼélan fol de lʼinstinct nʼexclut pas lʼéquilibre des propositions, (p. 143)

Mockel is evidently one of the latter group. In this he goes beyond Kahn, who tackled the problem of the rhythmic integrity of the (p.138) line only from a musical point of view, but otherwise was content that the line should be the providential product of the combination of any number of measures. While a measure-verslibriste like Kahn looks, for rhythmic shaping, to the recurrence of a dominant measure, to a constante rythmique, within the framework of the stanza, Mockel, as a verslibriste of the measure with a leaning towards cadence accentual verslibrisme, looks for an intermediate rhythmic integrity at the level of the line without in any way restoring authority to syllabicity. One of the attractions of accentual consistency is precisely that it can maintain itself despite considerable syllabic variation:

La mesure de quatre temps, par exemple, peut contenir des vers de sept à douze syllabes, selon que les rythmes sont binaires, ternaires ou alternants,2 et davantage si à chaque syllabe forte correspondent trois atones; et quelle richesse de mélodie variée, par lʼentrelacement des tonales et des semi-toniques, des longues et des brèves, des notes graves ou subtiles! (pp. 134–5)

But then the principle of accentual consistency need not, and indeed should not, be rigidly adhered to (‘un nombre plus ou moins constant de syllabes toniques’); in two stanzas from Vielé-Griffin's ‘Vous si claire’ (Joies, 1889) quoted approvingly by Mockel but not analysed by him, we find a dominant tetra-accentual pattern, interspersed with four lines of three accents, and one of two:

‘Vous, | si claire | et si blonde | et si femme,

4

Vous | tout le rê: | ve des nuits | printanières,

4

Vous | gracieu: | se comme une flamme

3

Et svelte | et frêle | de corps | et d’âme,

4

Gaie | et légère | comme les bannières;

3

Et ton rire | envolé | comme une gamme,

3

En écho, | par les clairiêres—’

2

‘Vous | ma fierté | tout enorgueillie,

3

Vous | seul but, | seule voie | et seule fin,

4

Vous | de qui seul | je me rêvais | cueillie,

4

Vous | mon poème | et ma soif, | et ma faim,

4

Quel soir | est tombé, | quelle heure | est vieillie?

4

(p.139) More closely in line with Kahnian precepts are Mockel's words on verbal music. He criticizes Vielé-Griffin for his neglect of the integrative function of sound and his continuing promotion of rhyme as an independent device:

Mais en général M. Vielé-Griffin oublie la puissance expressive de lʼharmonie. Le plus souvent les sons de ses vers ne réunissent leurs voix que par la rime, laquelle est presque puérilement gardée; car si sa position en évidence la doue dʼune importance spéciale, la rime nʼa pourtant pas, dans le vers moderne, un rôle indépendant du rôle des autres sons. Elle doit, pour acquérir toute sa valeur, sʼallier avec les tons syllabiques voisins ou se fondre en leur rumeur quʼelle peut alors synthétiser par sa note vive. (p. 129)

Harmony, in this musical sense, is an uninterrupted and seemingly organic concordance of sounds, acting cohesively on the stanza, producing an indissoluble network of echoes and modulations. And, as we mentioned in our treatment of Kahn, acousticity becomes a vital ingredient in the projection of accent. Sometimes a successively repeated sound will progressively intensify accentuation; sometimes a set of acoustic kinships will redistribute the expected accentual hierarchy, giving priority to intraphrasal or secondary accents; in this latter sense sound patterns are a method whereby the free-verse poet can play down grammatical demands in favour of subtextual psychic or emotional ones.

The notion of the ‘accent oratoire’, which seems to have confirmed Kahn in his perception of an ‘accent dʼimpulsion’, is touched on only passingly by Mockel, in relation to the Romantic poets’ treatment of the classical line:

Sʼils gardérent intacte sa charpente, ils voulurent au moins renouveler lʼintérieur décor et toute lʼatmosphère; le vers, conservé en ses strictes limites, put désormais faire mouvoir entre ces bornes solides les rythmes internes qui le varièrent, tandis que lʼaccent oratoire, issu de la logique même de la phrase, passait à travers la mesure en se combinant avec elle ou en sʼy opposant à la manière de ce quʼon nomme en musique la Syncope. La grâce et la force des œuvres de ce temps fut de cacher souvent la mesure arbitraire révélée par la seule rime, de la voiler sous lʼondulant tissu de vers réels, de vers inégaux et libres ayant dʼune et de deux jusquʼà seize syllabes ou davantage. (pp. 130–1)

This is all the help that Mockel gives us. The ‘accent oratoire’, it seems, is the accent which defines spans of syntax and thus the (p.140) periodicity of an utterance. The Romantics use the trimètre and enjambement to create a counterpoint between the ‘accent oratoire’ and the pattern of metrical accents; or, to use Mockel's own term, in these instances of the non-congruence of syntactic juncture and metrical juncture the ‘accent oratoire’ produces an effect of ‘syncopation’. The ‘accent oratoire’, whether it coincides with tonic accents or not, serves harmony by establishing patterns of cadence, by engendering an ‘ondulant tissu’ within verse; it is that which registers the respiratory configurations of the voice as it gives shape to its own enunciation.

In view of the sketchiness of Mockel's account, it is more likely that Kahn owed the theoretical substance of his ‘accent dʼimpulsion’ to Robert de Souza's more detailed analysis of the ‘accent oratoire’ in Le Rythme poétique (1892: 242–54). Like Kahn, de Souza identifies the ‘accent tonique’ with the accent which is intrinsic to the individual word, the accent which has weakened with time and which is superseded in connected discourse by a phrasal accent. This latter is what de Souza calls the ‘accent oratoire’, which will always coincide with a tonic accent and is a blend of the ‘logical’ (the accent demanded by grammatical segmentation) and the ‘pathétique’ (the accent demanded by sense, where sense includes feeling, psychophysiological impulse, and other manifestations of self): the ‘accent oratoire’ is ‘à la fois tonique, dʼacuité, dʼintensité, rythmique, cʼest-à-dire ramenant à lui, concentrant toutes les énergies’ (1892: 245); de Souza further adds: ‘Pathétique et logique ne prennent rythmiquement toute leur force quʼen une alliance étroite’ (p. 247). Rhythm, therefore, is as much derived from variations in the degree and nature of accentuation as it is from variations in the position of accent: ‘Ce nʼest en effet que le jeu des accents FORTS ET FAIBLES qui peut achever dʼanimer le mouvement rythmique’ (p. 253), an observation which bears out our findings in Chapter 1. Like many of his verslibriste contemporaries, de Souza is anxious to establish that rhythm is not the result of a purely syllabic principle whereby the fixity of a number of syllables compensates for accentual weakness and imposes a rhythm merely by segmenting its own number according to predétermined rules (the caesura); rhythm is the result of accentual activity of an intense and varied kind and it is only by demonstrating this accentual activity that the tyranny of the line can be broken; for there is obviously a close connection (p.141) between a lack of accentual activity and a justifiable insistence on the sanctity of syllabic number. As a footnote, it should be emphasized that de Souza only applies his concepts of the ‘accent oratoire’ to regular verse; whatever his debt to de Souza, Kahn certainly deserves the credit for having adapted the concept to free verse and for having integrated it into his own prosody of the stanza, with its generative principal line.

It is quite probable too, then, that Mockel's presentation of the ‘accent oratoire’ is shadowy because he has not confidently digested what de Souza has to say about it. Much of Mockel's argument consolidates Kahnian positions. Where Mockel differs from Kahn is in his greater emphasis on the musical affiliations of free verse, due in some measure no doubt to his enthusiasm for Wagner—

Enfin, lʼon a senti lʼinfluence de Wagner qui développant le récit beethovénien, supprime la carrure de la phrase au profit dʼun rythme large et continu et, devancé en cela par les Romantiques, comme on lʼa vu, juxtapose parfois des mesures aux nombres divers. Tout semble enfin sʼunir pour favoriser le developpement libre du rythme. (1962/1894: 132)

—and in his pursuit of a cadence accentual prosody of the line. Mockel is obviously afraid that a free verse of the measure will only exacerbate that tendency towards fragmentation and rhythmic jaggedness which he intermittently finds in Vielé-Griffin's work. He does not share Kahn's faith in a prosody dispensing with the line and emerging from the polarization of measure and stanza. Free verse for Mockel means replacing an old Procrustean prosody of the line with a new prosody of the line, elastic but aspiring towards a describable, objective, aesthetic status.

Something of all these issues can be seen, I think, in Mockel's own free verse, and I would like to bring this section to a close with a brief consideration of the first part of Mockel's most anthologized poem ‘Ange’(Clartés), in which a mysterious youth, found asleep under some holm-oaks, turns out to be a stranded angel who will soon return to the empyrean: (p.142)

No. of accents

No. of syllacles

Scansion

2

1. Ouelquʼun ici sʼest endormi.

8

4+4

3

2. Dans le matin léger, parmi les dômes des yeuses

13

6(4+2)+4+3

3

3. il repose, innocent et las, sur lʼherbe heureuse,

12

3+5+4

3

4. et lʼombre, à peine mouvante sur lui,

10

2+5+3

3

5. autour de son sommeil prolonge un peu de nuit.

12

6(2+4)4–2+4

3

6. Quel est-il, cet enfant apparu tout à coup?

12

3+3+6(3+3)

3

7. Un seul de vous sait-il ici

8

4+2+2

3

8. dʼoù vient ce blanc voyageur juvénile

10

2+5+3

2

9. qui sʼest arrêté parmi nous?

8

5+3

3

10. Est-il parti des mers au loin, où sont les lies,

12

4+4+4

3

11. ou là-bas, des forêts? ou des plaines stériles

12

3+3+6(3+3)

3

12. dont nul, jamais, nʼa sondé lʼétendue?

10

2+2+6(3+3)

4

13. Il est blanc; il est nu. Toutes les pierres de la route

14

3+3+4+4

4

14. nʼont pas blesse ses pieds ni meurtri ses genoux;

12

4+2+3+3

4

15. il y a sur son front quelqué chose que Ton Redoute…

14

3+3+3+5

3

16. Dʼoù vient-il, avec sa parure de beauté,

12

3+5+4

3

17. lui qui sʼest arrêté parmi nous?

9

1+5+3

2

18. Sa chevelure sʼest repandue

9

4+5

2

19. comme une vague de clarté;

8

4+4

3

20. sa main reclosetient une fleur inconnue,

12

4+5+3

2

21. et toute sa candeur enchantée

9

6+3

2

22. est comme une image des nues

8

5+3

3

23. que lʼeau mirante redouble en elle…

9

4+3+2

3

24. Fréres, gardons quʼil ne sʼéveille!

8

2′+2+4

Preliminary scansional note: Having found a tri-accentual pattern recurring with some frequency, I have treated as tri-accentual all lines which could be so treated without syntactic distortion. This does not mean to say that many of the lines I have read tri-accentually are not susceptible of tetra-accentual or bi-accentual readings. In my syllabic count, I treated the e atone in the traditional way, largely because, by this device, the recurrence of the dodecasyllabic line, with its inescapable recognizability, with its attempt to install a familiarity within an enigmatic experience, is maintained (I think particularly of ll. 11, 16, and 20). I envisaged the possibility of treating the e atone inconsistently, to increase the incidence of parisyllabic or more traditional lines; such a manœuvre might be defensible in lines 13 and 23, for example, where apocopes would produce a dodecasyllable and an octosyllable respectively:

  • Il est blanc; il est nu. Tout(es) le pierr(es) de la route
  • que lʼeau mirant(e) redouble en elle…
But there are powerful acoustic and expressive reasons for not pursuing such a policy; in line 15, for instance, one can hardly entertain:

il y a sur son front quelquʼ chosʼ que lʼon redoute,

while the articulated ‘e’ is necessitated by the sense in line 18 (‘sʼest répandue’) and is needed in lines 2 (‘dômes’) and 21 (‘toute’) to capture the sense of envelopment or suffusion. Equally I have assumed that all post-accentual unelided ‘e's create coupes enjambantes, with the exception of ‘Frères’ in line 24, where the case for a coupe lyrique seems undeniable.

(p.143) The question I want to address myself to in this analysis is single and simple: according to what free-verse ‘metric’ should the poem be read: the free verse of measure, cadence accentual free verse, or the free verse of line?

The free verse of measure can, I think, be quickly disposed of. Looking at the tabulation of measures, I see no single measure which I could call rhythmically dominant, no constante rythmique; in fact trisyllabic, tetrasyllable, pentasyllable, and hexasyllabic measures all seem to have a considerable part to play in the rhythmic construction. It would be more accurate perhaps to say that, in this poem, the measure has more intralinear significance than it has interlinear significance; by this I mean that any given measure is more significant by virtue of the sequence it creates with other measures in the same line, than by virtue of the pattern it produces with similar measures in other lines. Thus, for example, the pairs of tetrasyllables in lines 1 and 19 are more important in their self-completing function, in their function of mutual reinforcement, than as éléments in a larger configuration; this is borne out by the internal rhyme of line 1 (/i/) and the internal assonance of line 19 (/a/). Similarly the trisyllables in line 15 are more important as a movement of rhythmic stabilization, of perceptual confidence, undermined by the destabilizing apprehen-siveness of the final pentasyllable (‘| e que lʼon redoute’), than as units in a trisyllabic context. Similarly the sequence of tetrasyllables in line 10 is significant only for the duration of the line, as the poet ‘prend le large’ and is rocked in the expansive undulations of the sea, while in the line following, taking into account the intraphrasal accent on ‘plaines’, the horizons have shrunk and the movement, though equally expressing the uniformity of terrain, is more short-winded. And further, taking into account the activity of intraphrasal accents, we might point to the line-internal symmetries and continuities of lines 5 and 6.

There seems to me, then, no doubt that Mockel's free verse is a (p.144) free verse based on the rhythmic integrity of the line, whose field of influence is enlarged not by rhythmic cross-reference so much as by the concertedness of acoustic echoes—one might note for example the thoroughgoing activity of /y/, /u/, /o/, /ã/, /a/ among vowels—and by the interstanzaic play of rhyme and line-terminal assonance. But both cadence accentual free verse and the free verse of line look to the continuing integrity of the line for their rhythmic effects. Which of these does Mockel's verse gravitate towards?

Given Mockel's favouring of the parisyllabic line, and particularly the dodecasyllable and the octosyllable, the conclusion might appear to be foregone. And accordingly my treatment of lines 2, 5, 6, 11, and 12 as tri-accentual, with intraphrasal accents in the hexasyllabic segments, rather than as tetrametric, might appear to be a case of unwarranted special pleading for the cadence accentual version. In defence of my conduct and in an effort to refute the idea that Mockel's is a free verse of the line, I would like to recall the ‘accent oratoire’ and Mockel's objection to Vielé-Griffin's excessive and anti-harmonious decoupage of the rhythmic period. Mockel's particular objection is, paradoxically, closely related to a more general objection that one might level against regular verse. On occasion, and regardless of rhythmic coherence and the sustaining impulse of enunciation, Vielé-Griffin mechanically chops up his verse at every grammatical juncture and allots a line to each fragment. But is it not true to say that regular verse, with its eye on metricity rather than rhythmicity, is inclined to do the same, albeit within the span of the line? Does not a deeply ingrained and unquestioning need to find a tetrametric configuration wherever possible in the alexandrine lead the traditional analyst to practise an arbitrary and mechanical segmentation of the dodecasyllable, which ignores the cohesiveness of feeling and idea, ignores the informing cadences of voice, ignores, in short, the ‘accent oratoire’, in the name of metrical convention? If one combined a preoccupation with rhythmic regularity with a hatchet-attitude towards the phrase, if one let an obsessive concern with the recurrence of the tonic accent have its way, what kind of rhythms would lines be subject to? Perhaps

il repose, | innocent | et las, | sur lʼherbe | heureuse 3+3+2+2+2

or

(p.145) et lʼombre | à pei: | ne mouvan: | te sur lui 2+2+3+3

or

Un seul | de vous | sait-il | ici 2+2+2+2

or

dʼoù vient | ce blanc | voyageur | juvénile 2+2+3+3

These segmentations may make some grammatical sense, and in some contexts they might make metrical sense as well; but they are uninhabitable by the voice; they show absolutely no sensitivity to the variable and expressive periodization of utterance. Cadence accentual free verse, by replacing metrical segmentation with rhythmic segmentation, allows the dodecasyllable to escape its stifling tetrametricity once and for all, and at the same time admits the ‘accent oratoire’ as its animating principle. The ‘accent oratoire’ is integrative at a suprasegmental level, is not subject to metrical predetermination, but free to synthesize the differentiated local impulses of utterance into a larger, overarching unity. In this sense Mockel's ‘Ange’ is the battleground of two conflicting principles of rhythmic organization, a battleground on which the recognizably regular line has to surrender its traditional mode of rhythmic realization to a new mode, more variable and expansive in its designs.

(iii) Édouard Dujardin

To close this short sequence of theoretical writings, we turn to a retrospective work, Dujardin's Les Premiers Poètes du vers libre (1922). This survey grew out of a series of lectures delivered at the Sorbonne in October 1920, and the text first appeared in the Mercure de France of 15 March 1921.

Dujardin makes clear from the outset that he has no ambitions as a theorist; he sees his task as providing a historical mise au point, tracing the steps in free verse's appearance and getting rid of some of the misconceptions that free verse has been a prey to—he notes that Camille Mauclair has spoken of René Ghil's vers libre and of the vers libre of Laforgue's Les Complaintes; and that, according to Ernest Raynaud, Charles Cros wrote vers libre; Verlaine's verse has been subject to similar misnomers. For all his (p.146) disclaimers of theoretical intentions, Dujardin finds the need to define free verse inescapable, and he proceeds to this definition on the basis of differentiations between regular verse, vers libéré, and vers libre. I quote his initial distinction between vers libéré and vers libre:

Le vers libéré est celui qui nʼexige plus la césure à lʼhémistiche dans lʼalexandrin, admet lʼhiatus et la rime pour lʼoreille seule sans distinction des masculins et féminins ni des pluriels et des singuliers, et use couramment des chiffres (peu usités dans les vers réguliers) de 9 et de 11 syllabes et ceux de 13, 14 et même 15 et 16 syllabes. Verlaine en a été le promoteur; la plupart des poetes le pratiquent plus ou moins aujourdʼhui.

Le vers libre, enfin, est celui qui, poussant à lʼextrême la libération, est susceptible dʼun nombre de syllabes indéterminé, ne compte (selon certains) lʼ muet que lorsquʼil se prononce, admet lʼassonance à la place de la rime, et se caractérise en ce que, semblable en cela au vers libre classique, il sʼemploie le plus souvent groupé en séries de vers inégaux. (1922: 8–9)

There is nothing to quarrel with in these definitions, but Dujardin finds them inadequate, even unacceptable, because they do not engage with the rhythmic foundations of French verse. Like many of his verslibriste contemporaries, Dujardin rejects the idea that because of the weakness of French accent, French verse is essentially syllabic, that the syllable is the basic rhythmic unit. Accent is a real acoustic phenomenon in French and it is accentual recurrence that creates rhythmic pulse in French verse; the basic rhythmic unit in French is therefore the ‘pied rythmique’, which Dujardin describes thus:

Un pied rythmique, exactement, se compose dʼun mot ou dʼun ensemble de mots, lequel 1° porte un accent à la dernière syllabe (à lʼavant-dernière, si la dernière est muette); 2° peut porter un ou plusieurs accents secondaires ou demi-accents sur une ou plusieurs autres syllabes, et 3° comporte par lui-même assez de signification pour permettre un minimum dʼarrêt de la voix. (p. 10)

This definition puts Dujardin's thinking very much in line with Kahn's, though Dujardin makes more explicit allowance for the presence of intraphrasal accents. We have encountered the intraphrasal accent fairly frequently in our analyses of vers libre, but we should remind ourselves of its purely polemical function in the prosody of free verse: it is a way of drawing the attention of the (p.147) die-hard syllabic school to the insistent accentuatedness of the verse-line; it is a way, too, of combating an over-exclusive preoccupation with the tonic accent, a preoccupation which tends to minimize the expressive function of accent in favour of its metrical job, which is to provide assimilable fractions of the line's overall syllabic count; indeed, going further than Dujardin, we have found instances, particularly in cadence accentual free verse, in which these so-called secondary accents act as the expressive nuclei, the accentual centres, of the measures in which they appear.

The crucial question, then, which confronts the analyst who would differentiate between regular verse and vers libéré on the one hand, and vers libre on the other, is: according to what principle does a succession of ‘pieds rythmiques’ constitute an ensemble, a line of verse? Before answering this question, Dujardin introduces two further factors, the unity of thought and the number of syllables.

For Dujardin the line of verse can be considered as a formal unity which corresponds to ‘une unité intérieure’ deriving from lʼunitè de signification, lʼunitè de vision, lʼunité musicale’ (p. 12). In support of this precept, Dujardin approvingly quotes Kahn's definition of the line as ‘un fragment le plus court possible figurant un arret de la voix et un arrêt du sens’ (1897: 26). The difference between vers libre and regular verse or vers libéré is that while the lines of vers libre always constitute such a unity, those of regular verse or vers libéré often do, but equally often do not; in other words the much-vaunted metrical unity of these latter forms of verse is frequently a pretext for the creation of lines woefully deficient in other kinds of unity. Here Dujardin is implicitly pursuing a line of argument explicit in Kahn's ‘Préface’: that the practice of enjambement, by which the Romantics were ostensibly liberating verse, is in fact injurious to verse, a false route. Though Dujardin's continuing concern with the line is greater than Kahn's, both associate enjambement too exclusively with questions of rhythmic (and thus syntactical) integrity, with an infringement of that integrity. However, not only does Dujardin's promotion of intraphrasal, ‘secondary’ accents itself suggest the kind of intraphrasal juncture we associate with enjambement, but neither to Dujardin nor to Kahn does it occur that free verse, with its emphatically oral orientation, with its privileging of énonciation (p.148) over the énoncé, might derive particular expressive benefits from those vocal hesitations, apprehensions, and anticipations which enjambement enacts. While it is true that both the free verse of the independent measure and cadence accentual free verse tend to discourage enjambement as habitual practice, we have found examples enough in Alain-Fournier and will find others in Claudel which testify to enjambement's continuing significance as a verse-resource; and, in that free verse which plays with the ghosts of metricality (Laforgue, Henri de Régnier, Apollinaire), enjambement, perhaps more pointedly, can continue to flaunt its metrical excuse.

As far as the number of syllables is concerned, the difference between regular verse or vers libéré and vers libre is absolute: ‘Le vers régulier ou libéré est celui qui tient compte du nombre des syllabes; le vers libre est celui qui nʼen tient pas compte’ (p. 14). So the law which orders the ‘pieds rythmiques’ of regular and liberated verse is the total number of their syllables, while the only principle which presides over the ordering of ‘pieds rythmiques’ in vers libre is that of unity, regardless of syllabic number. The verslibriste does not count syllables. This remark might be compared with Rémy de Gourmont's ‘il ne faut plus, ici moins que jamais, compter les syllabes, il faut les nombrer’ (1955b/1899: 162–3). While it is not absolutely clear what Gourmont means by ‘nombrer’, it would seem to mean something like ‘mark the time of. Of the following two lines:

  • lis virent les pins sévères de la mélancolie
  • barrer les blancheurs septentrionales
Gourmont says ‘le plus long, si lʼon nombrait avec une précision arithmétique, serait peut-être le second’ (p. 163). In other words, one does not count syllables as digits but as note values, some having more duration, being double or triple the value of others, some subject to a slower tempo or more pausing. Gourmont's reader is one who practises catalexis and other forms of temporal and rhythmic compensation, to arrive at a rough isochrony between lines.

So Dujardin is in a position to arrive at his final conclusions:

Les caractéristiques qui distinguent le vers libre du vers régulier ou libéré sont done, en fait:

  1. (p.149) 1. Lʼunité formelle du vers (correspondant à son unité intérieure) devenue obligatoire;

  2. 2. Lʼignorance du nombre des syllabes; et jʼajoute, comme donnée accessoire;

  3. 3. La libération dʼun certain nombre de petites règies, telles que césure, rime, hiatus, etc. (1922: 14–15)

All it remains for Dujardin to do, before moving on to his investigation of the first examples of free verse, to which we have already referred, and to some speculation about the different paths by which the first free-verse poets arrived at their verslibrisme, is to provide two further definitions, one of the prose poem, the other of the verset. The prose poem ‘diffère du vers, dʼabord en ce que ses pieds rythmiques sont généralement dʼun rythme moins prononcé, ensuite et surtout en ce quʼils ne sont pas ordonnés dans lʼunité resserrée du vers’ (p. 17). For these reasons the difference between vers libre and the prose poem is greater than that between vers libre and regular verse. Our own findings in section (iii)(e) of Chapter 2 would lead us to contest this assertion. The verset is not so much a form intermediate between vers libre and the prose poem as an enlarged vers libre:

le plus souvent composé lui-même de plusieurs vers libres étroitement associes, à lʼexemple du verset biblique; il peut ainsi sʼallonger, ou se réduire à un seul vers libre, même à un seul pied rythmique. Commé le vers libre, qui est composé de pieds rythmiques et est lui-même une sorte de pied rythmique supérieur, le verset est essentiellement une unité. Nous retrouvons là, on le voit, la caractéristique du vers, et non celle du poème en prose, (p. 22)

This is an argument of convenience, whose persuasiveness lies much more in its logic than in its demonstrability. Certainly distinctions between vers libéré, vers libre, verset, and prose poem must be essayed if only to justify their different denominations; but their points of overlap and interpenetration are, as I hope has already been shown, manifold, and it is wiser and safer to think of these different forms as only roughly circumscribable points in a single rhythmic continuum. And indeed it is the form which would encompass this continuum, and its relativities, that Dujardin himself has it in mind to discover:

Et cʼest pourquoi jʼai cru (depuis longtemps et de plus en plus) quʼil était possible de trouver une forme qui passerait, sans transition et sans heurt, (p.150) de la forme vers à la forme prose, suivant lʼétat lyrique du moment, et, toujours sans heurt et sans transition, serait elle-même vers libre, verset et poème en prose, dans une succession de pieds rythmiques tour à tour serrés en vers, élargis en versets et dilués en quasi-prose. (p. 22)

Dujardin speaks rather disparagingly of his own free-verse poems, but, as in our treatment of Kahn and Mockel, we might profitably explore the relationship between Dujardin's definitions and his practice, bearing in mind the large temporal gap which separates them. Accordingly I would like to undertake a brief analysis of the sixth poem in the ‘Les Chansons du rivage’ cycle which opens the ‘Les Chansons du rivage’ (1889) section of the collection La Comédie des amours, first published in 1891:

No. of syllables

Scansion

6

1. Void les violons …

2+4

6

2. Oh! dans les clairs salons

1+5/4+2/1+3+2

2

3. Vaisons!

2

7

4. En ce vent virant qui vague,

5+2

7

5. Lʼâme sur des voiles vagues

1+4+2

7

6. Vole et, véloce, extravague.

1+3+3

12

7. Le rythme de la valse est un incantement

2+4+6/6+6

3

8. Et son chant

3

8

9. En un magique girement

4+4

11

10. Entraîne lʼesprit hors lʼespace et le temps.

5+3+3

11

11. Oh! pour nous en aller en des tourbillons

6+5/1+5+5

15

12. Et que tout sʼexalte et que tout sombre et que nous tournoyions,

5+4+6

2

13. Vaisons

2

14

14.Dans Textase de ces inextinguibles horizons!

3+7+4

Preliminary scansional note: I have counted ‘e's according to traditional principles. This strategy seems to be demanded by the poem's subject if nothing else: the articulated ‘e’conveys the supple and lilting undulations of the dance. And where an articulated ‘e’ follows an accentuated syllable I have practised coupes enjambantes, partly because this is the conventional method and partly because they, too, contribute to the dance, to its seamless continuity. Finally the monorhymed stanza used here is not, for Dujardin, an eccentric display of virtuosity; all poems in La Comédie des amours are written in monorhymed stanzas.

This poem is about the tension or interplay between the centripetal and centrifugal forces at work in the waltz. Against the gyroscopic stasis of ‘girement’, ‘tourbillons’, ‘tournoyions’, of the insistent rhymes in /Ʒ/ (stanzas 1 and 4, see also ‘sombre’, l. 12) which make the movement of the poem as a whole circular thanks (p.151) to their appearance at its beginning and end, and of the first-person plural form of the imperative, binding the couple together in an in-turned rotation, is set the dispersive movement, the out-turned dynamic, of words such as ‘virant’, ‘vague’, ‘vole’, ‘extravague’, and, indeed, as this last word intimates, of all other words which combine /εks/ or / εgz/ or / εsp/ with a following /α/ or /a/ (i.e. ‘espace’, ‘sʼexalte’, ‘extase’, and relatedly ‘inextinguibles’). This same tension is to be found in line 11, in the juxtaposition of an ‘en’ which signifies movement away (‘nous en aller’) and an ‘en’ of withinness (‘en des tourbillons’); and more imbricatedly it is to be found in the single word ‘incantement’, which is presumably a portmanteau coinage, a blend of ‘incantation’, the ritualized, hypnotic monotony of the dance's rhythm, and of ‘enchantement’, that ‘chant’ which, through its spell-bindingness, ‘Entraîne lʼesprit hors lʼespace et le temps’. Ultimately the poem Opts’ for the centrifugal; although the final stanza brings a return of the /δ/ rhyme, the predominant impulse is away, either laterally, in ‘extase’ (‘displacement’, ‘standing outside’) and towards the inextinguishable multiple horizons, or perpendicularly, since the gyrations of the dance have produced a vertical axis which allows escape upwards (‘sʼexalte’) or downwards (‘sombre’).

In examining the relationship between the poem's prosody and this tension between centripetality and centrifugality, I would like to take up the notion of ‘standing outside’ (‘extase’) in connection with Dujardin's intraphrasal accent. In line 2 the ‘ecstatic’ exclamation ‘Oh!’ can either be treated as a measure on its own, bearing an accent of both intensity and duration:

                      Theories and Enquiries

or as a word within a tetrasyllable, attracting an intraphrasal accent of intensity:

Óh! dans les claȉrs | salóns

But either way, viewed in the perspective of the line as a whole, it is an accentuated element within the hexasyllable, a hexasyllable which is reinforced by the hexasyllable of the first line. In the second stanza the prosodic status of the line-initial accents on ‘Lʼâme’(1. 5) and ‘Vole’(1. 6) is very different. This second stanza is unequivocally about centrifugal activity, a fact underlined both by the impair lines and by the eccentricity of the stanza's (p.152) rhyme-sound and rhyme-gender—all other stanzas have nasal masculine rhymes (/δ/ or /ã/). Here, then, the line-initial accents in lines 5 and 6 are not assimiliated, as intraphrasal accents, into hexasyllables, but occur outside the boundary of the hexasyllable as extraphrasal accents. Looked at another way ‘Lʼâme’ and ‘Vole’ are the extra measures which transform the parisyllabicity of the hexasyllable into the unstable ‘falseness’ of the impair. The prosody of this stanza thus itself encodes a centrifugal process. And what more fitting perhaps than that the ‘Oh!’ at the beginning of line 11 is prosodically so ambiguous, for we can either regard this as bearing an intraphrasal accent within an initial hexasyllable or as an initial self-sufficient measure ‘standing outside’ the decasyllabic configuration of 5+5 which follows it. Given the dominant centrifugality of this stanza, the latter reading is more likely, though the hexasyllable still continues to play a diminishing role (see l. 12). I hardly need add that all instances of the impair in this poem, and the one instance of a parisyllabic line (l. 14) which moves into the uncharted rhythmical territory beyond the dodecasyllabic limit, are intimately linked with the uncontrolled and dispersive dynamics of the centrifugal tendency; indeed the parisyllabic fourteener might itself be read as a combination of impairs, a trisyllable and an eleven-syllable measure.

The above observations lead me to the conclusion that Dujardin is much more a free-verse poet of the line than of the individual ‘pied rythmique’, and in this sense he belongs more with Laforgue and Henri de Régnier than with Kahn. This is not entirely so, because his long lines, the fourteen-syllable line 14 and the fifteen-syllable line 12, clearly derive from the accumulation of individual, separate syntactic/rhythmic units; this is particularly clear in line 12, which is made up of the triple incidence of the optative subjunctive each time introduced by the ‘et que’ construction. But with these lines, as with the hendecasyllabic ones, it could hardly be otherwise, since, as we have earlier observed, tradition has provided them with no lineal status; they are by nature rhythmic hybrids and cumulative products. Everywhere else, though, Dujardin seems to have a very precise apprehension of the linear span and seems equally to have made the creation of certain overall syllabic shapes his prime objective. Apart from those factors already adduced in the previous paragraphs, there are four other reasons for making this claim. (1) Dujardin's traditional (p.153) treatment of the unelided ‘e’—and we have encountered no cause to change our minds about this—tends, when it appears, to bind the different elements of the line together into a single rhythmic enunciation. (2) This tendency both reinforces, and is reinforced by, Dujardin's interest in the intraphrasal accent, because any measure-terminal accent is susceptible to transformation into an intraphrasal one; in other words the very notion of the intraphrasal accent makes any measure available to expansion, to a more extensive reading; I have tried to suggest something of this rhythmic elasticity in the alternative segmentations I have provided in my tabulation of measures. (3) Although the line-internal ‘music’ of the poem is obtrusive and highly developed, the very insistence of the rhyme in these monorhymed stanzas endows the line-ending with the same kind of prominence it enjoys in regular verse; as in regular verse, there is a very real sense in which Dujardin's lines gravitate towards, are orientated towards, their ends. (4) This poem, for all the variation in its line-lengths, is dependent upon the ‘fineness’ of its lines for its overall formal integrity, for we cannot help but be struck by the fact that it has fourteen lines and constitutes a ‘sonnet renversé’, a variation of sonnet structure to be found, arguably, in Baudelaire's ‘Bien loin dʼici’ and quite clearly in Verlaine's ‘Sappho’, to name but two examples.

That we are right to make this last identification is borne out by Dujardin's practice in his later collection Le Délassement du guerrier (1904), where we find several free-verse sonnets, the odd example of free-verse terza rima (‘Un jour, un rêvé chevalier’, ‘Vénus génitrix’), a free-verse villanelle (‘Vous qui mʼavez aimé’), a free-verse rondeau (‘Vous êtes femme, et dans mon âme’) and this poem:

No. of syllables

Scansion

3

Oui, je sais

3/1+2

10

Que vous êtes une petite reine

3+5+2

6

De naissance lointaine,

6/3+3

7

Enfant dʼobscures forêts.

2+5/2+2+3

6

Quels inconnus relais

6/4+2

9

Mènent en cette terre incertaine,

6+3/1+5+3

3

Oh! je sais

3/1+2

10

Que vous êtes une petite reine.

3+5+2

8

Voire primordial palais

6+2

8

Cʼétait quelque roche hautaine

5+3

8

Un champ pâle de marjolaine

3+5

9

Une lande dʼirréels genêts …

3+6/3+4+2

3

Oui je sais.

3/1+2

(p.154) This is a free-verse version of the standard thirteen-line rondel (ABba | abAB | abbaA, where the capitals denote refrain lines).3

In noting that Dujardin published a cycle of five regular rondels under the title of ‘Jeunes Filles’ in the 29 May–3 June 1886 issue of La Vogue—Charles Vignier had also published a regular rondel in the previous issue (13 May)—we cannot avoid asking the question which, had it an obvious answer, would have been more appropriately put in section (iii) of Chapter 2, namely: what is the connection between fixed forms and free verse? Why is it that the poets of vers libéré indulged in fixed forms (I think particularly of Charles Cros's Triolets fantaisistes’ and Corbière's irregular ‘Rondels pour après’)? How is it that Mallarmé, the poet of Un coup de dés, wrote five triolets and four rondels? What led Rimbaud to write the sequence of triolets that constitute ‘Le cœur volé’? Bearing in mind the free verse of Serres chaudes (1889), how do we account for Maeterlinck's triolet sequence of 1883 ‘Dans les joncs’? How is it that Jean Ajalbert's impressionist collection Sur le vif (1886) contains rondels, when he was to publish the free verse ‘Sur les talus’ the following year? How is it that Stuart Merrill's first free-verse poems (Les Quatre Saisons, 1900) were preceded, though not directly, by two collections (Les Gammes, 1887, and Les Fastes, 1891) in whose pages we find examples of the triolet, the pantoum, the sestina, the villanelle, besides the sonnet and terza rima? Were these the last convulsions of a hypertrophied regular verse, consuming itself in its own artificial difficulties, on the eve of the outbreak of free verse? Were these the ultimate standards of regular verse, ironically and expressly subscribed to by the poet, the better to demonstrate his subsequent distance from them (there certainly seems to be something of this motive in Dujardin's particular case)? Or was the complicity between the fixed form and free verse greater than we are accustomed to suspect? After all, the two-rhyme fixed forms (rondel, rondeau, triolet, villanelle) do in a sense exceed what (p.155) regular verse normally allows in terms of rhyme saturation, and we have found rhyme saturation to be a characteristic of the free verse of both Alain-Fournier and Dujardin (as it is of the free verse of Laforgue, to quote another example); furthermore, given the presence of refrain in these two-rhyme fixed forms, they are bound to introduce repetition into their rhyme-sequences (the rentrement of the rondeau is, in fact, a rhymeless element, or rather it only rhymes with itself). And even though these forms are in their overall patterns fixed, is it not true to say that several of them are heterostanzaic (sonnet, rondeau, rondel), that many, looked at from the perspective of regular verse, are free-rhyming or, at least, irregularly rhymed (sonnet, rondeau, rondel, triolet), that all of them, considered from the point of view of the ‘innocent’ reader, enjoy a high degree of structural unpredictability? I have yet to encounter any comments made by a, verslibriste on this subject, apart from Vielé-Griffin's remark: ‘Bref, à la forme fixe, nous opposâmes la forme mobile, à lʼattitude, le geste, à la statique, le mouvement—à la mort, nous opposâmes la vie’ (quoted by Cours 1930: 17). It is doubtful whether he has fixed forms themselves in mind when he refers to ‘la forme fixe’, so his declaration is only potentially relevant. In fact fixed forms are remarkably animate, mobile, unsettled, particularly when their structural eccentricities are combined with the mercurial octosyllable, which they so frequently are (though not in the sonnet; and the villanelle has a special relationship with the heptasyllable4). At first sight fixed forms and free verse seem to be diametrically opposed; but they are both, in their own ways, conspiracies against regular verse.

(iv) Jules Huret, Enquête sur l'évolution littéraire (1891)

On 3 March 1891, Jules Huret introduced his ‘Enquête sur l' évolution littéraire’ in the daily paper Lʼécho de Paris. There followed in subsequent issues, over the next five months, a series of interviews with literary personalities of the day, which were (p.156) published in book form the year after. The literary events which immediately provoked the enquiry were the publication, in December 1890, of Moréas's Le Pèlerin passionné, the banquet given in Moréas's honour in February 1891, and the publication, early in the year, of Barrès's novel Le Jardin de Bérénice. Huret's work is interesting not only as a series of valuable literary testimonies, but as a specimen of fashionable literary Darwinism, that kind of criticism that busied itself with the classification of species (genres, schools, trends), their heredity, and an assessment of their chances of survival. Part of Huret's enterprise was to map out the territories of these cohabiting species so that their conflicts and evolution might be better understood. Upon his initial broad classification—‘Les Psychologues’, ‘Les Mages’, ‘Symbolistes et Décadents’, ‘Les Naturalistes’, ‘Les Néo-Realistes’, ‘Les Par-nassiens’, ‘Les Indépendants’, ‘Théoriciens et Philosophes’—Huret, in his ‘Avant-propos’, superimposed another, based on the temperaments, the ‘attitudes dʼesprit’, of his interviewees, thus ‘Bénins et Bénisseurs’, ‘Acides et Pointus’, ‘Boxeurs et Savatiers’, and so on.

Huret had some reason to be dissatisfied with his results. The large questions to which he had sought answers (e.g. Was there a link between the ‘Psychologues’ and the Symbolists which would correspond to a link between the Naturalists and the Parnassians? Did the Parnassians consider the Symbolists as disciples or rebels? Would the Naturalists admit to a decline in their literary fortunes?) remained largely unanswered. His interviewees seemed to have little taste for a discourse conducted at this level of abstraction. But after all, there were compensations:

De là ce résultat inattendu, mais dont je me félicite pour les lecteurs (car il faut bien me féliciter de quelque chose), que si mon enquête nʼoffre pas à lʼhistoire littéraire de théorisations suffisantes, elle révèle à lʼhistoire générate les passions foncières, les dessous dʼesprit, les moœrs combatives dʼun grand nombre dʼartistes de ce temps. (1982/1891: 24)

although it is precisely the discovery of a Darwinian law of the jungle among the so-called defenders of the spiritual life that left Huret with a dispiriting after-taste:

Si les déprimantes constatations lʼemportent, cʼest hélas! que le métier littéraire nʼéchappe pas à la loi féroce de la concurrence vitale, et que là, comme en toute carrière, les intérêts matériels priment et tyrannisent les appétences spirituelles. Force mʼétait done de noter, sous les apparences (p.157) hautaines dʼune lutte pour lʼart, les âres et douloureuses et basses nécessités de la lutte pour la vie. (p. 29)

And Huret's index, acting rather like a primitive concordance, reveals the fluid relativity of literary history, the way in which issues personal and theoretical, as much as issues of literary ‘merit’, generate personalities, the speed with which an old order passes; reveals, in short, how misleading a consecrated ‘long view’ of literary history can be:

Cʼest ainsi:

Que M. Mallarmé, dont la haute personnalité littéraire ne se révèle que les mardis soir à quelques personnes choisies, a pourtant groupé plus de nominations que Victor Hugo, la plus populaire des gloires de la France moderne; encore faudrait-il ajouter que, sur les quarante-trois citations du poète national, dix au moins lui viennent de M. Auguste Vacquerie, son executeur testamentaire …

Que M. Jules Laforgue, probablement parce quʼil est mort, a recueilli 18 mentions, tandis quʼÉmile Hennequin, parce quʼil est mort aussi, nʼen a recueilli que 3, alors quʼil y a deux ans à peine, il était unanimement reconnu le plus considérable des jeunes …

Que si M. Jean Moréas, avec ses 54 voix, dépasse ses maîtres Verlaine et Mallarmé, il paraît le devoir aux appréciations généreusement défavorables de ceux de ses camarades qui lui ont offert un banquet en février dernier, (p. 25)

The Enquête was, then, a salutary reminder of the fierce partisanships and self-regarding whims which motivate literary opinion.

But at least one might have predicted that the strongest opposition to vers libre would come from the Parnassians, and this was indeed so. Aligning themselves behind Leconte de Lisle, Catulle Mendès, Heredia, François Coppée, Sully-Prudhomme, Armand Silvestre, and Edmond Haraucourt expressed, with varying degrees of indignation and derision, their view that vers libre derived from a profound misunderstanding of the durable flexibility of traditional versification. We have already had occasion (see above, p. 84) to quote Leconte de Lisle's insistence that parisyllabicity is the very life-blood of French verse, and he also asserts that vers libre is no more than rhythmic prose and that it promises only aesthetic anarchy. We have had occasion, too, to quote Heredia's opinion that vers libre is part of an international prosodic conspiracy (see above, p. 100)—though in an article in (p.158) the Entretiens politiques et littéraires (June 1891) Vielé-Griffin notes Heredia's pride in his ‘sang hispano-cubain’, Leconte de Lisle's in his ‘origine australe’, and Coppée's in the ‘avatares beiges de son patronomique’(sic)—and that the alexandrine is already the adequate ‘vers polymorphe’ (see above, p. 77). Both Mendès and Silvestre reiterate the racial argument, and Mendès, Coppée, and Haraucourt all subscribe to the view that there are no kinds of verse-music that the alexandrine cannot produce. Heredia returns to Leconte de Lisle's implicit charge, that vers libre is in fact rhythmic prose with a typographical face-lift, but concedes that it might have some part to play in lyrical drama where it will not have to rely on its own music (if it has any), but can be supported instrumentally. This apparent absence of music worries Silvestre, too, though he acknowledges that habit may prevent him from perceiving it. Such, in broad outline, are the objections of the Parnassians; but there is one more, formulated by Mendès, which deserves to be quoted, since it bears so directly on the difficulty confronting us:

‘Les Symbolistes ont cru inventer, dit M. Achille Delaroche, un vers, une strophe dont lʼunité fût plutôt psychique que syllabique, et variable en nombre et en durée selon les nécéssités musicales’.

Voilà bien, nʼest-ce pas, la théorie de ce que certains poèts nouveaux appellent le vers libre? Eh bien! jʼai une crainte: comment le lecteur, vous, moi, nʼimporte qui, sʼy prendra-t-il pour découvrir le rythme de cette strophe, ‘plutôt psychique que syllabique’, et comment en sera-t–il touché? Oui, où trouvera-t-il le point de repère qui lui permette de suivre le rythme choisi par le poéte? (Huret 1982/1891: 248)

For the verslibriste anxious to maintain the inimitability and inwardness of his own utterance at the same time as the reader's purely intuitive perception of it, no scansion may seem appropriate to free verse. But for Mendès this would be to make the business of communication much too haphazard and unreliable. Besides it requires too much to be taken on trust and leaves the initially unresponsive reader with no way of establishing contact with the text. But to agree that this is so does not make it any easier to decide what an appropriate scansion would be. This is the very problem we are in the process of investigating.

The exemplary solidarity of the Parnassians was in no way matched by those who might have been expected to champion free verse. In a sense this is not surprising given that, by definition, (p.159) there are as many versions of free verse as there are poets practising it; but even those who defend it in principle frequently do so only in passing and only in the most general terms. Jules Bois, for example, with his eye more firmly fixed on the issue of occultism, has only this to say:

Comme les casuistes sur un point de théologie, les contemporains se chamaillent sur lʼalexandrin et le vers libre. Sans un autoritarisme vain, le rythme ne saurait être fixé pas plus dans le désordre que dans la règie. Chacun selon son sujet et sa forme dʼesprit, adopte ou se crée un rythme. (Huret 1982/1891: 68–9)

Bois was one of the very few contributors to Huret's enquiry also to appear in Marinetti's Enquête Internationale sur le vers libre (1909) (see next section), where, as we shall see, his continuing support for vers libre is subject to important qualifications. Henri de Régnier's comments are equally brief and concern themselves predominantly with the stanza, whose lines are like the varied reverberations of a single image, idea or feeling:

—Sur la technique du vers, quel est votre avis?

—La liberté la plus grande: (quʼimporte le nombre de vers, si le rythme est beau?) lʼusage de lʼalexandrin classique suivant les besoins; la composition harmonieuse de la strophe, que je considère comme formée des échos multipliés dʼune image, dʼune idée ou dʼun sentiment qui se répercutent, se varient à travers les modifications des vers pour sʼy recomposer. (Huret 1982/1891: 101)

It should be said that throughout his poetic career Henri de Régnier remained reticent about the theoretical dimension of free verse; his brief preface to his Choix de poèmes(1932), whose principal paragraph we have already quoted (see above, p. 18), constitutes, to all intents and purposes, the totality of his theoretical output. Mallarmé, for his part, takes a long view of the matter: the present prosodic instability reflects a social instability; the alexandrine needs a rest from its protracted labours, though it still has an irreplaceable function ‘dans des moments de crise de lʼâme’ (Huret 1982/1891: 75); the new departures in verse-art correspond to new developments in music—‘aux mélodies dʼautre-fois tres dessinees succede une infinité de mélodies brisées qui enrichissent le tissu sans quʼon sente la cadence aussi fortement marquée’ (p. 75); above all, perhaps, our perception of rhythmicity (p.160) has become more inclusive, more subtle: all stylistically motivated language has a versification, to a greater or lesser degree, and thus all forms of language are coterminous:

Le vers est partout dans la langue où il y a rythme, partout, excepté dans les affiches et à la quatrième page des journaux. Dans le genre appelé prose, il y a des vers, quelquefois admirables, de tous rythmes. Mais en vérité, il nʼy a pas de prose: il y a lʼalphabet, et puis des vers plus ou moins serrés, plus ou moins diffus. Toutes les fois quʼil y a effort au style, il y a versification, (pp. 74–5)

The idea of a prose which is no more than a convenient medium for the embedding of different kinds of verse we have already explored in our section on the prose poem (Chapter 2, section (iii)(e)); we have also seen Dujardin converting his original free-verse conception of ‘A la gloire dʼAntonia’ into a suite of prose poems; and we might add what Maeterlinck confides to Huret, namely that his play La Princesse Maleine (1890) was ‘des vers libres mis typographiquement en prose’ (Huret 1982/1891: 127).

But if on the whole the verslibristes and their sympathizers did not put their case with the strength that it deserved, if the resistance of the Parnassians was well co-ordinated, if the unpredictable Verlaine turned out to be a real fly in the ointment—‘pour quʼil y ait vers, il faut quʼil y ait rythme. A présent, on fait des vers à mille pattes! Ça nʼest plus des vers, cʼest de la prose, quelquefois même ce nʼest que du charabia … Et surtout,ça nʼest pas français, non, ça nʼest pas français’ (p. 83)—support also came from unexpected directions, from Anatole France, for whom classical precepts had been so much eroded or so thoroughly hoisted on their own absurdities, that a new prosody was inevitable:

Je trouve que, lorsque Coppée a écrit un vers conformé de cette façon (je ne vous réponds pas des mots, mais de la contexture):

Je suis la froide et mechante souveraine

cʼest–à-dire depuis quʼil a mis la césure entre lʼarticle et le substantif, il a supprimé lʼhémistiche classique, et quʼon avait absolument le droit de mettre la cesure au milieu dʼun mot. Quant à la règie de lʼhiatus, on ne peut la trouver que bête, quand on pense quʼelle défendait de dire dans un vers: tu aimes, et quʼelle autorisait dʼécrire: au haut de lʼescalier! Les autres entraves, lʼélision de lʼe muet prescrivant [sic]: je prie Dieu, lʼalternance des rimes, etc., etc. … ne se soutiennent pas davantage. (p.161) Reste le nombre illimité des syllabes … je sais bien que cela équivaut a faire de la prose rythmée, mais quand elle est faite par un artiste, jʼy vois assez de charme. (p. 36)

But Huret was right to complain about the general absence of theory in the replies he received. Apart from the contribution of Kahn, dealt with in section (i) of this chapter, apart from Ghil's seizing the opportunity to canvass further for his ‘philosophic évolutive’ and ‘instrumentation verbale’—‘et de même que pour rendre un état dʼingénuité et de simplesse, par exemple, il ne voudrait pas évidemment des saxophones ou des trompettes, le poète instrumentiste pour ceci évitera les mots chargés dʼO, dʼA et dʼU éclatants’ (p. 115)—only Moréas himself offers remarks which seek to justify a practice theoretically. Confronted by Huret with the line of twenty-one syllables from ‘Épître’:

Et mes litières sʼeffeuillent aux ornières, toutes mes litières à grands pans

and anticipating the charge that this is in fact rhythmic prose, Moréas defends himself by remarking that Hugo's verse was often printed as prose by those of an anti-Romantic persuasion, and by reserving his right as a poet to make use of any verse-line as he sees fit. Besides, he continues, the emergence of free verse is no accident. And he goes on to supply a brief history of the alexandrine, in which, rather in the manner of Kahn, he maintains that the alexandrine couplet was in fact a quatrain of hexasyllables and that the two hemistichs of the alexandrine were frequently without any semantic connection. After reviewing Hugo's innovations and Verlaine's fragmentation of the alexandrine, he comes to this conclusion:

Paul Verlaine a respecté toutefois le nombre syllabique de douze. Respect illusoire, car, à la vérité, un alexandrin non fixement césuré nʼest quʼun octosyllabe allongé, les syllabes muettes prenant une valeur relative. Suite de cela: un lacis de vers inégaux déterminant leur réciproque quantité syllabique pour constituer la strophe. Cʼest autant lʼintuition de ce principe quʼun instinct auquel je me fie qui me révéla le système rythmique définitif de telles de mes pièces, comme Agnè et Galathée. (Huret 1982/1891: 90)

This passage, with its notion of the ‘octosyllabe allongé’, is very close to the paragraph on prosodic matters which appears in the preface to Le Pèlerin passionné and of which we have already (p.162) made mention (see above, p. 82), with two significant additions: the references to the ‘valeur relative’ of mute syllables and to the reciprocal influence of successive lines in determining syllabic quantity.

Since, in many respects, Moréas's Le Pèlerin passionnéis one of the principal subjects of Huret's enquiry, I would like to close this section with an analysis of one of its poems, paying particular attention to the two features we have just isolated in Moréas's reply. I have chosen for this purpose one of the four poems which appeared in La Vogue, 3/4 (15–22 Nov. 1886), under the title of ‘Lais’ and which reappeared, with considerable revision, in Le Pèlerin passionné under the titles ‘LʼInvestiture’, ‘Chanson’ (‘Les courlis dans les roseaux!’), ‘Chanson’ (‘On a marché sur les fleurs’), and ‘Historiette’; before examining this last, I give both versions of it, the La Vogue text first:

A. No. of syllables

Scansion

15

1. De sa hache, de sa lance, de sa hache ah quʼil Est las,

3+4+4+4

13

2. Le Chevalier-aux–blanches armes ah quʼil est las!

4+5′+4

9

3. A coups de hache, à coups de bastarde

4+5

8

4. Rompre des targes et des casques

4+4

12

5. Parmi les Alamans, parmi les Arimaspes,

6+6

13

6. Le Chevalier-aux–blanches armes ah quʼil est las!

4+5′+4

9

7. Et de la jolie fille de Perth,

6+3

8

8. Et de Béatrix et de Berthe,

5+3

10

9. Et des robes a bordure de perles

3+7/3+4+3

11

10. Et des cheveux sur le cou—ah quʼil est las—

4+3+4

11

11. Et des bras autour du cou—ah quʼil est las—

3+4+4

13

12. Le Chevalier-aux–blanches armes ah quʼil est las!

4+5′+4

10

13. De mourir, de mourir ah quʼil est las

3+3+4

13

14. Le Chevalier-aux–blanches armes,—ah quʼil est las!

4+5′+4

B.

7

1. De sa hache—ah quʼil est las —

3+4

8

2. Le chevalier aux blanches armes.

4+4

4

3. A coups de hache

4

9

4. Rompre des casques,—ah quʼil est las—

5′+4

8

5. Le chevalier aux blanches armes.

4+4

9

6. Et de la jolie fille de Perth,

6+3

8

7. Et de Béatrix et de Berthe,

5+3

10

8. Et des robes à bordures de pedes

3+7/3+4+3

11

9. Et des cheveux sur le cou—ah quʼil est las—

4+3+4

11

10. Et des bras autour du cou—ah quʼil est las—

3+4+4

8

11. Le chevalier aux blanches armes.

4+4

7

12. De mourir,—ah quʼil est las—

3+4

8

13. Le chevalier aux blanches armes.

4+4

Preliminary scansional note: The fact that Moréas does not object to Huret's calling the line from ‘Épitre’ a twenty-one–syllable line implies that it is right, as a first move, to count his ‘e's in the traditional way. I have treated post-accentual unelided ‘e's as coupes lyriques only where the syntax seems to demand it; I have not done so in lines 1 and 4 of A because, it seems to me, Moréas wants us to read these lines as headlong rushes (particularly 1. 1); it is the unremittingness of military activity that tires the knight, the dizzying speed of change, as much as any loss of spiritual commitment.

(p.163) I do not wish to delay over Moréas's medievalizing lexicon on which he so much prided himself and which was to be a first step in his creation of ‘lʼécole romane’—‘Je suis le seul à réclamer un renouveau de la langue poétique, un retour aux traditions, un style retrempé aux sources de lʼidiome roman’ (Huret 1982/1891: 90); ‘Pour qui sait, dans notre litterature medievale un riche heritage se recèle. Ce sont les grâces et mignardises de cet âge verdissant, lesquelles … nous constitueront … une langue digne de vêtir les plus nobles chimères de la pensée créatrice’ (Moréas 1891: iii). I only wish to point out the obvious, that in transforming A into B Moréas has jettisoned much of his archaizing local colour, his references to the ‘bastarde’ (medieval cannon), ‘targes’ (shields), ‘les Alamans’ (Germanic tribal confederation defeated by Clovis at Tolbiac in 496), and ‘les Arimaspes’ (a legendary one-eyed people of Scythia who fought the Griffons for possession of the gold to be found in Scythia's principal river, the Arimaspios). Moréas's revisions for the ‘Chanson’‘Les courlis dans les roseaux’ reveal a similar removal of meretricious detail. But this is part of Moréas's more general tendency in his revisions to de-rhetoricalize and de-historicize his text, to make it more subdued, more withdrawn, barer. By removing the capital and hyphens from ‘Le Chevalier-aux–blanches armes’, Moréas has discouraged us from trying to identify a particular knight and from investing him with an amplified and exemplary status. By making sure that the exclamation ‘ah quʼil est las’ occurs in the line immediately preceding ‘Le chevalier’, rather than succeeding him in the same (p.164) line, Moréas makes the knight the mute undergoer of his condition, an existential terminus, imprisoned, as it were, in the wearying destiny of his armour, rather than a figure who can expend his despair in a cry, transform a condition into an utterable feeling. By dispensing with the exclamation mark after ‘ah quʼil est las’ and always parenthesizing this phrase, Moréas discourages us from regarding it as something openly uttered, springing directly from activity, luxuriating in its own pathos; instead it moves into the dimension of the undertone, an undertone which interrupts text with inwardness, a confession which the knight's very code compels him to suppress. And by reducing the incidence of repetition, Moréas correspondingly reduces the élément of the possessed and rhapsodic in favour of the unaffected and undemonstrative.

These changes are of course closely allied with the generical change. The ‘lai’, with its variable line-length, its variable number of stanzas, its customary association with narrative, its frequent flirtation with the supernatural, and with the world of dream (see Deloffre 1969: 52), yields to the ‘chanson’, with its usually short sequence of identically structured stanzas, its lack of the learned, the intricate, and the other-worldly, and its dominant interest in affairs of the heart; is this why, of all his cultural and historical references, Moréas has most conspicuously retained Scott's fair maid of Perth, Béatrix (Dante?), and Berthe (au grand pied? de Bourgogne? de Hollande?)? It may seem strange to insist on the prosodic difference in this generical change, since Moréas seems only to have, as it were, shifted his syllabic range by an octave—in A the lines vary between 8 and 15 syllables, while in Bthey vary between 4 and 11, but in both there is a 7-syllable degree of flexibility. But this apparently cosmetic change conceals, I believe, a more radical shift, from a free verse of the measure, and in this case of a tetrasyllable leitmotif or constante rythmique, to a free verse of the line. Let me explain.

The second of the ‘Lais’ published in La Vogue, beginning with the line ‘Le train express sous le ciel de Véronèse’, has the following syllabic quantities over its three stanzas: 11.15.19.22.7 / 17.15.28.25.7 / 15.17.17.7; the 7 at the end of each stanza is the refrain ‘Ô vous jolie fée des eaux!’ There are some instances of repetition and rhyme at the line-endings, but these are few and unpatterned. In the revised version, the ‘Chanson’ (‘Les courlis (p.165) dans les roseaux!’) of Le Pèlerin passionné, on the other hand, we find three quatrains, on two rhyme-sounds only, all in heptasyllables, each stanza sharing the same second and fourth lines and the same syntactic structures; I quote the first and third stanzas:

  • Les courlis dans les roseaux!
  • (Faut-il que je vous en parle,
  • Des courlis dans les roseaux?)
  • Ô vous joli’ Fée des eaux.
  • . . . . . .
  • Mon cœur pris en vos réseaux!
  • (Faut-il que je vous en parle,
  • De mon cceur en vos réseaux?)
  • Ôvous joliʼ Fée des eaux.
From the La Vogue version, written in a free verse of the measure, Moréas has, it seems, lifted instances of the dominant heptasyllabic leitmotif, created new heptasyllables where necessary (‘Mon cceur pris en vos réseaux!’) and put them together in a repeated isosyllabic structure. A similar process is apparently at work in his reworking of the first of the ‘Lais’, to become ‘LʼInvestiture’ in Le Pelerin passionné: here a stanzaic/syllabic structure of 16.9.14.7 / 18.21.19.8 / 17.13.27.8 is converted into 9.9.9.9.7 / 8.9.8.8 / 8.8.8.8.8; the ‘Lai’ has three dominant measures, a heptasyllabic configuration (the repeated ‘au déclin de la Grande Ourse’, ‘La fleur nommee asphodele’, ‘tu regarderas mes yeux’, ‘mes yeux auront la couleur’), an octosyllabic configuration (the repeated ‘immuablement taciturne’, ‘comme le rocher de la fable’, the repeated ‘De la fleur nommee asphodèle’), and an enneasyllabic one (‘Nous longerons la grille du pare’, ‘Et tu porteras — car je le veux’, ‘parmi les bandeaux de tes cheveux’); as before Moréas has ‘lifted’ a selection of these, but has introduced greater syllabic consistency in ‘LʼInvestiture’ by converting ‘au declin de la Grande Ourse’ into an enneasyllable (Ά lʼheure ou la Grande Ourse decline’) and ‘mes yeux auront la couleur’ and ‘tu regarderas mes yeux’ into octosyllables (‘Et mes yeux auront la couleur’, ‘Tes yeux regarderont mes yeux’). In other words, in ‘LʼInvestiture’Moréas has produced a ‘chanson’ in which, after some initial syllabic vacillation, dominated by the enneasyllable, an octosyllabic consistency is achieved; the ‘chanson’ formally comes home to itself.

(p.166) What these findings lead me to propose about ‘Historiette’ is this: in transforming his ‘lai’ into a ‘chanson’, Moréas has once again transformed a free verse of the measure into a free verse of the line, where the line in question is the octosyllable, and where octosyllabic consistency is owed to the ‘valeur relative’ of mute syllables, itself détermined by the reciprocal influence of successive verse lines. In short, ‘Historiette’ is ‘un lacis de vers [ostensibly] inegaux determinant leur réciproque quantité syllabique pour constituer la [syllabically consistent] strophe’. We need therefore to envisage a scansion for ‘Historiette’ very different from the one we have already presented and bringing into full play the various responses to the e atone (with their concomitant scansional consequences) enumerated in Chapter 1. Thus:

1. Soudainement contre les vitres

4+4

2. lʼaverse sʼest jetée

3′+3/2+4/6 or 2+3

3. comme un vol dʼoiseaux en fuite

5 + 2/3+2+2

4. buté

2

5. soudainement contre les vitres.

4+4/8

6. Un assaut de bees crible le verre

5+4/3+2+1+3/5+1+3

7. et les bees fragiles se brisent

6′+2/5+3 or 5 + 2

8. et dʼautres viennent …

3′+1/2+2/4 or 2+1

9. —et lʼon nʼose ouvrir la chaumière grise

5+5/5+3+2

10. de crainte de les faire taire.

2+6

11. Ce sont des bees sonnant clair

4+3

12. des bees de cristal harmonique

5+3/2+6/2+3+3

13. qui frappent;

2

14. cristal contre cristal,

2+4

15. les vitres sonnent et les bees sonnent,

5′+4/4+5 or 4+4

16. double musique:

4

17. à la vitre

3

18. la pluie parle.

3

19. Or les gouttes

3

20. au heurt

2

21. sʼétalent,

2

22. coulent en longues larmes douces,

2′+6/1+7 or 1+6

23. comme un deuil le long dʼune jeune face …

3+7

24. Telle une destinée qui passe,

6+2

25. aux carreaux ingénus de lʼobscure demeure

6+6/3+3+3+3

26. la pluie tour à tour chante et pleure

5+3/2+3+1+2

Preliminary scansional note; Kahn's underlying object in treating measure-terminal unelided ‘e's in the same manner as line-terminal ones is not just a manoeuvre whereby the enunciation of vers libre can be brought more into line with current pronunciation; it is also, and more importantly from a prosodic point of view, a way of creating a clear-cut boundary between one measure and the next. Kahn's wish to ensure a rhythmic independence for the individual measure means that he can have no truck with the coupe enjambante, a feature which, as we saw in our analysis of Dujardin's ‘Void les violons’, contributes to the extension of the rhythmic phrase beyond the accentual boundary, and thus to the establishment of the overall line of verse as the fundamental rhythmic perspective. It is no accident, I think, that the couplet from Racine's Athalie which Kahn chooses to segment rhythmically, contains no problematic, post-accentual unelided ‘e's. Now it is clear that Ghéon wishes tout simply marking a pause in its place) or by treating each measure-terminal unelided ‘e’ as a coupe lyrique, which would achieve the same end. For this reason, in those lines in which this problem arises (11. 2, 7, 8, 22), I have provided three types of scansion, a scansion with the unelided ‘e’ treated as a coupe lyrique, a scansion with the unelided ‘e’ counted as a coupe enjambante (the least likely reading), and a scansion with the unelided ‘e’ not counted, in the Kahnian manner.

(p.167) Before drawing any conclusions, we should note that the two hendecasyllabic lines (9, 10) remain in place; these might be regarded as octosyllables interrupted, or interfered with, by a trisyllabic measure; what shakes these two lines out of the octosyllabic mould is their vestigial erotic charge; the knight, for all his weariness, is still haunted by images of sensual enticement and the rhythm momentarily deviates out of its settled pattern to register the possibility of a fuller experience. The isolated tetrasyllable of line 3 is, I take it, unproblematic, an incipient octosyllable perhaps presupposing an absent ‘ah quʼil est las’.

If my second scansion of ‘Historiette’ seems plausible, and justified by the prosodic tendencies discernible in Moréas's revisions of other ‘Lais’, then we must conclude this about the free verse of the line: the free verse of the line uses its proximity to regular verse to draw the reader towards a regular-verse reading and in so doing activates all or many of the subtle expressive nuances, of the enunciatory potentialities, concealed in problematic syllables, and particularly in the e atone. We recognize that uncertainty about the treatment of the e atone is one of the principal sources of free verse's syllabic indeterminacy, but we are inclined to assume, for the sake of our own scansional sanity, that each free-verse poet at least treats the e atone consistently; we have only to discover, by various kinds of non-definitive evidence, what principle is being adhered to in any given example. Morier's studies of the verse of Verhaeren, Henri de Régnier, and Vielé-;Griffin have demonstrated the groundlessness of that assumption. Conversely, therefore, we should recognize that a free-verse poet may persuade us to discover a syllabic regularity within sequences (p.168) of apparently irregular lines even before we have made decisions about the specific value of the e atone and precisely so that we have the authority, as it were, to treat the e atone in a whole variety of different ways in the same poem. The e atone remains for the verslibriste a syllable whose very relativity makes it a rich mine of expressive effects, but these can only be brought into play, in their diversity, if the reader has some warrant for unearthing them, and this warrant is provided, paradoxically, by the line's capacity to imply its own regularity. I have tried, in my second scansion of ‘Historiette’, to indicate what some of these expressive effects are. In converting ‘lai’ into ‘chanson’, a free verse of measure into a free verse of line, therefore, Moréas has converted a potentially consistent treatment of the e atone into an unmistakably inconsistent treatment of it. And he has done this in the interests of the greater inwardness of the ‘chanson’; the relativized e atone functions like an expressive subtext, revealing the less visible impulses and inner struggles of the knight. The relativization of the e atone is, in short, one of the means whereby free verse psychologizes itself.

In Huret's enquiry the verslibristes were an embattled minority, thanks to the scepticisms and incredulities of Parnassians and Naturalists alike, and thanks also to the fact that those who, as Symbolists, might have been expected to support free verse vigorously either expressed themselves lukewarmly or were more concerned with other issues (occultism, ‘instrumentation verbale’, the symbol, the nascent ‘école romane’). In Marinetti's Enquête Internationale sur le vers libre (1909), on the other hand, they had the stage to themselves, and the added voices of many who had not participated in Huret's enquiry (e.g. Verhaeren, Vielé-Griffin, Mauclair, Ghéon). But, as we shall see, verslibrisme had by this time, and for a variety of reasons, lost some of its crusading momentum.

(v) F. T. Marinetti, Enquête Internationale sur le vers libre (1909)

As we come towards the end of our chosen period of study, we find Kahn still firmly in the saddle as the presiding genius of free verse. When Marinetti set about organizing his international enquiry into the state of the art of free verse, in the pages of his Milan journal Poesia, in 1905, he referred there to Kahn as ‘il creatore del verso (p.169) libero’ (Feb. 1905), and, when all the replies had been collected and were published, together with the Manifeste du futurisme, in 1909, it was Kahn's response which, by an apt stroke of chance, appeared first (the replies were arranged in chronological order). And, as we shall see, what theoretical assertions there are among French contributions seem to have their source in Kahn's thinking, though its éléments have, in the hands of others, undergone something of a polarization.

Reading through the results of Marinetti's enquiry one cannot but be struck by their rather weary, desultory nature. No doubt, by this time, the nagging habit of literary enquiries, so prevalent in the late nineteenth century, was beginning to take its toll. But much of the steam had gone out of the Franco-Belgian verslibriste movement, thanks to a falling off of production in the early years of the new century, and a series of disaffections: Charles van Lerberghe loudly repudiated free verse after La Chanson dʼEve (1904); in the collections that followed Jeux rustiques et divins (1897), Henri de Régnier gravitated back to classical forms; Le Pèlerin passionné (1890) was Moréas's only foray into free verse, as was ‘Sur les talus’ (1887) for Ajalbert. Has this lull in free verse's existence occurred because it has failed to live up to poets’ expectations of it, because it runs against the grain of the French temperament, Jules Bois asks? ‘Pas précisément: il a simplement pris sa place réelle, moyen dʼexpression intermédiate entre le vers traditionnel et la prose, mais incapable de se substituer à lʼun ou à lʼautre’ (Marinetti 1909: 47). But if free verse is simply finding its level, defining its area of expressive usefulness, Bois is none the less of the opinion that free verse has been blessed neither with an authoritative theorist—in his view recent theoretical pronouncements have been disqualified by their confusions of the arts (and particularly by the confusion of prosody and musical rhythm)— nor with a poet of imposing stature. Camille Mauclair, too, is anxious to show how naturally vers libre finds its place in a whole gamut of rhythmic resources upon which the poet draws in the way that his psychophysiological impulses dictate, at any given moment. And so he argues that nobody has invented free verse ‘parce quʼil y a autant de vers libres quʼil y a de poètes, et que leurs musiques ne se ressemblent pas’ (Marinetti 1909: 65). Later in his reply, he illustrates this principle of the self-selection of rhythms, by reference to his own practice:

(p.170) Il mʼest arrivé de faire des vers réguliers, et Poesia en publiait récemment. Je ne les ai pas faits exprès. Jʼai suivi mon sentiment, qui determinait une certaine musique verbale. Quand jʼai eu fini, jʼai vu des strophes alexandrines sur mon papier. Le poème disait un état rythmique très normalement cadencé: si jʼavais transcrit un état dʼagitation, de réticences suivies dʼelans, jʼaurais été amené a des rythmes polymorphes. Cʼest le poème qui se fait tout seul selon sa logique intime … (pp. 66–7)

For this reason, the consistent rhythmic symmetries proposed by the alexandrine are absurd; eloquence may wish to speak in this way, but the emotions do not. But the alexandrine, used as an occasional resource, still has its part to play:

Lʼalexandrin est un vers libre, tous les vers sont libres, lui comme les autres, ni plus ni moins. On ne peut pas plus le supprimer que le nombre douze dans la numération: mais le nombre douze nʼa pas plus de valeur spécifique que les autres, et voilà toute la querelle. (p. 67)

Like Kahn, Mauclair looks upon the presence of enjambement in regular verse as a proof of the necessity of free verse; like Mauclair, Kahn maintains that there is still a home for the alexandrine in free verse (p. 28), thus moderating his view of 1891 (see above, section (i) of this chapter). But there is nothing in Kahn's reply that we have not already encountered in his earlier writings, apart perhaps from his claim that vers libre ‘est dʼaccord avec la phonétique ou avec la pronunciation française actuelle à Paris, sauf au Théâtre Français où elle est archaʼique et où dʼailleurs le triomphe des tragedies ou de lʼauteur comique est dʼarriver a supprimer dans le dialogue lʼapparence du vers’ (pp. 26–7), a remark which I find oddly self-contradictory. Otherwise Kahn's historical account, his praise of Banville's far-sightedness, and his insistence that free verse's ‘jeu de strophes et de rythmes permet à tout poète dʼexprimer sa personnalité par le choix quʼil en fait’ (p. 27) are familiar.

It is this defence of free verse as the direct rhythmic embodiment of the poet's deepest and most authentic self which provides one élément of the polarization intimated earlier. For while in Kahn's theoretical works the notion of free verse as psychophysiological enactment can happily coexist with the notion of the ‘strophe analytique’, which registers segments or facets of thought, largely because, for Kahn, thought is an integral part of the psychophysiological ground, there are signs in the Marinetti enquiry of a (p.171) divorce between these two notions, engendering two different ‘schools’ of vers libre, a divorce akin in some respects to the divorce between Impressionism and Neo-Impressionism. On the one hand, the ‘school’ of Èmile Verhaeren, Louis Le Cardonnel, Marie Dauguet, and Albert Boissière, promoting vers libre as the verse form whose rhythms are sufficiently supple to capture, with unparalleled immediacy, the poet's inner, fluid, instinctual, emotional, organic life:

Le vers libre est un esthétique littéraire, le dernier effort de 1ʼévolution individualiste commencée par le romantisme. Il est le rendu de ce quʼil y a de plus indépendant dans lʼhomme, non pas même de la pensée — elle se dresse jusquʼà un certain point—mais de la façon de sentir. Cʼest justement pour cela quʼil a excité de telles protestations. Il est la forme même du moi intérieur émancipé. (Marie Dauguet, Marinetti 1909: 39)

For this school, the discipline of the verse derives from the responsibilities that freedom confers; the poet has no right to the arbitrary or the fantaisiste (Verhaeren, Marinetti 1909: 36); he has a right only to a rhythm which is what it is with an aptness of 100 per cent. On the other hand, steering clear of these begged questions, the ‘school’ of Francis Vielé-Griffin and Henri Ghéon, the school of the ‘strophe analytique’. One should say from the outset that Vielé-Griffin is ill-adapted to the school he finds himself in: ‘dʼintellectuelle, raisonneuse, logicienne, soucieuse, du fini et du defini, notre muse est de venue sensuelle, impressionnable, amoureuse de lʼheure, suggestive dʼinfini, suspensive’ (Marinetti 1909: 33). These are not the words of a Cartesian. And yet Vielé-Griffin claims that the anarchic early days of embattled verslibrisme have been superseded, and that free verse is now permeated by a new positivistic, not to say scientific, spirit:

Ainsi est née, du fait dʼœuvres signées des noms les plus beaux, la grande strophe analytique, moderne laisse rythmique, familière désormais à toute personne curieuse de la littérature française contemporaine. Cette laisse a ses, lois non plus individuelles mais générates, lois vitales, organiques, comme en comporte tout être viable.

Il serait absurde, immoral et barbare de parler de fantaisie quand on traite dʼune des plus importantes pages de notre littérature, en perpétuelle genése dʼelle-même, suivant des lois évolutives, séculaires et nécessairement logiques. (p. 34)

(p.172) Here there is a curious interweaving of lexicons, but Vielé-Griffin begins to sound like a Charles Henry. Can one really believe that verslibristes, of one accord, have rallied to the cause of the strophe analytique, that all see vers libre as the instrument of an evolutionary purposefulness, that logic has totally displaced the unconscious as the source of rhythmic configuration?

Henri Ghéon has no doubt about it, nor about his high calling. Declaring that vers libre is nothing more nor less than the ‘strophe analytique’, he goes on: ‘La critique, mal informée, en est encore à la conception négative des premiers jours, incompatible désormais avec notre souci cartésien de construire: nous en sommes au rationalisme, Messieurs!’ (Marinetti 1909: 68). Kahn becomes an ancestor, one of the early ‘instinctifs purs’, like Jammes. It is to the logically minded, classically intentioned poets like Vielé–Griffin, Verhaeren(!), and van Lerberghe that one must now turn for guidance. But Ghéon's new programme for vers libre is only reheated Kahn. He argues that vers libre is neither libre (freedom in art is the choice of a discipline) nor vers:

A moins dʼappeler vers, sur la ressemblance typographique, les unités rythmiques, les unités logiques (cʼest tout un) que nous avons pris lʼhabitude — moi du moins — dʼisoler chacune sur une ligne comme on fit des vers jusquʼici, et qui ne valent jamais par elles-mêmes, mais par leur groupement, leur proportion, leur relation réciproques dans la strophe organisme qui les unit. (pp. 69–70)

Like Kahn, Ghéon then goes on to demonstrate the fundamentality to all French verse of the ‘unité rythmique’ by speaking of the alexandrine as a ‘strophe’ of two hexasyllables, and of the alexandrine couplet as a ‘strophe’ of four or more rhythmic units, depending on the degree of rhythmic fragmentation within the line. Thus, like Kahn, he can assert that these ‘unités rythmiques et logiques’ are not the invention of the verslibristes; every French poet worthy of the name has composed his verse on the basis of these units; it is just that the external forms imposed by prosodic fashion have hidden them from view. Ghéon then proceeds to formulate the rules, ‘non plus empiriques, mais rationnelles’, of the new school. The first follows on directly from what he has just said: ‘Chaque unité expressive de la pensée, chaque unité logique du discours, créera une unité rythmique dans la strophe. Corollaire: la strophe sera lʼexpression totale analytique, harmonique de la (p.173) pensée’ (p. 71). The second rule concerns that dimension of verse which complements and supports rhythmicity, namely acousticity, the ‘rappel des sons’:

Comme on en soulignait, tout arbitrairement, la fin périodique de chaque vers, nous en soulignerons chaque unité rythmique avant le temps de repos qui la suit, accentuant ainsi, et rationnellement, chaque mouvement de la strophe, chaque progrès de la pensée et chaque moment du discours. Et ce ne sera plus de façon symétrique, uniforme comme autrefois, mais suivant les milles ressources du clavier infini des assonances et des rimes: vagues ici, précises là, sourdes, éclatantes, lointaines, proches, sans dépasser la limite dʼécart où lʼécho cesse dʼêtre perçu. (pp. 71–2)

This again coincides with Kahn's programme for a wider distribution of the poem's acousticity, for the transformation of the tonic into the phonic, for a diminution of rhyme's privilege, for a multiplication of acoustic resources. Ghéon closes his outline with certain concessions: in exceptional circumstances, analytical vers libre will admit the old ‘vers-organisme à césures’ and rhythmic units without acoustic kinships.

Following a procedure used in previous sections, I would like now to test Ghéon's system against his own verse practice, and have chosen for this purpose a poem from his 1898 collection La Solitude dʼété, ‘Soudainement contre les vitres’:

1. De sa hache—ah quʼil est las—

4′+4 Non-elision of ‘e’. Hiatus. Pain in transition between phrases, and deceleration (fatigue).

2. Le chevalier aux blanches armes.

4+4

3. A coups de hache

4

4. Rompre des casqu(es),—ah quʼil est las—

4+4 Apocope or sourdine (i.e. extrametrical ‘e’ barely heard, absorbed by silence 4 … +4). Sourdine more likely; attenuation of coupe lyrique of l. 1—process of enfeeblement, loss of will, action dying away in its own pointlessness.

5. Le chevalier aux blanches armes.

4+4

6. Et de la jolie fill(e) de Perth,

6+2 Apocope. The knight enumerates the ladies without wishing to dwell on, or inflate, their images.

7. Et de Béatrix et de Berthe,

5+3

8. Et des rob(e)s à bordur(es) de perles

3+5/3+3+2 Apocope. Continuation of deflationary process of l. 6. Knight positively resists feminizing ‘e's, particularly after/R/. He is unmoved by the finery.

9. Et des cheveux sur le cou—ah quʼil est las—

4+3+4

10. Et des bras autour du cou—ah quʼil est las—

3+4+4

11. Le chevalier aux blanches armes.

4+4

12. De mourir … e—ah quʼil est las—

4′+4 Muette dʼappui (what Morier calls ‘épenthèse’), also non-elision of ‘e’, coupe lyrique. This takes up the movement of l. 1. The knight is wearied by his constant engagement with death and his never being able to die. The muette dʼappui conveys the knight's inability to let go of life; there is no end to the process of dying.

13. Le chevalier aux blanches armes.

4+4

(p.174) This poem is a fantasy on rain beating against window-panes. The first three stanzas compare the beating of the rain with the collision of birds’ beaks, seen both as an aggression (‘Un assaut’) and, more predominantly, as the panic-filled disorder of flight from an enemy. In the third stanza the collision of the beaks becomes a music, glass striking glass. The fourth stanza introduces the second image-field: the panes become a face down which, spreading as they land, the raindrops roll like tears. In the final stanza the two strains of imagery are brought together as alternating modes of the rainfall, though, thanks to the association of the birds’ beaks with fragility and fear, the ‘chante’ is closer to the ‘pleure’ than the poet might wish. Throughout the poem the cottage itself is presented as an image of concealed and na engineer the same self-sufficiency for the rhythmic unit as Kahn does: to Kahn's ‘simple intervalle’ corresponds Gheon's ‘temps de repos’ (see quotation above, p. 173), but it is not clear whether he wishes to do this by adopting Kahn's method (not counting the measure-terminal ‘e’, naïve inwardness, helplessly responding to the forces which act upon it; and as an image of the poem itself, (p.175) refusing to interrupt its own activity by too investigative an attitude—
  • —et lʼon nʼose ouvrir la chaumière grise
  • de crainte de les faire taire
—it aligns itself more with the intuitional school of free verse than with the analytical.

This covert betrayal of his own principles is to be found also in Ghéon's handling of the poem's prosody. It is true that, in conformity with the ‘strophe analytique’, each stanza is a Complete sentence. It is true that the lineation frequently corresponds to a rational, logical ordering of material: each formulation of simile, for example, occupies its own line (see ll. 3, 23, 24) as do the major metaphorical statements (ll. 12, 18, 22, 26); successive lines are frequently the product of syntactical analysis inasmuch as they contain similar grammatical sequences: thus lines 6–9 all contain the sequence subject + verb + object (where the verb is transitive), while lines 11–12 each include a noun followed by an adjectival phrase. One should note that lines 6–8 also contain a poignant syntactic progression, reflected in the gradual diminution of their overall numbers of syllables: the aggressiveness of ‘Un assaut de bees’ is borne out by the fact that it is an agent directly acting on an object; in line 7 however the transitivity of the verb is directed against the agent, in the pronominal form; finally, in line 8, the verb has become intransitive as the birds (beaks) lose control of their destiny, become randomly subject to their own actions. As further evidence of Ghéon's systematic prosodic method, one might mention a factor which also relates his verse to a cadence accentual mode: leaving aside the single-measure short lines, all other lines are susceptible, without any forcing of the rhythms, to a biphrasal, bi-accentual reading, with optional variations in intraphrasal accents; thus the last two lines are either tetrametric but biphrasal, thanks to intraphrasal accents on ‘carreaux’, ‘obscure’, ‘pluie’, and ‘chante’, or the doubling of biphrasality into tetraphrasality as cottage, window-panes, rain, beaks, and tears are assembled in a final, ordering, lapidary couplet.

But if we confront the notion of syntactical parallelism with that of intraphrasal accent we can begin to see how Ghéon's analytical project falters. Lines 1 and 5 are lexically identical, and thus made (p.176) up of identical grammatical components, but they do not operate in the same way syntactically, as the comma in the first line makes clear; the first line means, ‘ A sudden phenomenon (the downpour) occurred against the window-panes’, while the fifth line means ‘ A flight of birds suddenly struck the windows’. In order to make this syntactical distinction clear, we would need to read line 1 as 4+4 and line 5 as 8; and yet line 5 could still maintain the shade of its meaning with a light intraphrasal accent on ‘soudainement’, thus

                      Theories and Enquiries

The self-sufficiency of the ‘unités rythmiques et logiques’ depends on our being able to identify rhythmic boundaries with confidence, on our being willing to resolve ambiguities. But once intraphrasal accentuation is admitted, the position of boundaries becomes conjectural and the only way that the various possibilities of combination and separation can be maintained is by an ‘inclusive’ reading which treats the line boundary as the only reliable boundary. This is a tendency we traced in our analysis of Dujardin's ‘Voici les violons’. But, as my scansional tabulation makes clear, the room for rhythmic fluctuation and the constant redefinition of boundaries within Ghéon's poem is considerable, not only because of the doubts we may have about the treatment of the e atone, but also because there is no sure way of knowing where the boundaries of measures fall, and thus which accents are intraphrasal and which are phrase-terminal. And it seems to me that Ghéon's own poetic instinct impels him to equivocate about boundary, not only about the boundary of measure, but about the boundary of stanza as well.

I say this about the stanza for two reasons. It is true that Ghéon does not want the divisions between stanzas to be hiatuses, does not want his stanzaic sequences to be consistently discontinuous:

Et comme se groupent les pensées en sʼappelant lʼune lʼautre, en sʼopposant Tune à lʼautre, ainsi; la strophe naîtra de la strophe précédente, prendra sur elle son point dʼappui ou sʼen écartera, en antithèse — et jʼentends rythmiquement. (Marinetti 1909: 71)

And it is clear that the rhymes of ‘taire’/‘clair’(11. 10–11) and ‘face’/‘passe’ (an allowable near-homophony /fas/ /pas/) are designed to provide stanzaic transitions, just as the absence of such a rhyme (p.177) between lines 18 and 19 and the interjection of the conjunction ‘Or’ is designed to push stanzas three and four apart, as the poet installs his second thread of imagery. But in fact, on close inspection, we discover that Ghéon uses wide-ranging and wide-interval acoustic kinships in the way that Alain-Fournier does, to create a seamless fluidity between stanzas, to call their typographical shaping into question, to build up a network of more distant echoes and near-echoes, so that stanzas reach deeply into each other. Though Ghéon speaks of an acoustic framework in which connected sounds will not ‘dépasser la limite dʼécart où lʼécho cesse dʼêtre perçu’ (see above, p. 173), the evidence of this poem suggests otherwise, suggests that many acoustic connections here work at a level below conscious perception and in such a way that stanzaic outline becomes submerged. If the /εR/ of ‘verre’(1. 6) is taken up in the half-rhyme /vjεn/ (1. 8) and the full-rhyme ‘taire’(1. 10), it equally throws us back to ‘averse’(/avεRs/) (1. 2); if ‘heurt’(1. 20) seems initially to be an orphan, it ultimately finds partners in the last two lines of the following stanza (25–6); if ‘sʼétalent’(1. 21), in the structure of its own stanza, half-rhymes with ‘face’(1. 23), it has much stronger acoustic claims to a rhyme with ‘cristal’(1. 14 and 1. 12) and indeed with ‘parle’(1. 18); and further, by added virtue of their common disyllabicity, ‘sʼétalent’ and ‘qui frappent’(1. 13) propose themselves as acoustic mates; though ‘douces’(1. 22) clearly half-rhymes in the first instance with ‘gouttes’(1. 19), it actually half-rhymes more convincingly with ‘double’(1. 16); and so on.

Secondly, there are rhythmic reasons for claiming that Ghéon's poem is a single intricate web in which the overt divisions are only provisional, only one possible version of a rhythmic and acoustic complex, whose function is not to validate its own priority but to activate other versions. It is not just that the sequence of disyllabic lines (4, 13, 20, 21) creates a cross-pattern, but that the whole sequence of single-measure lines (16–21) creates its own stanza, embedded in and overlapping two other stanzas, a stanza with its own music (‘double’/‘goutte’, ‘musique’/‘vitre’, ‘parle’/'sʼétalent’) and with its own rhythmic principle of progressive syllabic reduction, in which the rain loses its voice, and its concerted collectiveness (‘la pluie’), to become fragmented plurality (‘les gouttes’) and arbitrary noise (‘au heurt’); in this stanza, too, there is a movement towards liberating paradox: as the diminishing (p.178) measures reach a minimum, so they find a source of expansion (‘sʼétalent’). These kinds of rhythmic redistribution also occur locally, within the stanza, largely because of potential ambiguities of syntactic collocation. Thus, as the last élément of lines 11–13, ‘qui frappent’ is a discord, both by its brevity and because its disyllable is out of key with the trisyllabic final measures of lines 11 and 12; it becomes a momentary dead end, unmusical, with falling intonation. But if we read lines 11–14 like this:

  • Ce sont des bees sonnant clair
  • des bees de cristal harmonique |…
  • qui frappent, →
  • cristal contre cristal |
then the disyllable of ‘qui frappent’ is the impulse for the disyllable of the first ‘cristal’ which then doubles in ‘contre cristal’, and correspondingly ‘qui frappent’ will invite a rising intonation. Similarly, if we take ‘double musique’(1. 16) as the last élément of lines 14–16, it is a tetrasyllable reinforcing and consolidating the tetrasyllables of lines 14–15, bringing a movement to a close in its own dying fall; as the first element of the new ‘stanza’ we have just identified (lines 16–21), it is an impulsive declaration which then undergoes a gradual erosion.

The conclusion I would draw from this analysis would be that the ideal of a rationally controlled and controllable ‘strophe analytique’, a stable and systematic versification of free verse, is a chimera; that, as in Kahn's theoretical writings, the free verse of the individual measure must derive from a free exchange between the analytic and the intuitional, between the relative structural authority of the measure and its susceptibility to a whole range of distributional adaptations in the reading process; that in the free verse of the individual measure, the line can never be just the aggregate of a sequence of measures, but does interfere with and complicate the rhythmic activity of the measure; that Ghéon's own poetry bears this out, despite his efforts to give ultimate authority to measures and to an analytical method of combining them.

Notes:

(1) In a letter of 2 Sept. 1920, Mockel confides to Dujardin (1922: 62): ‘Mais en 1887 je lus les Palais Nomades, et je considère comme un devoir de stricte honnêteté de proclamer que lʼexemple de Gustave Kahn fut pour moi décisif. Dans les Palais Nomades—que me révéla un article de Wyzewa dans votre Revue Indépendante et que jʼachetai presque aussitôt—je voyais se réaliser par miracle ce que depuis deux ans, en gosse mal initié, perdu dans une province lointaine, je cherchais avec des tâtonnements pleins de gaucherie.’

(2) ‘Alternants’ here presumably means ‘alternating measures of two (binaire) and three (ternaire) syllables’.

(3) For treatments of the villanelle and rondeau, see Scott 1980a: 157–71; of the rondel, see Scott 1980b; of the triolet, see Scott 1980c.

(4) Joseph Boulmier, whose Villanelles of 1878 popularized the 19-line variety of villanelle derived from the work of Jean Passerat (1534–1602), a variety adopted by the English, asserts in his ‘Notice historique et critique sur la villanelle’: ‘Le vers de sept syllabes, pimpant et dégagé dʼallure, est le vers attitré de la villanelle’ (1878: 13).