UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Music and poetry in Mallarmé and Debussy Wilson, Geoffrey Allan 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2007-319611.pdf [ 16.8MB ]
JSON: 831-1.0100820.json
JSON-LD: 831-1.0100820-ld.json
RDF/XML (Pretty): 831-1.0100820-rdf.xml
RDF/JSON: 831-1.0100820-rdf.json
Turtle: 831-1.0100820-turtle.txt
N-Triples: 831-1.0100820-rdf-ntriples.txt
Original Record: 831-1.0100820-source.json
Full Text

Full Text

MUSIC AND POETRY IN M A L L A R M E AND DEBUSSY by GEOFFREY A L L A N WILSON B. Mus., The University of Calgary, 1999  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Music)  THE UNIVERSITY OF BRITISH COLUMBIA September 2007 © Geoffrey Allan Wilson, 2007  11  ABSTRACT My dissertation re-evaluates music and poetry in the works of Claude Debussy and Stephane Mallarme. Often in such collaborations, critics assume that the music mimics various aspects of the texts it engages. Instead, I argue for a more nuanced paradigm that values both concurrences and antagonisms between the two media, in light of the specific systems of thought characterizing, respectively, the poet and the musician. Chapter One re-evaluates the role of music in Mallarme's oeuvre. Mallarme imagined an original language in which individual phonemes created the meaning of words. As languages evolved and multiplied, the sound-sense relationship in words became increasingly arbitrary. Traces of this original language are visible in contemporary idioms when a group of words share both a phonemic and a semantic link. For him, poetry exists to reconstruct soundsense relationships in modern language. These relationships, and the patterns of thought they enact, are music for Mallarme, a music which the sound of instruments and singers merely implies. Drawing evidence from Mallarme's letters and critical writings, I establish the "musical" nature of his language and show its use in analyses of selected poems. In the remaining chapters I examine each of Debussy's compositions that engage a Mallarme text: the songs Apparition, Soupir, Placet futile, Eventail, and the Prelude a I'apresmidi d'un faune. I chronicle  Debussy's early exposure to Mallarme's poetry through Verlaine's  essay "Les poetes maudits" and the song Apparition that resulted. Here, Debussy responds to the semantic content of Mallarme's poem, constructing a musical "apparition" to parallel the poetic one. I offer a new reading of Prelude that relates it to Mallarme's dramatic theories, and not as a mimetic illustration of the poem's text. I argue that Debussy's later song settings allow Mallarme's poetic "music" to be perceived alongside his own. In Soupir, this is manifested through a series of mirror images in both music and text. In Placet futile, I show how the music  Ill  alters the semantic message of the poem. In Eventail, I compare the interaction of wholetone, octatonic and diatonic pitch collections to the interaction between the phonetic and semantic layers of the poem.  IV  T A B L E OF CONTENTS Abstract  ii  Table of Contents  iv  List of Figures  vii  List of Musical Examples  viii  Preface  x  Acknowledgments  xii  Dedication  xiii  INTRODUCTION  1  Debussy's Relationship with Mallarme Debussy's Opinion of Mallarme Literature Review: Debussy Literature Review: Mallarme Sound in Mallarme Music in Mallarme Problems and Solutions in Mallarme and Debussy Structure of the Dissertation Methodology CHAPTER ONE  The Music of Language: Sound and Sense in Mallarme  Introduction Music in Mallarme's Early Writings The Song of the Sailors: Mallarme's "Brise marine" Mallarme's Silent Muse: "Sainte Cecile" Music and the Crisis Music and Epistemology Mallarme's Demons An Aborted Dissertation Music after the Crisis Mallarme's Great Work Music in the Mature Poetry The Case of Wagner Music in the Late Critical Works La Musique et les Lettres The Music of Reading Poetry as Music Conclusion  4 7 13 22 25 30 36 39 42 49 49 52 60 65 72 74 77 82 87 92 ...97 102 Ill 114 118 121 122  V  CHAPTER TWO  Debussy's First Contact with Mallarme: Apparition  125  Mallarme's "Apparition"  131  Debussy's Apparition  138  CHAPTER THREE  The Faun's Arabesque  149  A Failed Theatrical Project Mallarme's Ideal Theater Mallarme's Arabesque The Arabesque in "L'apres-midi d'un faune" Mallarme's Ideal Overture Debussy's Prelude Debussy and the Arabesque The Composition of the Prelude Debussy's Flute The Faun's Arabesque Conclusion  151 154 157 161 171 171 172 177 179 181 197  CHAPTER FOUR  Le Chant sous le Texte: Debussy's Troispoemes de Stephane Mallarme  Part One: Un soupir qui mire: Reflection in Soupir  198  201  Mallarme's "Soupir"  201  Debussy's Soupir  211  Part Two: Invocations of a Pastoral Past: Placet futile Mallarme's "Placet futile" Debussy's Placet futile  Part Three: The Song of the Fan: Eventail  222 222 230  246  Mallarme's "Autre eventail" Vers de circonstances Mallarme's Eventails "Autre eventail" Revisited Sound in "Autre eventail"  247 252 256 257 259  Debussy's Eventail Tonality in Eventail  262 267  The Wholetone Collection in Eventail Conclusion: Eventail and Apparition  275 281  vi CONCLUSION  283  Bibliography  290  Appendix A  Chronology of Mallarme's Writings Cited  298  Appendix B  IP A Phonetic Symbols  300  Appendix C  Mallarme's "Le demon de l'analogie"  301  Appendix D  Debussy's Apparition  303  Appendix E  Mallarme's "L'apres-midi d'un faune"  309  Appendix F  Debussy' s Soupir  312  Appendix G  Debussy's Placet futile  316  Appendix H  Debussy's Eventail  321  vii LIST OF FIGURES Figure 1.1.  Excerpt from Mallarme's Letter Tables (W) in Les mots anglais  90  Figure 1.2.  Diagram from Mallarme's "Notes en vue du Livre" OC I, 567  93  Figure 1.3.  Diagram from Mallarme's "Notes en vue du Livre," OC I, 572  95  Figure 1.4.  Diagram from Mallarme's "Notes en vue du Livre," OC I, 578  95  Figure 1.5.  Schematic of a Hegelian Triad  96  Figure 1.6.  Schematic of Conn's tetrapolar representation of Mallarme's epistemology  96  Figure 2.1.  "e" phonemes in Mallarme's "Apparition"  135  Figure 2.2.  Formal Design of Debussy's Apparition  146  Figure 3.1.  Diagram from "Notes en vue du Livre," OC I, 572  160  Figure 3.2.  The arabesque of the flute in "L'apres-midi d'un faune"  167  Figure 3.3.  The arabesque of the flute in Prelude a I 'apres-midi d'un faune  Figure 4.1.1.  Phonemic palindrome in "Soupir," line 7  210  Figure 4.2.1.  Debussy's scansion of the text in Placet futile  230  Figure 4.2.2.  Adjusted scansion in Placet futile  232  Figure 4.3.1.  Mallarme's "Autre eventail (de Mademoiselle Mallarme)" and its significant revisions  248  Figure 4.3.2.  Eventail,  parsing of stanza four  186  271  viii LIST OF MUSICAL EXAMPLES Example 2.1.  Apparition, mm.  1 -14 reduction  142  Example 2.2.  Apparition, m.1-14.  Example 3.1.  Prelude a I 'apres-midi d'un faune. Schenkerian graph  Derivation from Gt major  143 of a  "typical" reading of mm. 1-4 Example 3.2.  182  Prelude a Vapres-midi d'un faune, alternative  reading of mm. 1-4,  exposing the two contrasted tritones Example 3.3. Example 3.4.  183  Two tritone axes, one possible parsing for the four pitches emphasized in mm. 1-4  184  Various versions and harmonizations of the first theme of the Prelude a I 'apres-midi d'un faune  Example 3.5.  Prelude a I'apres-midi d'un faune, mm.  37-39, oboe  Example 3.6.  Prelude a I 'apres-midi d'un faune, mm.  Example 3.7.  Graph of tonal closure in the Prelude a I'apres-midi d'un faune  Example 4.1.1.  Soupir,  55-60, winds  mm. 20-22, Octatonic setting of Mallarme's phonemic palindrome  187  190 191 193 215  Example 4.1.2.  Soupir,  reduction of piano chords, mm. 9-15  217  Example 4.1.3.  Soupir,  piano, mm. 20-21  218  Example 4.1.4.  Soupir,  mm. 27-31, third-related dominants and cadence in At  220  Example 4.2.1.  Placet futile, mm. 9-10,  voice  232  Example 4.2.2.  Placet futile, mm. 22-3, voice  233  Example 4.2.3.  Placet futile, mm.  30-3, voice  234  Example 4.2.4.  Placet futile,  mm. 1-3, piano  236  Example 4.2.5.  Three possible versions of Debussy's "unsounded" original gesture  236  Example 4.2.6.  Placet futile, mm. 4-7, piano  238  Example 4.2.7.  Placet futile, mm.  239  Example 4.2.8.  Placet futile, m.  Example 4.2.9.  Placet futile, mm.  3-10, harmonic reduction  14-5, piano 20-1, piano  Example 4.2.10. Placet futile, mm. 22-23, piano  240 242 244  ix Example 4.3.1. Example 4.3.2. Example 4.3.3.  m. 1. Symmetrical disposition of the pentatonic collection  265  mm. 4-8 reduction. Symmetrical arrangement of 014 sets  266  mm. 19-23, piano, reduction. Symmetrical arrangements of 026 sets  267  Eventail,  Eventail,  Eventail,  Example 4.3.4.  Eventail,  mm. 12-24  270  Example 4.3.5.  Eventail,  mm. 50-65, with harmonic reduction  273  Example 4.3.6.  The tonal scheme of Eventail  Example 4.3.7.  Eventail,  mm. 25-36 with analysis of wholetone subsets  277  Example 4.3.8.  Eventail,  mm. 28-33 reduction  278  275  PREFACE  Throughout the dissertation, quotations from Mallarme's letters and critical essays are given in translations in the body of the text, while the original French is given in footnotes. The translations of the critical prose are my own. Given the difficulty of the prose, my translations are not meant to be artistic. They sometimes violate the syntax of the original in the service of clarifying the point I wish to make. Wherever possible, I have compared my translations to those published by others, in particular the translation of letters by Rosemary Lloyd and the translations/commentaries of the critical prose by Robert Greer Cohn. When they have been particularly influenced by (or where my interpretations depart significantly from) these sources, an acknowledgement is given in the notes. What is lost in translation by adopting this strategy is, I hope, counterbalanced by greater clarity. For the poetry, however, I have placed my translations beside the original text in the body of the paper. These translations are literal rather than poetic. They strive to match the grammar and syntax of the original without trying to approximate its prosody, since the phonetic aspects of Mallarme's poetry that I discuss are truly untranslatable. I have followed the same procedures with quotations from Debussy, even though the language is not often difficult. For the titles of the poems and essays, I follow the somewhat idiosyncratic capitalization scheme of Marchal's QLuvres completes. In the chapters that trace the interaction of Mallarme's poetry and Debussy's music, a problem exists in the fact that two separate artworks are known by a single title. Therefore, throughout the dissertation, I have chosen to give all of the titles of Mallarme's poems in quotation marks, while rendering all of Debussy's composition titles in italics. The only exception in this regard is Un coup de des, which stands so far beyond any of Mallarme's previously published poems in complexity and ambition that it deserves this distinction. While  xi this scheme of quotation marks and italics is not entirely conventional, I hope that it will minimize confusion in distinguishing the poet's work from the composer's. The scores for the songs in the Appendices are all prepared from the original published editions (1913 for Soupir, Placet futile and Eventail; 1926 for Apparition). They are included here because they may not be readily available in smaller libraries or previously published compilations of Debussy's songs. I have made some editorial changes, mostly of a practical nature, clarifying the distribution of notes between the hands and removing or placing in parentheses numerous redundant accidentals. When these songs are finally published in the forthcoming Debussy critical edition, the reader is advised to consult those instead.  xii  ACKNOWLEDGEMENTS  This dissertation would not have been possible without the dedication, encouragement and wisdom of my advisor Dr. Vera Micznik. Her thoughtful and provocative questions were a constant source of inspiration during the research and writing process. It is my genuine hope that something of her passion for knowledge and academic enquiry has rubbed off on me during our work on this project. My heartfelt thanks go out also to the other committee members: Dr. William Benjamin, Dr. Richard Kurth and Dr. David Metzer. Their perspicacity, both in regards to this project and in their own work, sets a high standard that I have tried to emulate to the best of my ability. Their intellectual fingerprints are discernable on virtually every page of this work. A sincere word of gratitude must also go to Dr. Sima Godfrey and Dr. John Roeder, whose thoughtful reading of the dissertation provided fresh insights. Dr. James McCalla's probing questions helped me to clarify much that was once fuzzy, and have opened up to me new horizons for continued work on Debussy's relationship with nineteenth-century French literature. Finally, I must thank my family, especially my wife Emily, who was patient, loving and supportive in difficult times and who celebrated with me in times of joy. To my children Megan and Hayden, this project has often been a greedy third sibling that took more than its fair share of my time. Their innocence and laughter helped me to put some of the more frustrating days of research into perspective. As for the rest of my family, they have supported me emotionally and financially for many years. Without their enduring patience, this project would have been unthinkable.  xiii  For Emily  1 INTRODUCTION  Music and poetry are two songs that search in vain to accord with one another, and even in the very rare cases where they are in accord, have the effect of a bad pun. 1  Debussy, Letter to Pierre Louys, 10 April, 1895  This is not the kind of epigraph one usuallyfindsintroducing a discussion on music and poetry. It is one that is especially surprising from the pen of Claude Debussy (1862-1918), a composer whose debt to the poetry of the Symbolist movement has been so firmly ensconced in conventional wisdom that it hardly receives any critical reflection these days. A more typical epigraph would have been Walter Pater's dictum that "all art aspires to the condition of music" or Paul Verlaine's "de la musique avant toute chose," or even an appeal to the Wagnerian Gesamtkunstwerk;  something that really plays up the happy concordance of the arts in the  French fin-de-siecle. Yet I begin here because I think that Debussy has hit upon something important about the relationship between music and poetry, something that has gone largely overlooked both in Debussy's case in particular, and in music-and-text relationships in general. This is the tension between music and language, a tension that gets progressively more intense as the two arts are brought into closer contact. On one hand, there is a genuine desire in song 2  to find an expressive resonance between the music and the text. Yet, at the same time, there is the fear that, if pushed too far, this resonance becomes mere imitation; that the music loses itself in the process. It is this mutual attraction and its concomitant anxiety, more than a happy  Debussy, Correspondence, ed. Francois Lesure and Denis Herlin (Paris: Editions Gallimard, 2005), 249: "D'ailleurs, la musique et les vers, c'est deux chansons qui cherchent vainement a s'accorder, meme dans les cas tres rare ou cela s'accorde, 9a fait I'effet d'un mauvais calembour." Subsequent references to Debussy's corrspondence shall use the abbreviation Corr. Lawrence Kramer is one of the few scholars to give significant attention to the complexities of the music and text relationship in song. See Music and Poetry: The Nineteenth Century and After (Berkeley and Los Angeles: University of California Press, 1984). 1  2  2 concordance of the two arts, that characterizes the relationship between music and poetry in the Frenchfin-de-sieclein general and Debussy's oeuvre in particular. By 1895, when Debussy wrote the letter quoted above, he had already composed well over 60 songs (more than two-thirds of his total output) and substantial portions of two operas, the abandoned Rodrigue et Chimene and Pelleas et Melisande, as well as the prize cantatas written for the Prix de Rome competitions. He had set a considerable number of poems by some of the most important French poets of the recent past, including Baudelaire, Verlaine, and Theodore de Banville, a prominent Parnassian poet. The vast majority of his instrumental music, including the most important pieces (save the Prelude a Vapres-midi d'un faune) was yet to be written. The remaining years of Debussy's life would see an increasing inability to complete vocal works, an affliction from which his instrumental works seem relatively immune. Among the projects contemplated but abandoned, notable examples include dramatic projects on texts by various prominent writers: Cendrelune (Pierre Louys, 1895-8), Les uns et les autres  (Verlaine, 1896), Comme il vousplaira (Shakespeare, 1902-4), L'histoire de Tristan  (Gabriel Mourey, 1907-9), Orphee-roi (Victor Segalen, 1907-9), Siddartha (Segalen, 190710) and Crimen Amoris (Verlaine, 1914). The two major opera projects on Poe texts, Le diable dans le beffroi  and La chute de la maison Usher both remained unfinished at Debussy's death.  His inability to complete these works in particular, which he had contemplated for at least 25 years, speaks to an increasing difficulty he had in finding a suitable music to set alongside these texts. Why would a composer who had been writing music in response to poetic texts for the bulk of his early career hold such a dim view of the possibility that music and text could or even should be in mutual accord? And why would his output, which until that point had consisted almost entirely of vocal music, shift to include an increasing number of instrumental  3  works that refer to text only obliquely, if at all? The answer, I think, involves Debussy's growing awareness of the complexities of music and text relationships, and the increasing pressure he put on himself to make his musical responses reflect this situation. Tensions between music and poetry characterize Debussy's later writings. His choice of blank verse or a kind of rhythmical prose in his own texts for Proses lyriques (1895) was motivated by a belief that prose was more suitable than verse for settings as song. The 3  preference was not aesthetic, but practical: he simply found prose more amenable to, or at least less resistant to, the kind of music he wanted to write than poetry. Debussy held this view, more or less intact, for the rest of his life. Fifteen years later, responding to an enquete by Ferdinand Divoire, he lays out his ideas on the subject: [1 think] that musicians who understand nothing about poetry should not set it to music. They can only spoil it.... Truly beautiful poems, we must not exaggerate, there are not many of these. Who writes them these days? But when one does find them, it is better not to meddle with them. Henri de Regnier, who writes full, classic poetry, may not be set to music. And can you envison music setting the verse of Racine or Corneille? ... Real poetry has its own rhythm, which is rather difficult for us [to set musically]. ... It is very difficult to follow well, to "veneer" the rhythms while still preserving one's inspiration. If one fabricates, if one is content with a work of juxtaposition [of the two arts], evidently this is not difficult, but then it is not worth the trouble. 4  Two things emerge from this passage. The first is that a composer must understand the poetry that he intends to set. In support of this point, Debussy singles out Schumann's settings of Heine that, in his opinion, fail to capture the inherent irony of the texts, a quality related mostly to the semantic dimension of the poems. The second and possibly more important point to emerge from the passage deals with the sonorous elements of poetry. For Debussy, it is the rhythm of the poetry—the materiality of its words, the way that they sound—rather than Debussy, Corn, 249. Letter to Louys, 10 April 1895. Debussy, Monsieur Croche et autres ecrits, ed. Francois Lesure (Paris: Gallimard, 1987), 206-07: "...les musiciens qui ne comprennent rien aux vers ne devraient pas en mettre en musique. lis ne peuvent que les gacher... Les vrais beaux vers, il ne faut pas exagerer, il n'y en a pas tant que 9a. Qui en fait aujourd'hui? Mais quand il s'en trouve, il vaut mieux ne pas y toucher. Henri de Regnier, qui fait des vers pleins, classiques, ne peut pas etre mis en musique. Et voyez-vous de la musique sur des vers de Racine ou de Corneille ?... Les vrais vers ont un rythme propre qui est plutot genant pour nous.... c'est tres difficile de suivre bien, de « plaquer » les rythmes tout en gardant une inspiration. Si on fait de la fabrication, si on se contente d'un travail de juxtaposition, evidemment ce n'est pas difficile, mais alors ce n'est pas la peine." 3  4  4 simply their meaning that is the most difficult element of verse to set as music. By rhythm, of course, Debussy means much more than just the poem's meter, but also the rhythm created by rhyme that operates inside the verses themselves. The difficulty comes in creating a counterpoint of music and poetry that both allows the original sonorous material of the poem to speak and preserves the inspiration of the composer. This suggests that in song, for Debussy at least, the music should not follow the poem too closely for fear of becoming its mimetic echo, a "bad pun" of the poetic text. Likewise, the music cannot simply be unrelated to the poem, a "juxtaposition" of the two arts. Instead, Debussy seems to argue for an interactive model of music and text relationships in which the music and the poem read across one another, configure and reconfigure their shared space. This model maintains a certain level of independence for both art forms and, in the process, creates a "polyphonic" set of meanings for the work: the poet's, the composer's, and that of the work itself, whose total significance resides at the intersection of these two strands of meaning. The relationship that Debussy advocates between music and poetry in this and other late essays shows a different side of his aesthetic than the image of the composer as erstwhile symbolist poet that some studies have tried to put forward. It seems no coincidence that the increase in instrumental music and Debussy's struggle to complete various texted works later in his career both follow a period of relatively close personal contact with many members of the Parisian literary milieu, and with the poetry and aesthetic of Stephane Mallarme (1842— 1898) in particular.  D E B U S S Y ' S R E L A T I O N S H I P WITH M A L L A R M E  The reasons why Debussy chose to direct his creative efforts more consciously towards instrumental music around 1895 were never definitively stated by the composer. Certainly, the  5  difficulties he experienced during the composition of Pelleas et Melisande played a role in this, as did the surprising popularity of the  Prelude a L'apres-midi d'un faune,  the premiere of  which (in December 1894) opened up both new compositional possibilities and new sources of income for Debussy. His close personal association with figures of Parisian literary scene also cooled noticeably during this time, with the exception of his good friend Pierre Louys. He stopped attending the  mardi  gatherings of Mallarme, where he had been an occasional guest  since 1890. There, he would have heard Mallarme speak on poetry, art, music and aesthetics. 5  Mallarme was one of the first to suggest that Wagner's genius was not an example that French artists ought to follow. Debussy, whose enthusiasm for Wagner in the late 1880s is wellknown, seems to have heeded this call in the early 1890s. The culmination of this may well be the Prelude, which can bee seen as a re-writing of Tristan that recasts both its emotional power (from emphatic desire to limpid sensuality) and its harmonic language. It is no coincidence that Debussy's increasing difficulty with texts occurs after his closest personal and artistic contact with Mallarme. Mallarme was, with the possible exception of Verlaine, the most important French poet of his generation, rising to prominence first as the head of the Decadents in the wake of Joris Karl Huysmans' novel A  rebours,  figure in the Symbolist movement. Despite his public position as 6  and as a major  chef d'ecole  of the  Symbolists, Mallarme's poetry bears only a surface resemblance to that of his contemporaries. His passion for the supremacy of verse and the stubborn complexity of his poetry and critical writings drew young writers and artists to his humble home on the Rue de Rome, where Mallarme held forth on a wide range of topics. Central among these was the relationship  Rosemary Lloyd has summarized the main points of Debussy's personal contact with Mallarme. See Lloyd, "Debussy, Mallarm6, and 'Les Mardis'," in Debussy and His World, ed. Jane F. Fulcher (Princeton, NJ: Princeton University Press, 2001), 255-69. Huysmans' valorization of Mallarme's poetry in the novel shot the then-obscure poet to national fame in 1884. This is detailed by Guy Michaud, Mallarme, trans. Marie Collins and Berha Humez (New York: New York University Press, 1965), 102-3. 5  6  6  between music and poetry, and the way that poets must turn away from imitating musical effects and discovering the inherent music of language. He advocated a return to a pure language, one set apart from the common speech and writing of the world. In response to the Gesamtkunsrwerk,  Mallarme offered Le Livre, his unfinished Great Work, which would  complete the synthesis of the arts begun by Wagner. Le Livre was to be a work that would supplant music, theater and dance, and distill their essences into verse. If Debussy was initially drawn into Mallarme's sphere in search of an alternative to the conservative aesthetic of the Conservatoire and the Academie, he would have found it there in spades. In both Mallarme's poetry and critical writings, Debussy would have heard a unique perspective, one whose deep respect for the traditions of French verse was tempered with a quiet contempt for anything derivative or commercial. Debussy had access to these views in the numerous essays that Mallarme published in Parisian journals from 1885 through 1898, the substance of which was often incorporated into the mardi speeches. Debussy also had a personal friendship with the poet beyond the mardi gatherings, and one can only imagine the substance of their conversations. Debussy, Mallarme and Pierre Louys were frequently found together at various artistic events: concerts of Renaissance polyphony at St. Gervais, performances at the Concerts Lamoureux, even the premiere of Maeterlinck's Pelleas et Melisande.  Debussy and Mallarme even collaborated on a proposed theatrical production of  "L'apres-midi d'un faune" in late 1890 or early 1891 that was the genesis of Debussy's Prelude J  It would seem, in light of this, that Debussy had ample opportunity to absorb Mallarme's aesthetic, to become another acolyte in Mallarme's circle (like Paul Valery) whose Mallarme was planning a theatrical performance of "L'apres-midi d'un faune" at the Theatre d'Art in Paris. Impressed by a recent performance of Debussy's Cinq poemes de Charles Baudelaire, Mallarme arranged to meet Debussy and engage him to write the necessary incidental music. The collaboration was short-lived, but the Prelude was certainly conceived during this meeting in some form. See Francois Lesure, Claude Debussy avant Pelleas, ou les annees Symbolistes (Paris: Klincksieck, 1992), 98-99. 7  greatest glory would be found in the shadow o f the master. Yet this did not happen. Despite their close contact, Debussy made no settings of Mallarme texts at this time (the unpublished setting o f Apparition in 1884 significantly predates Debussy's personal contact with the poet). This is particularly surprising in light of the regularity with which Debussy set prominent poets like Verlaine, Banville and Baudelaire. Even the Prelude a I'apres-midi d'un faune does not engage a Mallarme text directly, but stands in opposition to it. Although several scholars have attempted to line up particular passages o f the poem with particular measures o f the music, there is no general agreement on how, or even if, the Prelude responds to the text of Mallarme's poem (rather than its meaning or idea). If Debussy considered writing songs on any other Mallarme text, no evidence of these compositions has survived. It would seem then that Debussy's increased interest in instrumental music, and his growing difficultly in producing texted work, results (in part at least) from a desire to interrogate the nature o f music itself, to purify it from derivative elements of various kinds. This also throws into sharp relief Debussy's decision, some twenty years later, to return to Mallarme's poetry. Perhaps his musical language had progressed to the point where he felt ready to take on the challenge posed by Mallarme, that he had achieved a purity o f expression that measured up to Mallarme's similar effort in verse. These songs are the only Mallarme settings published during the composer's lifetime. They also mark a final break with text: these are Debussy's final songs, his last completed "essays" on music-text relationships, before turning his attention to the late works: the piano Etudes, the chamber sonatas, and various unfinished dramatic projects.  DEBUSSYS ' OPINION OF MALLARME It is difficult to gauge Debussy's opinion o f Mallarme from his writings. Mallarme is only rarely mentioned by Debussy, and the few instances where the poet's name does emerge  8  are ambiguous. The first significant mention of Mallarme in Debussy's correspondence comes from a letter to Ernest Chausson. Here, Debussy offers a criticism of Chausson's Le roi Arlhus: One thing that I would like to see you lose is your preoccupation with the "underpinnings"; let me explain: I think that we have been led in this, as always, by the same R. Wagner, and that too often we dream of the frame  before having  the painting,  and sometimes the richness  of this frame  makes us  overlook the mediocrity of the ideal I'm not talking about cases where magnificent inner parts dress up ideas, like cheap dolls! one would be better off, it seems to me, to take the opposite viewpoint, that is to say to find the perfect outline for an idea and to place alongside it only those ornaments that are necessary, for truly "certain" of these are like priests vesting incomparable gems on wooden idols! look at the poverty  of symbol  hidden  in several  of Mallarme's  last sonnets,  where  despite  this  the  workmanship is carried to its furthest limits] and look at Bach, where Everything conspires prodigiously to highlight the idea, where the delicacy of the inner parts never absorbs the principal line. 8  Debussy suggests that overly complicated compositional technique ("underpinnings") obscures the fact that the main melodic ideas of the work are insufficient to express the subject matter. The images he invokes to support his point—the ornate frame that distracts from the poor quality of the painting, the cheap doll gussied up in fancy clothing, and the wooden idol decorated with jewels—all point to a lack in the quality of the main thing that cannot be recompensed by clever marketing. When Debussy invokes Mallarme and Bach in support of his point, the situation becomes much more complicated. Bach is clearly praised here as an example of a composer whose "frame" never conceals a second-rate idea. Rather, it is the inherent quality of the musical idea (the "picture") that the "underpinnings" work to display clearly. The treatment of Mallarme is less certain. The way that the sentence is constructed seems to set up a parallel between Bach and Mallarme as two examples of artists who always have, the "picture"  Debussy, Corr., 167: "Un chose que je voudrais vous voir perdre, c'est la preoccupation de « dessous » ; je m'explique : je crois que nous avons ete mis dedans, toujours par la meme R. Wagner, et que trop souvent nous 8  songeons au cadre avant d'avoir le tableau, et quelquefois  la richesse de celui-ci nous fait passer sur  I'indigence  de I'idee ! Je ne parle pas du cas, ou des dessous magnifiques habillent des idees, comparables a des poupees de treize sous ! on gagneraient, il me semble, a prendre le parti contraire, c'est-a-dire, a trouver le dessin parfait d'une idee, et de n'y mettre alors que juste ce qu'il faudrait d'ornements, car vraiment « certains » sont pareils a des pretres revetant de gemmes incomparables des idoles en bois de sapin ! regardez la pauvrete de symbole cachee dans plusieurs  des derniers sonnets de Mallarme,  oilpourtant  le metier d'ouvrier  d'art est porte it ses  dernieres limites! et regardez Bach, ou Tout concourt prodigieusement a mettre I'idee en valeur, oil la legerete des dessous n'absorbe jamais le principal." Emphasis mine.  9  foremost in their minds. But what Debussy actually says about Mallarme works against this: if there is indeed a "poverty of symbol" hidden in several of Mallarme's late sonnets, then it must be hidden by the "workmanship" of the verses, which are taken to their "furthest limits" despite the lack of symbol, rather than in service of some present idea.  9  Critical reaction to this passage has been mixed. Some, like Rosemary Lloyd, contend that the letter shows Debussy's detailed understanding of and sympathy for Mallarme's aesthetic. Others, like Jean-Michel Nectoux, see the letter as a rejection of Mallarme's late style in favor of his earlier poetry. I think that the latter position is more likely true. I cannot 10  get past the phrase "poverty of symbol" which, to my mind, can not be a term of praise." Debussy seems to equate this with the "mediocrity of the idea" in certain of Wagner's compositions, where a prodigious technique (here poetic, there compositional) obscures the fact that the idea itself is not particularly compelling. Mallarme's earlier works, which include all the poems for which Debussy crafted a musical response, are not at issue here, and nowhere does Debussy speak ill of the early works. At issue here, I think, are the divergent notions of the word "idea" in Debussy's and Mallarme's thinking. For Debussy, the idea of a composition is reducible to an actual musical line, which ought to be expressed according to its own inherent qualities, and not subjected to an external rigorous technique, no matter how beautiful the result. Elsewhere, Debussy writes, "A musical idea contains its own harmony (or at least that's my opinion); without this,  "Derniers sonnets" is the subtitle of the ninth book of Mallarme's Poesies (1887). It contains the sonnets "Le vierge, le vivace et Ie bel aujourd'hui," "Quand l'ombre menca," "Victorieusement fui le suicide beau," the Sonnet enyx, "Mes bousquins refermes sur le nom de Paphos," "Quelle soie aux baumes de temps," "Tout orgueil fume-t-il du soir," "Surgi de la croupe et du bond," "Une dentelle s'abolit," "M'introduire dans ton histoire" and the Hommages to Poe and Wagner. See Lloyd, "Debussy, Mallarme and 'Les Mardis'," 263; and Nectoux, "Debussy et Mallarme" Cahiers Debussy 12/13 (1988/89): 62. " This phrase calls to mind Debussy's comments about chords losing their "symbolic value" through being used in music with a commercial aim, written six weeks before the letter in question. See Debussy, Corr., 155. y  10  10  harmony is no longer anything but parasitic and inelegant." This same attitude is expressed in 12  a letter from Debussy to Raoul Bardac dated 24/25 February 1906.  13  After criticizing Bardac  for not exercising enough care in choosing among the various pathways that his musical ideas suggest, he continues:  You know how little I like that parasitic development, which has for too long served the glory of Masters. We need to replace this development with a more rigorous selection of ideas, a line more conscious of the value of those ideas on the orchestral and ornamental front, and a line above all that allows the ideas to breathe, as they are overwhelmed so often under the richness or the banality of t r  14  frame.  Here, as before, the contrast of the idea and its expression ("the frame") is drawn. Debussy's aversion to the German symphonic tradition as an example for French composers to follow is largely rooted in that tradition's predilection for developmental procedures that are applied indiscriminately to a diverse group of musical themes, rather than bom of an inner necessity. Debussy seems to locate this same fault in Mallarme's late sonnets. For Mallarme, the idea is not immediately readable in the work itself, but located beyond the work, in the realization that the work is ultimately about nothing other than its own processes, revealing the essential nothingness of the universe. The "principal piece or nothing" at the center of Mallarme's late poems is therefore, for the poet, precisely the point.' By 5  crafting works that do not seek to act as windows onto a semantic content, but rather try to place the relationships between words in the poem in the foreground, with the "meaning" of the poem existing, if at all, through the network of images and sounds that the words engage. So for Mallarme, the distinction that Debussy makes between the frame and the idea is irrelevant, since the poem's meaning cannot be abstracted from the particular circumstances of Debussy, Corr, 688: "Une idee musicale contient sa propre harmonie (ou du moins il me semble); sans cela, Pharmonie n'est plus qu'une chose parasite et maladroite." Debussy, Corr., 940-42. Ibid., 941: "Vous savez combien j'aime peu le developpement parasite, qui a trop servi a la gloire des Maitres pour que nous ne cherchions pas a le remplacer par un choix d'idees plus rigoureux, une ligne plus normalement soucieuse de la valeur qu'elles prennent sur l'horizon orchestral ou ornemental, et surtout, qu'elles y respirent, tant elles succombent si souvent sous la richesse ou la banalite du cadre." Emphasis mine. Mallarme, " L a Musique et les Lettres," OC II, 67. 12  13  14  15  11 its expression in language. In other words, it is not simply what the poem means, but how the poem means that is crucial. For if the poem springs forth sui generis from the relationships inherent in language itself, then there should be no distinction possible.  16  So both Debussy and Mallarme argue for essentially the same aesthetic: that the idea of an artwork should be expressed with a minimum of extraneous material. The only difference is that where Debussy locates the idea as a physical and acoustical reality, Mallarme's idea is a metaphysical absence that is expressed through verse. It is not surprising that Debussy does not see this aspect of Mallarme, for it was given formal expression mainly in the critical work of the twentieth century. For Debussy, in the poems of Mallarme's mature style, the means of expression do not find resonance with the subject matter of the poem itself: the "poverty of symbol" does not match the "workmanship" of the verse. What Debussy did not see, or did not value in Mallarme, was the poet's increasing abstraction from reality in the later works, his firm belief that the idea of a poem could in fact be a significant nothing rather than a significant something.  17  The only other significant mention in the correspondence of Debussy's attitude toward Mallarme's poetry and its possible influence on his music comes from 1913. Shortly after Debussy had set the Trois poemes de Stephane Mallarme, he discovered that the copyright for these texts rested with Edmond Bonniot, Mallarme's son in law. Writing to obtain permission to publish these settings, Debussy writes: I have piously conserved the most fervent admiration for the one who was "our master" ... He was, — without knowing it perhaps, —a considerable influence on the very quiet musician that I was in the era when he did me the honor of receiving me into his home. And this fervor grows as I recall in my memory  For instance, take the poem "Brise marine" (page 60 of the present study). The total significance of the poem is neither the failure of the poet to write poetry that the semantic dimension of the poem describes, nor the potential mastery of language that the play of letters and phonemes enacts, but a combination of these two elements that oscillate in the mind and ear of the reader. Jacques Derrida's seminal essay of Mallarme makes much of this clear, although Derrida's desire to emphasize the non-representational aspects sometimes obscures the relationship with objective reality that Mallarme is always at pains to preserve. See "The Double Session," 174-285. 16  17  12 his benevolent welcome to the music for L'apres-midi d'un faune [sic]. These memories, among many others, will entitle me, I hope, to ask you to facilitate the publication of these three songs? 18  The sentiment expressed in the letter certainly seems genuine, but we should not completely ignore the fact that Debussy has an ulterior motive here. Letters to Jacques Durand suggest that he was not amused by the situation, and may have been slightly offended at having to ask permission to publish these settings: "I had written to Dr. Bonniot before receiving your letter, — which is not important, and will never change my position vis-a-vis the solitary Mallarme!" Also, it seems strange that Debussy's "most fervent admiration" for Mallarme 19  would not have found some expression (compositional, literary or otherwise) in the fifteen years since the poet's death. If Mallarme was an enduring influence on Debussy after the composition of the Prelude, he  seems to have kept this fact to himself. This view is supported by the recollections  of Robert Godet, a close friend and confidant of Debussy. He writes: Did he [Debussy] spend much time with the Symbolists? Was he enthusiastic about Impressionism? The Debussy we knew never mentioned a word about such things, any more than he did about his visits to Mallarme's salon, which were perhaps not as frequent or as fruitful as is generally made out. One remembers only the respectful but amused look in his eye when he used to see the poet concentrating on the exploits of the Lamoureux Orchestra, and noting them down instantly in a little book whose pages were black with penciled jottings. 20  There is little reason to doubt either the accuracy of Godet's memory or his sincerity here. His remarks here do not imply any disdain on Debussy's part for Mallarme's poetry, but merely that his attendance at Mallarme's salon and its influence on his musical style needs careful  Debussy, Corr., 1650. Letter to Edmond Bonnoit, 7 August 1913: "J'ai conserve pieusement la plus fervente admiration pour celui qui fut « notre maftre »... II eut, — sans le savoir peut-etre, — une considerable influence sur le tres silencieux musicien que j'etais a l'epoque ou il me faisait 1'honneur de me recevoir chez lui. Et cette ferveur s'augmente de reconnaissance au souvenir de son bienveillant accueil a la musique pour L 'apres-midi d'un faune. Ces souvenirs, parmi tant d'autres, m'autoriseront je Pespere a vous demander de me faciliter 1'edition de ces trois melodies ?" Debussy, Corr., 1652. Letter to Durand, 8 August 1913: "J'avais ecrit au Docteur Bonniot avant de recevoir votre lettre, — ce qui n'a pa d'importance, et ne peut rien changer a ma position vis-a-vis du seul Mallarme!" Robert Godet, "En marge de la marge" in Revue Musicale 7 (1 May 1926): 63-64, reproduced in Debussy Remembered, trans, and ed. Roger Nichols (London: Faber and Faber, 1992) 36. 18  19  2 0  13  examination. The absence of Mallarme from Debussy's critical writings supports Godet's point about the composer's tight-lipped stance on these issues. In light of this, it is likely that Mallarme's influence on Debussy has been overstated by those who, desperate to locate Debussy's break with musical convention, have sought to make him a "figure" of Mallarme: a disciple whose ultimate failure to understand the full scope of Mallarme's vision mars his compositional achievement to some extent. This is the position taken by Laurence Berman, who argues that Debussy, unable to deal adequately with Mallarme's poetry during the composition of the  Prelude a L 'apres-midi d'un faune,  was  finally able to match the poet's complexity in Jeux. Berman simply assumes that, twenty 21  years after the composition of the  Prelude,  Debussy would still desire to produce in Jeux a  work that "seems to be the musical counterpart of Mallarme's poem ["L'apres-midi d'un faune"]." This attitude can be traced all the way back to the early 20 century. The desire to 22  th  see Debussy through a Mallarmean lens has thus coloured the critical appreciation of his music from its beginnings.  L I T E R A T U R E REVIEW: D E B U S S Y  Many critics have tried to show that Debussy's aesthetic was either molded in imitation of Mallarme, or at least shares an exceptionally close similarity to his works. Paul Dukas was one of the earliest critics to make this connection when he claimed that it "was the writers, and not the musicians, who exerted the strongest influence on Debussy." Arthur Symonds went 23  even further when he wrote that Debussy is "the Mallarme of music, not because he has set "L'Apres-midi d'un faune" to sound, but because the music has all the qualities of the poem Laurence Berman, "Prelude to the afternoon of a faun and Jeux: Debussy's summer rites" 19th-century Music 3/3 (March 1980): 225-238. Berman, "Debussy's summer rites," 227. Quoted in Robert Brussel, "Claude Debussy et Paul Dukas" La revue musicale 7 (1926): 101. 21  2 2  2 3  14  and none, for instance, of Verlaine." Symonds never actually specifies what these shared 24  qualities are, or what distinguishes them from Verlaine's poetry. Nevertheless, his assertion and many others like it have found their way into the musicological consciousness regarding Debussy. This sentiment is continued in the work of Stefan Jarocinski, who claims that Debussy's attendance at Mallarme's mardis was particularly important to the development of his aesthetic.  He sees Debussy's contact with Mallarme and with the Symbolists in general as  positive in the sense that Debussy's musical language was derived from "an absolute submission of music to a text." Perceptively, he recognizes the importance of both sound and sense in Mallarme's poetry. He argues that Debussy found musical correspondences for these categories, with the precepts of traditional harmony being equated with semantic meaning and the use of sound for its own sake as a symbol evoking a hidden idea. However, Jarocinsky 27  sees no fundamental connection between sound and sense in Mallarme's poetry. Instead, he sees merely a juxtaposition of elements, an alternation between passages driven by semantics and those driven by sound itself, rather than a consistent counterpoint of those two elements. In this, Jarocinsky does not always fully appreciate the intimate relationship between the sound of words and their meaning that is critical to understanding Mallarme's poetry. So when he sets up a similar duality in Debussy's style—where passages of traditional harmony (semantic) alternate with those conceived as pure sound—he bases his comparison on an incomplete understanding of Mallarme. If Jarocinsky's ideas about the juxtaposition of these elements in  Quoted in Edward Lockspeiser, Debussy: His Life and Mind, 2 vols. (London: Cassell, 1962-65), I, 120. As an interesting aside, Henri Gauthiers-Villars, who wrote under the pseudonym "Willy" made essentially the opposite claim: that Debussy was the "Verlaine of music, the equal of that other Verlaine, and one who also heard voices that no one else before him had heard." Quoted in Stefan Jarocinski, Claude Debussy: Impressionism and Symbolism, tr. Rollo Meyers (London: Eulenberg, 1976), 123. Jarocinski, Debussy: Impressionism and Symbolism, 88-89. Ibid., 118. Ibid., 149-50. 2 4  2 5 2 6  2 7  15  Debussy are indeed accurate, then he has proven exactly the opposite of what he intended. Mallarme's entire work is fundamentally concerned with the total integration of everything in poetry, and he accepts no less from himself or others. Francois Lesure has definitively established Debussy's biographical connections with 28  the major figures of literary symbolism in Debussy avant  Pelleas, ou les annees symbolistes.  His work completes and corrects many of Jarocinsky's observations, bringing them in line with the existing documentary evidence. Although he offers little in the way of musical analysis, his observations on Debussy's activities during these years give a more balanced picture of Debussy's relationship with Mallarme, one that posits the poet as one potential influence among many. More recently, Mallarme scholar Rosemary Lloyd has pointed out aesthetic similarities in Mallarme and Debussy's thinking that she implicitly attributes to the composer's attendance at Mallarme's  mardis.  She goes so far as to write that "the lack of dissonance in the thinking  of these two major figures of the late nineteenth century is exceptionally striking." Lloyd 30  explores Mallarme's and Debussy's shared mistrust of the influence that Wagner's music dramas were having on French artists. She also notes their respective mistrust of dogmatic theoretical explanations of their respective arts. However, much of her essay is devoted to brief biographical summaries of people whom both Mallarme and Debussy counted as friends or influences, including Pierre Louys, Andre Poniatowski, Verlaine, Henri de Regnier, Baudelaire and Poe. These vignettes may point future studies in interesting directions, but they add relatively little to understanding how Debussy's music relates to Mallarme's poetry, either in the specific settings of Mallarme texts or more generally. 2 8  2 9 3 0  Lesure, Debussy avant Pelleas. passim. Lloyd, "Debussy, Mallarme and 'Les Mardis'," 255-269. Ibid., 256.  16  From this biographical scholarship, a tradition of criticism has sprung forth that looks for common stylistic features between Mallarme's poetry and Debussy's music. Arthur Wenk's Claude Debussy and the Poets,  which includes significant chapters on both Prelude a I 'apres-  midi d'un faune and the Trois poemes de Stephane Mallarme  (1913), places heavy emphasis on  Debussy's formal responses to Mallarme's unusual syntax and grammar.  31  While Wenk's  analyses have much merit, they generally do not consider the possibility that there could be a fundamental difference between a Mallarme text and one by Verlaine, Baudelaire, or Banville. They do not consider the phonetic qualities of Mallarme's poems, the relationship between these phonetic qualities and their "meaning," nor how Debussy's musical setting responds to these elements in particular. For him, it seems that the function of a song is to frame an enunciation of the poem, and that the semantic and syntactic dimensions of the poem are of primary importance. For example, in his analysis of the song Soupir, Wenk shows all of the digressive clauses in the Mallarme poem and how Debussy's setting performs a "reading" of the syntax of the poem, clarifying the main clauses from those that are parenthetical. In 32  Wenk's analysis, Debussy does this through a series of harmonic digressions that parallel the syntax of Mallarme's original and allow a musical reconstruction of the "kernel" sentence on which the poem is based. He demonstrates Debussy's sensitive reading of Mallarme's grammar, but does not consider Mallarme's prosody, or the ways that Mallarme's poem seems to exist independently inside Debussy's setting. In order to understand whether there is, indeed, a distinction to be drawn between Debussy's treatment of a poem by Mallarme and one by another poet, we must begin with the fullest understanding of that text and its place in the poet's osuvre. Only then can we be reasonably accurate in describing the various ways that Debussy's music interacts with its text.  3 2  Arthur Wenk, Claude Debussy and the Poets (Berkeley and Los Angeles: University of California Press, 1976). Ibid., 246-54.  17  From there, it becomes possible to consider settings of other poets in a similar way and then finally to draw some conclusions. This level of detail is beyond the scope of Wenk's project, and indeed also beyond the scope of the current one. Verlaine's poetry, for example, is phonetically dense, and musical, in a way that is appreciably different from Mallarme, and Debussy's settings of Verlaine should probably be re-examined in that light also. Marianne Wheeldon has taken steps toward the approach that I am advocating here.  33  Looking at Soupir through the lens of Boulezian fragmentation, she argues that Debussy's setting invites an actual reordering of the musical elements of the piece, so that multiple pathways through the song are at least virtually possible, if not actually intended by Debussy himself. She calls this analysis "permutational," and attributes her method to Boulez, who in turn took it from Mallarme. Many aspects of Wheeldon's arguments are compelling, especially the sensitive way in which she describes the interaction of the actual musical narrative and the virtual pathways that both poem and music suggest. Her treatment of Mallarme's syntax and the way that it forces the reader to cast forwards and backwards across the text builds on similar ideas in Wenk, and I offer further comments on the poem later on. In the process, she gives the reader real insight into how reading a Mallarme poem is potentially different from one by another poet, and her work shows that Debussy seems aware of this particularly Mallarmean feature. However, she characterizes Soupir as built from sections that are "musically autonomous and nonteleological, so that the sections do not contribute to an overall contour or dynamic shape." I think her case is somewhat overstated in this regard, and in 34  Chapter Four I offer various kinds of continuities that link the various sections and work to counterbalance the disruptive musical surfaces that Wheeldon notes.  Marianne Wheeldon, "Debussy's 'Soupir': A n Experiment in Permutational Analysis" Perspectives of New Music 38/2 (Summer 2000): 134-160. . Ibid., 140. 3 3  34  18 In his dissertation A  Song Not Purely His Own: Modernism and the Pastoral Mode in  Mallarme, Debussy and Matisse,  David J. Code examines the relationship between Mallarme  and Debussy through "L'apres-midi d'un faune."  35  He argues that the unusual spacing of  certain elements in the typography of Mallarme's 1876 edition of the poem are designed to create a "virtual book," implying the recto and verso of pages independently of the actual pagination of the poem. He then argues that Debussy's  Prelude  "reads" this virtual book by  making corresponding shifts in orchestration in structural places that match, with unusual numerical precision, the "virtual book" that he has uncovered in the poem. His analysis of Debussy draws several exact parallels between individual measures in the  Prelude  and specific  lines of text in Mallarme's original. By expressly lining up particular measures in the  Prelude  with passages in the poem,  Code's analysis is a more elegant version of the mimetic model that has been used by other writers to align music and verse lines in the  Prelude.  Although he does not explicitly state it,  his idea of a virtual book in the Faun poem seems designed to evoke unfinished "Great Work." However, the structure of  Le Livre  evoking a virtual book inside a physical one. Instead, if Un  Le Livre,  Mallarme's  is not really concerned with  coup de des  is an accurate example  of the kind of writing that Le Livre would contain, the actual spacing of the text and its actual pagination would become significant elements, and the virtual elements they imply would not be structures inside the book—these would be explicit—but things external to it, particularly theatre and drama, which  Le Livre  was intended to replace. Of all of Mallarme's works,  "L'apres-midi" stands somewhat apart from  Le Livre.  It does manifest some of the same  Code, A Song not Purely His Own: Modernism and the Pastoral Mode in Mallarme, Debussy and Matisse, Ph. D Diss. University of California (Berkeley) 1999. A similar argument is presented in an article in the Journal of the American Musicological Society. See Code, "Hearing Debussy reading Mallarme: Music apres Wagner in the Prelude a l'apres-midi d'un faune" Journal of the American Musicological Society 54/3 (Fall 2000): 493-554.  19 concerns with theater and drama as the latter, but it stands more firmly in the French poetic tradition. In her 2003 book  Mallarme and Debussy: Unheard Music, Unseen Text,  Elizabeth  McCombie has called for an approach to music and text relationships in Mallarme and Debussy that abandons the notion of imitation. In its place, she argues for a more flexible approach that is neither properly literary criticism nor musicology, but a "mobile textual approach that is able to reconstruct the particular force of the intermediate ground [between music and literature] and its underlying dialogue of slippages and collusions, while at the same time insisting on the independence of the arts." Her primary interest is to develop a critical 36  language capable of exploring the ways that music and poetry interact. Mallarme and Debussy are test cases for this language, which is drawn in part from Mallarme's own critical writings and from other sources only tangentially related to the subject, like the works of Boulez, whose debt to both Mallarme and Debussy does not necessarily make his aesthetic relevant to a study of the two. McCombie's work is strongly influenced by Roger Pearson's homophonic approach to Mallarme's poetry. For this reason, she is sensitive to the phonetic aspects of Mallarme's verse, and particularly to moments where common phonemes unite various key words in some of Mallarme's poems. She makes useful observations about Mallarme's phonetic patterning, as when she identifies the common phoneme / O R / in several key words from the "Sonnet en >x:" "un or," "licorne," "decor," though her inclusion of "miroir" as an inversion is only partially true—the letters are reversed, but the phoneme itself changes. This confusion between letter 37  and phoneme is inherent to some degree in Mallarme's own writing on the subject, and he  Elizabeth McCombie, Mallarme and Debussy: Unheard Music, Unseen Text (Oxford: Clarendon Press, 2003), xvi. "' McCombie, Mallarme and Debussy, 28. 3 6  7  20  freely switches between various kinds of relationships that letters trace between words, whether phonetic or graphic. In other places, McCombie's observations seem forced, as when she analyses an excerpt from "Herodiade." She writes: In the 'Ouverture' d'Herodiade the simultaneous presentation and cancellation of an image and the multiplication of interpretative possibilities surrounding certain words creates an overdetermination of signification: Abolie, et son aile affreuse dans les larmes Du bassin, aboli, qui mire les alarmes, De l'or nu fustigeant l'espace cramoisi, Une Aurore a, plumage heraldique, choisi Notre tour cineraire et sacrificatrice This passage creates points of immobility through magnetic pulls of attraction and repulsion. The semantic space opened up by the initial act of repression or abolishing ('Abolie') is drowned again by its pursuit of a partner. The feminine 'Abolie' finds its mirroring reflection in the masculine version in line 2. The repetition both fills the emptiness created by the opening, in a matching sonority that has the effect of a double negative, and reiterates the sense of emptiness in a string of negative statements. Finding a masculine equivalent amounts to a reciprocal cancelling-out, yet the partnership gives birth to an overload of reflections in the line-final rhymes ('cramoisi', 'choisi'). 'aile' is 'Abolie' with its centre removed and its internal elements juggled. Sound patterns offer the promise of possible thematic centres, refuges from the pull between volume and emptiness. Yet to follow the path suggested by the phonetic patterning is to be misled. They are loci of stabilizing and destabilizing reflection, pools of verbal heterogeneity. The abolished pool is and is not reflecting 'les a' of'aboli(e)' ('dans les larmes', mire les a-larmes').  Here, the masculine-feminine pairing of aboli(e) is certainly relevant to Mallarme's prosody and the excerpt's negative semantic message. But to seriously suggest a connection (purely orthographic) between "abolie" and "aile" stretches the reader's credulity somewhat, and her assertion about the pool both reflecting and not reflecting the /a/s depends on a homophonic reading that she does not fully support. The passage can also (and more simply) be interpreted thus: The gold and crimson of dawn disappear ('abolie') as the sun rises, in a common Mallarmean theme of self-consumption. Its wing-like streaks are reflected in the pool, where they are also disappearing. This disappearance reflects our own fears about the impermanence of physical reality (life, death, etc.). This process is enacted phonetically, as the /a/s of the opening—aboli(e), affreuse, alarmes, etc.—fade into the his of "Aurore" (/ORDR/) as the dawn  21 spends itself. A's that no longer sound like /a/s simply enact the semantic message of the passage: a phonetic disappearance that matches the physical one described in the text. This phonetic game is not all-encompassing, nor is it meant to be. When it comes to the actual relationship between Debussy's music and Mallarme's poetry, there are some gaps in McCombie's work. Of the four Debussy songs that set Mallarme poems, she examines only Soupir and  Eventail  in detail. Instead, when she treats Debussy's  music, she prefers to deal with other works not specifically related to Mallarme: La mer, and some of the piano  Preludes.  Jeux,  Rather than looking for the actual intermediate ground  between Debussy and Mallarme, these studies are more concerned with the interaction of music and poetry in general. McCombie's fifty-nine page comparison of Mallarme's de des  Un coup  and Debussy's Jeux (1913) is based on the fact that they both disrupt established codes  of reading/listening in their respective fields. Yet again there is no particular reason to suggest that the composition of Jeux owes any particular debt (biographical, technical or otherwise) to Mallarme's poem. It is possible to cite numerous examples of discontinuous musical textures in twentieth-century music that would match up equally well, by McCombie's own criteria, with  Un coup de des.  The only possible reason for the comparison of these two works must  then rest on the tacit assumption that Berman also makes: that Debussy somehow aspired to the condition of Mallarme. In order to find an alternative to the various, essentially mimetic approaches described above, we must strike a balance between understanding what Debussy's setting of a Mallarme poem attempts to do, and how this reads through and across what the original poem attempts to  And unless Debussy had seen the original 1897 edition of Un coup de des in the journal Cosmopolis—which differs significantly in graphic appearance from Mallarme's intentions—he likely would not have known the poem until at least 1914, when Mallarme's collected works were published.  22  do. This requires a detailed understanding of Mallarme's poetry, particularly the ways that it differs from his contemporaries. For this reason, I now turn to Mallarme's works.  LITERATURE REVIEW: M A L L A R M E  Despite its age, Henri Mondor's  Vie de Mallarme  remains the standard resource for  biographical information on Mallarme. Mondor had access to numerous unpublished and 39  otherwise unavailable documents that have more recently become available through the work of Carl Barbier  (Documents Stephane Mallarme),  editor of the 8 volume collection of  Mallarme's documents, and Lloyd Austin who, along with Mondor, edited the massive 11 volume collection of Mallarme's letters. Recently, Mondor's 1945 edition of Mallarme's 40  CEuvres completes  Marchal.  has been supplanted by a two-volume CEuvres  completes  edited by Bertrand  41  During the century after Mallarme's death, Mallarme's thought has inspired a wide range of critical responses. The breadth and depth of this critical tradition far exceeds the space available in this forum to treat them fully. For this reason, I will only mention those works most relevant to my project. Several exegetical studies have shed much light on the question of meaning in Mallarme's oeuvre. Among these, works by Robert Greer Cohn  (Mallarme's Un  coup de des: An exegesis; Toward the Poems of Mallarme; Mallarme's Prose Poems, Igitur)  are significant. Colin presents a unified theory of Mallarme's poetry that assigns an absolute 42  semantic value to individual letters, based in part on their phonetic sound, in part on their Mondor, Vie de Mallarme, 2 vols. (Paris: Editions Gallimard, 1941-42). Carl Barbier, ed., Documents Stephane Mallarme, 8 vols. (Paris: Nizet, 1968-1998; Austin and Mondor, eds. Correspondence, 11 vols. (Paris: Editions Galimard, 1959-1985). A l l subsequent references to Mallarme's correspondence shall use the abbreviation Corr., followed by the volume and page numbers. Mallarme, OC I & II. Robert Greer Cohn, Mallarme's Un coup de des: An exegesis (New Haven: Yale French Studies, 1949); Toward the Poems of Mallarme 3ed (Houston: Scriverny Press, 2003); Mallarme's Prose Poems: a critical study (Cambridge: Cambridge University Press, 1987); Igitur (Berkeley and Los Angeles: University of California Press, 1981). 3 9  4 0  41  4 2  23  graphic shape. He deduces the signification of letters from their use in Mallarme's oeuvre as a whole, with pride of place given to Un coup  de des.  He then shows how, in various poems,  Mallarme reinforces the semantic message of the poems by expressing them through words that contain a conspicuous number of a particular letter (or phoneme). Occasionally, Cohn argues that the letter content of a passage modifies, or even contradicts, its semantic meaning. Cohn's theory of letters is laid out most clearly in  Un coup de des: An exegesis,  which is an  enlargement of his doctoral thesis, but the strategy informs virtually all of his subsequent work on Mallarme. Guy Michaud's Mallarme is organized as a biography but its true value lies in his thoughtful and detailed explications of Mallarme's poems. More recently, Bertrand Marchal published Lecture  de Mallarme,  an exegetical study of the major poems, in which he  summarizes' much of the work of previous critics like Emile Noulet and A. R. Chisolm and A.  offers many new insights of his own. Marchal's approach is mainly semantic, and he does not generally treat the prosodic elements of the poems as particularly significant.  43  Another important strand of Mallarme criticism deals with Mallarme's thought and aesthetic, which is expressed both through the poems and also in the numerous critical essays that Mallarme wrote in the 1880s and 90s. Jean-Pierre Richard's Mallarme  L'univers imaginaire de  is an intellectual topography of Mallarme's poetry. He systematically treats all of 44  Mallarme's favorite symbols and traces their nuanced meaning in a variety of poems to expose the more basic nature of Mallarme's aesthetic and epistemology. Reacting against Richard, Jacques Derrida has called into question the very notion of thematic criticism in Mallarme, suggesting instead in "The Double Session" that Mallarme's writing is essentially a-referential; that it sets up a series of intra- and inter-textual networks that constantly refer to other writings  Bertrand Marchal, Lecture de Mallarme: Poesies, Igitur, Un coup de des (Paris: Jose Corti, 1985). Jean-Pierre Richard, L 'univers imaginaire de Mallarme (Paris: Edition du Seuil, 1962).  24  (or objects) without the possibility of arriving at a stable and secure meaning. Derrida finds in 45  Mallarme the intellectual source of deconstruction, and an escape from writing as a system of representation that traces its heritage all the way back to the philosophy of Plato. Robert Greer Cohn has also written extensively on Mallarme's thought. For him, Mallarme's essential innovation was to modify the Hegelian dialectic to include a fourth pole, which he calls "antisynthesis." This "tetrapolar" pattern is essentially relational, since either of the pairs of polar opposites engaged in the dialectic may be brought together, considered for their similarity, then torn apart as the dialectic brings together the terms in the perpendicular pole  4 6  Cohn sits somewhere between Derrida and Richard. Although he believes that there is still a singular absolute meaning in a Mallarme poem (as does Richard), Cohn's own tetrapolar schematic comes very close to Derrida's notion of "undecidability" in Mallarme. Although Mallarme's critical essays are usually cited in support of a particular exegetical point in a poem, there are also important studies dedicated to the essays themselves. Cohn's Mallarme's Divagations: A Guide and Commentary explicates the essays collected in Divagations (1897). Essentially a work of translation, but peppered with explanatory notes, 47  Cohn takes pains to show how consistently Mallarme's critical writing supports his own ideas about Mallarme's epistemology. Cohn's work is nearly as difficult to navigate as the original Mallarme essays, but offers a way through the texts that is indispensable for anyone encountering these works for the first time. Judy Kravis has also written on Mallarme's prose  Derrida, "The Double Session," 173-285. Cohn's conception of Mallarme's thought is explored in Chapter One. Cohn, Mallarme's Divagations: A Guide and Commentary (New York: Peter Lang, 1990). A new translation of Divagations by Barbara Johnson was published too late to be incorporated into this study, but should be consulted by those interested in these works. See Mallarme, Divigations, trans. Barbara Johnson (Cambridge, M A ; Belknap Press, 2007).  4 6  25  essays. Her work has the added advantage of dealing with significant works like "La Musique 48  et les Lettres" that do not appear in Conn's book. Finally, Mallarme's unique use of the French language is explored by Gerard Genette in his brilliant book Mimologics. Genette explains Mallarme's conception of language in light 49  of Cratylism, which can be loosely defined as the belief that the sounds of words carry some authentic trace of the objects and ideas that they name. He locates Mallarme's particular brand of Cratylism in the context of numerous theories of the French language that engage this particular type of mimology. Ultimately, Genette argues that Mallarme sees the French language as derived from an original language that was essentially mimetic. This original language has been lost through the passage of time, through the grafting of one language into another, to the point that contemporary language no longer functions mimetically. Further, this language—or fragments of it—can be discovered in contemporary language. Genette claims that Mallarme's Cratylism is essentially worked out at the level of the verse line in his poetry, where various rhyme and rhythmic gestures compensate for the mimetic defects of contemporary French. The verse line is very important for Mallarme, but Genette's dismissal of the word as an important element in this regard seems unnecessary.  S O U N D IN M A L L A R M E  Like all poets, Mallarme works essentially in two media at once. On one hand, he is a word-smith, and his poems have an acoustic reality that cannot be denied. Even a silent reading of Mallarme's poetry calls forth the sonorous nature of his words, words that have a particularly poignant sound. On the other hand, Mallarme works with meanings. His poems are  Judy Kravis, The Prose of Mallarme: The Evolution of a Literary Language (Cambridge: Cambridge University Press, 1976). Genette, Mimologics: Voyages in Cratylusland, trans. Thai's E. Morgan (Lincoln: University of Nebraska Press, 1995), 202-18.  4 8  4 9  26  all about something, even if that something is apparently trivial, and regardless of the layers of analogical reading that they invite (or that critical tradition has foisted upon them). What separates Mallarme from the other poets of his generation is not the complexity of his poetry: that is merely an effect of his aesthetic. Indeed, Mallarme always maintained that his poems were clear (not easy) to anyone who knew how to read. The particular quality that distinguishes Mallarme from his contemporaries is the exceptional care with which he employs the sounds of words so that they have a significant relationship to the semantic meaning that the poem carries, and by extension with its analogical meanings as well. Throughout his correspondence and critical writings, Mallarme constantly equates this aspect of his poetic practice with music, using musical metaphors and imagery to describe the phonetic relationship between words and referring to his poems as "musical." Although most Mallarme scholars mention the importance of individual letters or phonemes at some point in their work, there are three more or less systematic treatments of the subject. Robert Greer Cohn argues that Mallarme gives a particular semantic meaning to individual letters, a meaning that adheres not only to their sound but also their graphic shape as well. His letter table is largely based on Mallarme's use in Un coup de des, which he then 50  applies with a fair degree of consistency to Mallarme's larger oeuvre. Roger Pearson's interest in phonemes is directed mainly towards homophony, which Pearson posits as essential to Mallarme's pursuit of linguistic mastery in his poems. He adduces several convincing cases 51  where Mallarme's lexical range and syntax are designed to force the text to speak with a double voice, and consequently to force the listener to hear two simultaneous parsings of phonemes: one that Mallarme's spelling engenders on the page, and others, which the ear can  Cohn, Mallarme's Un coup de des, 35. Roger Pearson, Unfolding Mallarme: The Development oja Poetic Art (Oxford: Clarendon Press, 1996). By homophony, Pearson refers to homonyms - words that sound the same (or nearly so) but have different meanings. 5 0  51  27  perceive in spite of the written disposition of the word. There are moments in Pearson's work, however, where he sees the formal manipulations of individual phonemes for their own sake, rather than in the service of an alternate semantic reading. Graham Robb sees Mallarme's use of phonemes through the lens of French prosody. He argues that Mallarme's obsession with 52  words that have few (or no) rhymes forces him to consider other phonemic echoes (assonance, alliteration, etc.) that compensate for their relative poverty in this regard. These echoes, drawn from key words or images in the poem, bind the work together and are responsible for the unique and often bizarre sonic landscapes of Mallarme's poems. Isolated references to music can be found in all three approaches, but none of them demonstrate how Mallarme describes the phonetic relationships of language as music in a systematic way. For Mallarme, aural similarity between words (including rhyme, assonance, alliteration, and other resemblances) indicates a hidden connection between them.  53  This  connection is both semantic and phonetic, although the semantic connection may not be immediately obvious. Mallarme sees in these various forms of rhyme an outline of the essential unity of language in its original state. He imagines a single generative language that is the unique source of contemporary dialects, a language in which there was no distinction between poetry and language since poetry was language and language was poetry. At some point in the distant past, this original language was broken apart, scattered across and inside the world's languages through evolution, war, cross-cultural influence, and similar factors. Modern languages are therefore broken: they do not consistently display phonetic relationships between words that are semantically related, and vice versa. This defect of language creates the need for the poet, whose purpose is to reconstruct this generative language through an exploration of  Robb uses the term "prosody" to refer to the conventional codes of French poetry that Mallarme both knew and exploited in his poems. I will use the term in the same sense throughout the present study. See Robb, Unlocking Mallarme (New Haven and London: Yale University Press, 1996), 3-31. Mallarme, Les mots anglais, OC II, 1014. 5 2  5 3  28  those fragments that remain intact. This, at its core, is Mallarme's notion of musical language, and the basis for virtually all of his later writings on music and poetry. If, as Robb asserts, Mallarme had a preference for words that rhyme with few or no others, this may have been because the essential unity of language could best be reconstructed through these words. A word that rhymes with a few others, say a dozen or less, gives a relatively small group of semantic, etymological, orthographic and ideographic meanings from which to extract a unifying thread. Words with dozens or even hundreds of rhymes often fall 54  victim to language's propensity to lie, to suggest relationships that turn out to be dead-ends or endless semantic labyrinths that make extracting the essence of particular phonemes very difficult. For example, there are only eight possible rhymes in French on the phonemes  /IVR/:  ivre - guivre - vivre - givre - delivre - livre - cuivre - suivre. Of these, five stand out as 55  significant in Mallarme's poetry: ivre - vivre - givre - delivre - livre. The first four of these are used as rhymes in two of Mallarme's poems: "Le vierge, le vivace et le bel aujourd'hui" and the unpublished "Eventail" for Mery Laurent; "Brise marine" rhymes livre/ivre, while "L'apres-midi d'un faune" rhymes ivre/delivre. Among these five words, various semantic connections can be drawn. The intoxication of "ivre" is caused by the inner life of words ("vivre"), which are immobilized ("givre") in contemporary language until freed by poetry ("delivre"). The totality of the project is to be contained in the poet's Great Work, Le Livre. 56  The relation between the phonetic structure of words and their signification is the problem that Mallarme confronts in Les mots anglais, where he attempts to reconstruct, in the English language, the "relationships between the total signification and the letter" that point to By ideographic, I refer to the meanings that could be adduced from the shapes of letters in these groups of words. Derivative rhymes like "enivre" (from "ivre") and "poursuivre" (from "suivre") are not customarily used in poetry, and therefore are not counted in the eight rhymes. See Phillipe Martinon, Dictionnaire des rimes francaises, precede d'un traite de versification (Paris: Larousse, 1962) s.v. "/ivR/" 5 4  5 5  5 6  Although inspired by Robb's approach in Unlocking Mallarme, these ideas are my own.  29 their "common origin" in a primitive language. His strategy is to begin with words that share 57  an initial consonant, a significant letter (or phoneme) cluster in the remainder of the word, and a semantic link. These words call out to one another in constellation, suggesting a common origin that is more than etymological, but points instead to an original language. Yet this reconstruction fails, by and large, to produce stable significance for practically any letter, precisely because the proliferation of rhyme gathers so many diverse meanings around each initial consonant position that the various constellations cannot all be reconciled. The essential plurality of language casts a peculiar importance therefore on the isolated words, those for which Mallarme finds no significant constellation. Nevertheless, Mallarme believes that these isolated words include some of the most important ones in the language. Mallarme claims that it is the writer's duty to reconstruct these alliterative constellations, "to relate some terms whose unity contributes all the more to the charm and to 58  the music of language."  Of the three authors mentioned above, Cohn sees this duty most  clearly. However, by trying to force a stable signification retroactively on letters (from Un coup de des  backwards onto the larger oeuvre), his system becomes somewhat dogmatic, with  the individual letters and phonemes speaking the same message in a dizzying variety of contexts. Robb's approach is more flexible, and allows him to respond to the actual constellations that Mallarme's poems bring together without constantly relying on an a priori signification.  Mallarme, Les mots anglais, OC II, 969: "rapports entre la signification totale et la lettre"; "origin commune." Ibid., 967: "de rapprocher des termes unis avec d'autant plus de bonheur pour concourir au charme et a la musique du langage."  30 M u s i c IN M A L L A R M E  Throughout his life, Mallarme consistently compared his poetry to music. In this way, Mallarme emerges from and participates in a long tradition of nineteenth-century French poetry that found inspiration in musical works, notably Baudelaire and Verlaine. By the 1890s, he argued that his poetry was more "musical" than sounding music itself, and the confrontational nature of his attitude toward music sets him somewhat apart from many of his contemporaries. Since it is my purpose to trace the interaction of Mallarme's poetic "music" with Debussy's settings of his poems, critical works that examine the role of music in Mallarme's oeuvre are particularly relevant to my thesis. Perhaps the most comprehensive treatment of music in Mallarme to date is a 1959 dissertation by Suzanne Bernard, Mallarme et la musique!' A good portion of the work deals 9  with Mallarme's relationship with Wagner as mediated through French wagnerisme. Bernard recounts Mallarme's attendance at the Concerts Lamoureux, which she argues permitted Mallarme to "refine his ideas on music and on the relationships existing between music and literature." She argues that essential features of Mallarme's literary technique—his penchant 60  for alliteration and other alternatives to end-rhyme—"seem at times to apply the poetic precepts of Wagner." She connects the Mallarmean ideals of allusion and suggestion to 61  music, as well as the importance of silence. Some of her best insights relate to the essentially interconnected nature of all things in Mallarme's aesthetic, and the value that he assigns to music in this process. However, Bernard's treatment of Mallarme's "musicalization" of poetry is less successful. For her, Mallarme's poetic music is found primarily in the typographical  5 9  6 0  5 1  6 2  Suzanne Bernard, Mallarme et la musique (Paris: Nizet, 1959) Ibid., 22: "preciser ses idees sur la musique et sur les rapports existant entre la musique et la litterature." Ibid., 32: "parait quelquefois appliquer les preceptes poetiques de Wagner." Ibid., 43-49.  31  disposition of words on the page, both in  Un coup de des  and other poems. She calls this a  formal architecture evident on the surface of the page and in the ordering of the book of verse, which she compares to the physical disposition of the orchestra on stage. In the broken 63  alexandrines of "L'apres-midi d'un faune", Bernard sees a process of "musicalization" whose blank spaces are analogous to the silences in instrumental music. She also comments on the 64  numerous enjambments and broken rhythms that read across the alexandrines, which she correctly identifies as an example of Mallarme's "musical" approach to rhythm in this poem. Yet Bernard constantly lapses into vague metaphor: she claims that the poem's structure may be called musical because it combines three themes: sensuality, musicality and intellectuality. If the sole requirement for the "musicality" of a poem is that it combines multiple themes, then virtually any poem would be musical. More interesting is Bernard's treatment of music in the then-recently published sketches for Le  Livre,  Mallarme's unfinished masterwork.  65  She identifies this work as  Mallarme's ultimate response to the Gesamtkunstwerk, a synthesis of Theater, Music and Poetry. Here, she compares the combination of abstract ideas found in the geometrical 66  diagrams or "equations" of the notes for Le Livre to the essential nature of music, in that these figures combine various intellectual "themes" (Theater, Drama, Hero, Hymn, Mystery, Poetry, Dance, the Book) in binary oppositions. These are then varied and repeated, brought together to display the unity that underlies their apparent diversity. She asks "What does Mallarme do [in these "equations"] if not combine themes, forms, like a composer?" Bernard cites 67  passages in the notes that describe a double performance of the projected volumes of Le  Livre,  in which "each volume [is] 'presented two times like its two halves, first the one and the other 6 3  Bernard, Mallarme et la musique, 87.  6 4  Ibid., Ibid., Ibid., Ibid.,  6 5  6 6  6 7  104. 135-48. 136. 140.  32 last, juxtaposed with the last and the first of the one and the other'; each volume forms thus ... a symphony of 384 pages (4 x 96)." The relational patterns of this multiple reading would 68  thus recast each volume according to its surroundings, much like a musical theme's signification is dependent on its context and surroundings. Bernard concludes that Le Livre represents Mallarme's effort to fuse the mobility inherent in music with the permanence of literature. This mobility is enacted as a reading strategy in which "vertical" and "horizontal" readings of the text produce various nuances of meaning comparable to music. In this text, one can see the "poet's persistent desire to 'take back' from Music that which seems to be the very essence of this sonorous form, its movement and the perpetual transformation of its themes." This is much closer to Mallarme's vision 69  than the derivative Wagnerian pastiche that she invokes elsewhere in the work. As we shall see, Mallarme's poetry does in fact require such a novel approach to reading, although it need not be restricted to Le Livre or Un coup de des, but usefully informs his entire oeuvre. Phillipe Lacoue-Labarthe also sees Mallarme through a Wagnerian lens. He argues 70  that "between [the essay] "Richard Wagner: Reverie of a French Poet" [1885] and "Music and Letters" [1894], everything, or almost everything, was collected and comprehended in a project that seems to have found its origins in the 'singular challenge' launched by Wagner." His 71  essay, which focuses exclusively on Mallarme's critical writings, posits Mallarme's entire aesthetic as a sort of negative image of the Gesamtkunstwerk: formed primarily in response to what Mallarme perceived as Wagner's failings. Lacoue-Labarthe recognizes that Mallarme's ultimate rejection of music in favour of poetry is based on the idea that actual concerted music Bernard, Mallarme et la musique, 140: "chaque volume, « presente deux fois en tant que ses deux moities, premiere de I'un er derniere de l'autre, juxtaposees a derniere et premiere de l'un et de l'autre » ; chaque volume forme ainsi...une symphonie de 384 pages (4 fois 96)." Ibid., 141. Phillipe Lacoue-Labarthe, "Mallarme" in Musica ficta: Figures of Wagner, tr. Felicia McCarren (Stanford: Stanford University Press, 1994), 41-84. Ibid., 45. 6 8  6 9  7 0  71  33  is merely a copy of an ideal music (that Lacoue-Labarthe calls archi-music) that it is not fully equipped to represent. He identifies language as the true site of Mallarmean music, though in 72  his conception, the musical aspects of Mallarme's language are entirely rhythmic and take no particular account of language's phonetic properties. He concludes that, for Mallarme, "[v]ersification is thus the restitution of Literature as archi-music—this archi-music of which 73  music is itself only the imitation or the (too) sensual presentation."  Lacoue-Labarthe's basic  premise—that Mallarme responds to the challenge of music by subsuming it into verse—is correct in my view: Mallarme does see his poetry replacing orchestral music with the silenced "music" of written verse. However, by setting aside issues of rhyme in its most expanded sense and valorizing rhythm, Lacoue-Labarthe does not consider one of Mallarme's most crucial poetic techniques: the manipulation of phonetic elements and the letters that represent them, which he is constantly describing as a musical process. And since, finally, Lacoue-Labarthe provides no example of how one might read Mallarme's poetic works in light of the "archimusic" created by rhythm, his argument never leaves the theoretical plane, as if Mallarme were first and foremost a philosopher and not a poet. Several useful contributions to the study of music in Mallarme have also been made by musicologists. In his article "Sea-Changes: Boulez's Improvisations Sur Mallarme," James McCalla offers a sketch outline of music in Mallarme's thought.  74  He notes that in music  Mallarme detected certain literary procedures, and that the poet's duty was to return these to language, their original source. In a particularly suggestive passage, McCalla argues that 75  Mallarme saw the play of ideas in his poetry in musical terms, specifically invoking the notion Lacoue-Labarthe, Musica ficta, 79. Here, Lacoue-Labarthe invokes Derrida's notion of arche-writing. See Derrida, OfGrammatology trans. Gayatri Chakravorty Spivak (Baltimore: Johns Hopkins University Press, 1998), 27-73. Lacoue-Labarthe, Musica ficta, 81. James McCalla, "Sea-Changes: Boulez's Improvisations Sur Mallarme" Journal of Musicology 6/1 (Winter, 1988): 83-106. The portion dealing with Mallarme can be found on pages 83-91. McCalla, "Sea-changes," 85. 7 2  7 3  7 4  7 5  34 of counterpoint and harmony to describe this aspect of his works. However, the example he chooses for this is somewhat disappointing. In "L'apres-midi d'un faune," McCalla identifies a "literary counterpoint" within the poem's structure between pairs of literary themes: "passion and fear, colors, light and shade, and so on." A similar "counterpoint" of themes could be 76  traced in virtually any piece of literature, leaving the reader to wonder how unique "L'Apresmidi" is in this regard. More promising is McCalla's description of Mallarme's oeuvre as a reduction of language that concentrates on "constellations" of inter-related words. His brief analysis of Mallarme's "Sainte" traces the way that parallel words and images from the first half of the poem are made to disappear in the second half, and the peculiarly static quality that this creates in the work. He shows how Mallarme's syntax and the semantic plurality of some of the poem's key words make it more allusive than the poetry of his contemporaries. He calls this process of evocation and disappearance in the inexhaustible play of relationships the "music of 77  silence." Although he makes passing reference to Mallarme's use of phonemes, he does not really consider how these elements inform and shape Mallarme's notion of the constellation. One of the few scholars to fully recognize both Mallarme's attraction to and antipathy towards music is Mary Breatnach. In a study of the poet's influence on the composer Pierre Boulez, she criticizes approaches to Mallarme that "mistake the poet for some sort of composer manque."  She also warns of the dangers of misrepresenting Mallarme's ideas about music by  taking individual passages out of context, and stresses the need to approach these ideas from a comprehensive view of the poet's works. Breatnach recognizes that an essential part of 80  McCalla, "Sea-changes," 87. Ibid., 89. Mary Breatnach, "Music and Mallarme's Aesthetic" in Boulez and Mallarme: A Study in Poetic Influence (Aldershot: Ashgate, 1996), 20-69. Ibid., 21. Ibid.,22. 7 6  7 7  7 8  7 9  80  35  Mallarme's poetic technique depends on establishing "an intricate and complex series of relationships between the words [of a poem] themselves" in order to free them from their conventional use in ordinary language. She demonstrates, in various poems, how Mallarme's 81  syntax, grammar and punctuation create such relationships. Occasionally, she alludes to the importance of the sound of words and their individual phonemes, although this notion is pushed much farther by other critics. Perhaps her most valuable insight, from my point of view, is her recognition that Mallarme uses the term "music" to imply, at various times, either the sounding notes of instruments and voices or a series of (essentially mental) relationships. However, she insists on minimizing the role of sound in the second of these cases, claiming that Mallarme makes "a distinction ... between two uses of the word 'music,' one signifying the art of sounds, the other indicating the abstract, silent, structural force whose exploitation and revelation is the poet's duty." Focussing little on the sound of words, Breatnach proceeds with a reading of Un coup de des that has many merits, including a sensitivity to the visual layout of the words on the page, which she identifies as musical. But by not considering the materiality of Mallarme's language—the sound of the words themselves—she misses an important aspect of Mallarme's thinking about music. At the most basic level, a poet uses words as sounds. Everything flows from his ability to see connections in sound between words that can also be arranged to mean something semantically. For a poet who believed in the supremacy of poetic language as firmly as Mallarme did, and played with phonemes and other sound patterns (rhythm for one) as regularly, it is inconceivable that the sound of words did not matter to him. That this aspect is in dialogue with the other elements that Breatnach identifies is certainly true, but our 81  Breatnach, "Music and Mallarme's Aesthetic," 27.  36  understanding of Mallarme's poetic practice, his ideas about music, and their potential influence on later figures, especially musicians, is incomplete without reference to the sonorous material of language.  P R O B L E M S A N D SOLUTIONS IN M A L L A R M E A N D D E B U S S Y  From the preceding pages it would seem that Mallarme and Debussy have significantly different ideas about music, poetry and the proper relationship between the two. For Mallarme, the goal was to create poetry of such purity that it would itself be music. In one of his last published descriptions of the Le Livre, Mallarme claimed that "Poetry, close to the idea, is Music, par excellence—admits no inferiority." This poetic music, freed from the typical 83  execution of strings, brass and winds, absorbs all of the characteristics of instrumental music inside of itself in a total synthesis of everything: language, music, dance, the plastic arts and indeed the whole of the universe. Le Livre is to be nothing less than the tracing of all relationships of the universe contained in the book, that singular text that has been the unconscious effort of all writers throughout history. In Mallarme's aesthetic, music must be 84  subsumed by poetry, transposed from the symphony to the book. This is why all of the 85  musicians in Mallarme's oeuvre are silent: the "musicien du silence" ("Sainte"), the "creux neant musicien" ("Une dentelle s'abolit"), even Wagner himself, who is "mal tu par l'encre as  meme" ("Hommage"). It is not exactly that Mallarme values silence, but rather that he values that which once sounded but has been silenced through writing.  Mallarme, "Le Livre, instrument spirituel," OC II, 226: " L a Poesie, proche l'idee, est Musique, par excellence—ne consent pas d'inftriorite." Ibid., 224: "Une proposition qui emane de moi ... sommaire veut, que tout, au monde, existe pour aboutir a un livre." Mallarme, "Crise de vers," OC II, 212. Mallarme's doubts about whether poetry could actually accomplish the silencing of music—the larger issue for which Wagner was in some ways merely a symbol—explains the "mal" in this final quotation. 8 i  8 4  8 5  8 6  37 For Debussy, the goal in his texted musical works was rather to create an artistic response to a pre-existing poem that was neither a mere juxtaposition to, nor an imitation of, that text. Debussy does not argue for a systematic poetification of music, nor does he seek to replace poetry with music entirely. Instead, he is interested in the simultaneous presence of these two signifying systems, these two artistic "languages," and the ways that they write (and read) over one another in the shared space of song. So when a Mallarmean text comes into contact with a piece of music by Debussy, there is, necessarily, a complex interaction between the two systems that includes moments of congruity and discord, imitation and ambivalence. By studying the moments of tension and accord in these works, it should be possible to arrive at an accurate understanding of Mallarme's importance for Debussy's musical style. In this study, I have restricted myself to examining only those works in which Debussy's music explicitly engages a Mallarme poem. In this way, I hope to avoid falling into the trap that awaits more general stylistic readings of Debussy's larger output: mistaking similarity for causation. I do not mean to argue that comparisons between Debussy's compositional style in general can not be made with features of Mallarme's poetic style, or that such comparisons are not fruitful. However, when it comes to understanding which features of Debussy's style can be traced back to his contact with Mallarme's poetry, we must clearly distinguish between similarity and influence, even if that influence is manifested in a modification or outright rejection of the original source. In those works where the composer has authorized a comparison, whether by bringing music and text together in song, invoking poetry in the titles of pieces, or referring to a literary work in a body of paratextual evidence (letters, marginalia, etc.) we should bring our best critical faculties to bear on understanding both the compositional and aesthetic relationships between the two works. In all other cases,  38 we must be extremely cautious to avoid confusing similarity in the mind and ear of the listener for imitation or influence in the compositional process. In order to fully understand the artistic relationship between Debussy and Mallarme, the limits of their mutual understanding must be articulated.  First, the central features of  Mallarme's creative enterprise must be described. This must include not only the aesthetic positions outlined in Mallarme's critical essays, but also their specific application in his poetry. My approach to Mallarme's poetry is, therefore, situated somewhere between Cohn's and Robb's. Taking my cue from Mallarme's method in Les mots anglais, I look for coincidences of sound and sense that Mallarme consciously exploits in his poetry. In many cases, these relationships are introduced through end-rhyme and reinforced in the surrounding syllables of the verse; however, there are several cases (documented by Robb) in which a word with no real rhyme appears in the middle of a verse line, scattering its phonemic elements across the poem. From the sound-sense relationship in these clusters, I read outwards to the apparent significance of the poem, which is always in a dialectic relationship with its phonemic qualities. Secondly, the works by Debussy that use or refer to a specific text by Mallarme must be examined to see whether Debussy responds in any meaningful way (whether positively or negatively) to the features of Mallarme's work and aesthetic. There are five such compositions: Apparition Mallarme  (1884), Prelude a l'apres-midi d'un faune (1892-3), and Trois poemes de Stephane (1913; Soupir, Placet futile, Eventail). What emerges from this study is a more  complex image of Mallarme and Debussy than has previously been put forward, one where the attraction, interaction and tension between music and poetry in their respective works stands as a microcosm of the larger relationship between Music and Language in general.  39  Paradoxically, Debussy's understanding of Mallarme's poetry is intimately connected to his understanding of Mallarme's "music" as well, a music far different from anything he had encountered before. Whether he gained this understanding intuitively through an attentive reading of his poetry, gleaned it through repeated readings of Mallarme's critical essays, heard it at the mardi gatherings, or pieced it together from the opinions of those in the Parisian literary milieu, close readings of Debussy's Mallarme pieces show the composer by turns embracing and rejecting Mallarme's poetry, trying to find a musical counterpart for these poems that was more than just a bad pun. What emerges is an approach to text setting in which the essential qualities of the poems are retained inside the songs through a vocal line that acts like a recitation, while the surrounding musical lines speak with their own logic and purpose.  S T R U C T U R E OF T H E DISSERTATION  Chapter One examines Mallarme's ideas about music and language. First, I locate Mallarme's essential conflict with music in the context of his spiritual-epistemological crises in the late 1860s, fully fifteen years before his "discovery" of Wagner. 1 argue that the music of 87  Mallarme's language is based first and foremost on the phonetic similarity between words. Shared phonemes between words—rhyme in its most elemental form—trace hidden relationships between ideas and objects in the world. These relationships point to an original language in which sound and sense were in perfect accord. Therefore, thought and language were also in perfect accord, and language was the expression of thought in the world. While I agree that there is a surface resemblance between some of Mallarme's techniques and the ideas of Wagner, there is little evidence that he actually read any of Wagner's writings (in French translation) until 1884. While some of his close friends—Villiers de L'Isle Adam, for example—had made the trip to Bayreuth, there is no indication in the frequent letters between the two that they ever spoke about Wagner's music, much less his more theoretical discussions of drama, poetic language, Stabreim, etc. As Genette has shown, many of these ideas stretch back to Plato and run through French theories of language for hundred of years before Mallarme. Very few people (Genette and Jacques Michon excepted) have tried to see Mallarme's treatment of language in light of these traditions. See Genette, Mimologics and Michon, Mallarme et "Les mots anglais" (Montreal: University of Montreal Press, 1978). 8 7  40 Somewhere in the unconscious history of humanity, languages began to multiply and change, a Tower of Babel scenario that allowed chance to infiltrate words. The unity of sound and sense in language was fractured and scattered across and inside the world's languages through the passages of time, war and conquest, and linguistic evolution. The poet's task is to reunite these fragments of language in which sound and sense are conjoined, and to work thus towards uncovering the essential signification of phonemes and the letters of the alphabet charged to represent them. Verse, which Mallarme understood broadly to include all forms of literature, 88  has always done this whether explicitly (through rhyme) or implicitly (in prose). In the process, I draw on evidence from Mallarme's letters, sketches for a planned dissertation on the word published posthumously as "Notes sur le langage," and his explication of similar phonetic reconstructions in Les mots anglais. I then turn to the more famous critical essays in order to show how consistently Mallarme sticks to his guns in the face of the Wagnerian movement. I also demonstrate how Mallarme's musical language functions in verse through a critical examination of selected poems, thereby grounding the more philosophical discussions that have received considerable critical attention in a poetic practice that has received relatively less attention. The remaining chapters examine all five of Debussy's compositions to Mallarme poems chronologically. Chapter Two treats Debussy's early exposure to Mallarme, which resulted in the song Apparition (1884). Therein, I provide a summary of all of the Mallarme poems that we can reasonably assume that Debussy read to that point in time, and argue that his choice of "Apparition"—a poem in Mallarme's early style—shows a degree of ambivalence towards Mallarme's mature style. I then show how Debussy's harmonic language in the song is designed to respond analogically to the semantic meaning of the poem, but not necessarily to This is Mallarme's position, expressed in many of the late essays, including " L a Musique et les Lettres" OC II, 64: "le vers est tout, des qu'on ecrit" ("verse is everything, from the moment that one writes"). 8 8  41 its particular phonetic effects. Chapter Three offers a new and non-mimetic reading of the Prelude a I'apres-midi d'un faune  (1892-4). Rather than following the poem line by line or  section by section, I see the Prelude in the context of its origins as a theatrical production. In late 1890 or early 1891, Mallarme engaged Debussy to write an overture, and perhaps incidental music also, for an upcoming performance of "L'apres-midi d'un faune." Although the performance never in fact took place, and Debussy completed the score much later, I argue from biographical and musical evidence that the Prelude is related to Mallarme's unique conception of theater as he must have imparted it to Debussy for their collaboration. Basically, I see the Prelude as a representation of the Idea of the poem rather than a representation of its text, for which Debussy's opening flute solo functions as an abstract musical symbol. The way that this melody and its important pitches recur throughout the Prelude is explained as an example of the arabesque, an enactment of essential processes of thought, which can also be traced in the poem and in Mallarme's aesthetic. Chapter Four examines the Trois poemes de Stephane Mallarme  (1913), which represents Debussy's final engagement with Mallarme's  poetry. In these songs, I show the degree to which Debussy understood the phonetic elements of Mallarme's style and how this awareness, along with more traditional notions of text-setting, shaped the individual compositions to allow the music of the poems to be audible inside the songs. I will also show moments where Debussy's music reads across Mallarme's texts, and moments when the reverse is also true. In the process, I sketch the limits of the relationship between music and poetry in Debussy and Mallarme, and between the two arts more generally. Chapter One stands in a somewhat uncomfortable relationship to the other three. The ideas I explore there are necessary to establish Mallarme's basic attitudes towards music and language, and therefore inform all of the texts that Debussy set to some extent. However, the extent to which Debussy knew about, understood, or agreed with Mallarme's theories is  42 impossible to establish directly, and must be inferred from the pieces he composed. For this reason, not every idea in Chapter One finds expression in the later chapters. The twin structure thus implied is, I think, emblematic of the way that music and text relate in these pieces in particular, and on more abstract levels. I hope that those with an interest in Mallarme will find some of the observations in this chapter useful on their own merits. For those interested in Debussy, the later chapters offer some new insights into the music itself, and the way that it interacts with the texts.  METHODOLOGY  In dealing with Mallarme's oeuvre, I use close readings of a number of primary texts. In this way, I hope to show that anyone who reads Mallarme patiently and carefully—as the poet demanded—may well draw conclusions similar to my own. I have also restricted myself to those texts that it is reasonable to assume that Debussy might actually have read (i.e. those published in Parisian journals), or to those unpublished texts that provide such profound insight into Mallarme's aesthetic or poetic practice that they are indispensable to the topic. As such, it is reasonable to assume that Debussy would have had access to them from Mallarme himself, or from one of their several mutual friends and acquaintances. Of course, my ideas about Mallarme have been indelibly shaped by my contact with several important secondary sources, most notably the work of Robert Greer Cohn, Guy Michaud, Jacques Derrida, Roger Pearson, Graham Robb and Bertrand Marchal. In their work, I continue to find revelations and new directions to pursue. In approaching Debussy's music, I use a rather free interpretation of traditional harmonic theory, Schenkerian perspectives and pc-set theory wherever they produce useful insights. I have no particular theoretical position to espouse, nor do I believe that one particular  43  analytical technique is universally valid for Debussy's oeuvre. It seems to me that a composer who scorned system as thoroughly as Debussy cannot be completely explicated from any single perspective, and I freely admit that there are additional insights to be gained from analytical perspectives that I have not used here. The techniques I use in the pages that follow simply seem to be the most effective ones to communicate the particular point I wish to make. The Schenkerian-style reductive graphs here are not intended to imply the same set of assumptions about the nature and function of the music as would be the case for Mozart, Beethoven or Brahms; I use them to identify important pitches and voice-leading. Likewise, my use of Roman numerals is intended to imply discernable tonal-functional progressions in the music, even in cases where the surface of the music is more chromatic. Perhaps the most useful model for me in this regard has been the 1965 dissertation of Laurence Berman,  The Evolution of Tonal Thinking in the Works of Claude Debussy,  whose  ability to trace how Debussy's music works inside of (and occasionally against) the traditional tonal system has not received the attention it deserves. Richard Parks' 89  Debussy,  The Music of Claude  a pc-set approach that uses Allen Forte's set-class numbers, has been by turns  enlightening and frustrating. While his observations about the non-tonal aspects of Debussy's 90  pitch organization are insightful, Parks has a habit of asserting that Debussy used Forte's pc collections as part of his compositional practice, which he clearly could not have, since many of the more exotic collections that Parks sees as central to Debussy's oeuvre had not received theoretical description during the composer's lifetime. Now I do not contend that theory is or ought to be prescriptive, but it is difficult to image how Debussy could have conceived of  Laurence Berman, The Evolution of Tonal Thinking in the Works of Claude Debussy. Ph. D. diss., Harvard University, 1965. Richard Parks, The Music of Claude Debussy (New Haven, CT: Yale University Press, 1989)  9 0  44 many of these collections as collections, rather than as enrichments of more traditional scales and harmonies with which he was certainly familiar. To my mind, Debussy's compositional practice involves an underlying tonality that is distorted on the surface by various non-traditional scalar collections, most often the wholetone and octatonic collections. So a complete understanding of the music rests on seeing the underlying tonal syntax and the way that these other collections enrich and obfuscate that syntax. In Debussy's music, the teleological nature of tonality is counterbalanced by the more static and circular pc collections. Parks tends to treat the pc collections as independent of the tonal structure of the individual pieces, which makes it hard to see how the two are related. Models of music and text relationships in song over the past half-century have moved steadily towards a model of reading. Critics from Edward T. Cone to Lawrence Kramer have put forward and developed the notion that in a song, what the composer sets is not the poet's poem, but the composer's reading of the poem. Kramer's conception is more fully developed 91  from a literary perspective than Cone's, but both are eager to imaginatively reconstruct the compositional process as a model for how song should be understood. There is at the heart of this conception a notion of mimesis that makes me uncomfortable. Specifically, it is the assumption that the music of a particular song was composed entirely in response to the text being set, or to a reading of that text. Following this logic, the aesthetic value of a song should be judged by how well the composer matches, follows, or illustrates the text. For convenience, I will call this type of approach to song composition mimetic. Even if it were possible to prove that this mimetic model explains what a particular composer actually does in the compositional process, it would remain a poor basis for critical evaluation. No matter how carefully a composer attempted to follow the syntax, form, sound and meaning of a poem, there would See Edward T. Cone, The Composer's Voice (Berkeley and Los Angeles, 1974); Lawrence Kramer, Music and Poetry: The I9' century and after (Berkeley and Los Angeles: University of California Press, 1984). 91  h  45  necessarily be numerous elements of the text that found no particular expression in the music, and a similar number of purely musical elements that would not engage with the text. Kramer recognizes this, and is willing to concede that the process of setting a poem as song does a certain amount of violence to the text itself, regardless of how closely a composer tries to follow the text. He coins the term transmemberment to represent the often disharmonious way that poetry and music engage and read across one another in song. The music and the text of a song will each organize time in their own way, both as independent elements and also as the hybrid that is song. Kramer's transmemberment suggests that a more complete critical reaction to song would be not only to note those places where the music and text are in accord, but also places where the musical and textual logic are unrelated or even at cross purposes. Richard Kurth has taken this notion one step further, and argued for a view of music and text relationships in song that avoids the idea that the music represents the text. Instead, 93  he advances the idea that, in song, the music and the poem engage one another on a more abstract level. Rather than seeing the music as dependent on the poem for its meaning, Kurth argues that the music already has its own meaning, generated through its own intrinsic processes.  94  This musical meaning is not created "after" the poem (although the specific  composition may be), but instead is an integral part of music itself that exists prior to its actual instrumental or vocal expression. Much as language engages a system of meanings that significantly predate the composition of any particular poem, music also emerges from a tradition with its own grammar and syntax. In this way, he argues that both the music and the poem be conceived as examples of writing as Derrida uses the term: as a forms of "generalized  Kramer, Music and Poetry, 125. Richard Kurth, "Music and Poetry, A Wilderness of Doubles: Heine-Nietzsche - Schubert - Derrida" Century Music 21/1 (Summer 1997): 3-31. See especially pages 26-33. Ibid., 31.  9 2  9 3  9 4  I9 lh  46  writing" that are not transcriptions of prior acoustic events (speech in one case and instrumental sound in the other), but "any sort of inscription in any sort of graphic, sonic or conceptual medium." As such, the music and poetry of a song can be understood to write 95  over (or under) one another. Kurth coins the term counter-writing to describe the ways that music and poetry write across one another in song, the one constantly informing and reconstructing the other. My own approach to music and poetry in song borrows rather freely from Cone, Kramer and Kurth. It is also informed by Mallarme's notion of music and poetry as "the alternate faces of the Idea," a concept that will be treated more fully in Chapter One. Whatever a composer imagines the process to be, we should never lose sight of the fact that what the composer actually sets is not a reading of the poem: it is the text itself. By "text," I mean not only the words of the poem, although these are fundamental and often to my mind strangely overlooked in many approaches to song, but also the network of meanings and relationships that accrue to the words through their engagement with literary tradition. Thus, on one level, song is characterized by counter-writing, in that both the music and the poem are self-contained cognitive acts whose individual meanings enter into a supplementary relationship. However, song is also, simultaneously, the presentation of a particular utterance of the poem: an opportunity for language and music to be heard in tandem. The way that the composer crafts this utterance may privilege some aspects of the poem, be they semantic, rhythmic or phonetic, without effacing those aspects that it does not treat as significant. Debussy's Mallarme' settings often seem crafted with this idea in mind. They preserve, to a degree uncommon in song, the rhythms and pacing of the poems, in order to ensure that the collateral damage done to the poem in the process of its setting is minimal.  Kurth, "Music and Poetry," 27n.  47  The text itself is a single unit of significance, which endures its setting in song much as a diamond endures its setting in a ring. Likewise, the music with which the composer sets the 96  poem is a single unit of significance that suffers the presence of the text. What really matters in song is the relationship between these two independent units; the way that the music and the text read (and write) across and through one another. Therefore, we cannot restrict ourselves to considering only those moments where they seem to be in accord, lest we make of the music simply a "bad pun" of the text (or vice versa). We must strive, instead, to understand the 97  complex dialectic of music and poetry that makes up song, tracing both congruity and disjunction, whether intentional or incidental. Both the poet's voice and the composer's voice deserve recognition, even if they speak at cross purposes or with mutual indifference. That this often leaves a textural gesture without a musical counterpart or vice versa does not concern me at all. Therefore, in the analyses that follow, I am not fundamentally concerned with showing that Debussy understood all of the finer points of Mallarme's linguistic theories as they are formulated in his critical essays and revealed in his poetry. Likewise, I do not argue that Debussy's musical language owes a particular debt to Mallarme, a position that I find impossible to prove in any significant way. What interests me, instead, is the way that music and text work in these pieces, producing both moments of accord and discord in the process. In the case of the songs, it is Debussy's extraordinary willingness to allow Mallarme's text to emerge from the setting with its rhythm and syntax intact that preserves the poet's voice in the song. This allows Debussy's musical response, and particularly the piano parts, to convey the  Significance is used here in the sense defined by Michael Riffaterre, Semiotics of Poetry (Bloomington, IN: Indiana University Press, 1978), 8. Cone's approach is particularly problematic here. He claims that, in a song, the composer steals the poet's voice, climbs inside the poet's words and changes their meanings to reflect the composer's understanding of the text. A l l traces of this violence must then be ignored by the critic. The only meanings that count are the ones that the composer has recognized and chosen to respond to in the music. 9 6  9 7  48 composer's voice most directly. Debussy's typical sensitivity to syntactic and semantic issues in these poems also foregrounds those moments in which the music contradicts the text, writes over it and transmembers it. In the case of the Prelude, the fact that the words of Mallarme's poem are neither explicitly heard in the piece or imagined in the listener's mind during its performance allows Debussy to focus more directly on the aesthetic ideas of Mallarme, particularly those concerning music, the theater, and the way that Music expresses and figures e  the Idea.  49 CHAPTER ONE T H E M U S I C O F L A N G U A G E : S O U N D A N D S E N S E IN M A L L A R M E  INTRODUCTION  From his earliest published poems in the 1860s through his final critical essays in the 1890s, Stephane Mallarme wrote constantly about music and its importance to his oeuvre. Music and musicians figure prominently in several early poems, and Mallarme frequently relied on musical metaphors to describe his poetic technique in letters to close friends. In the late 1860s and throughout the 1870s, music played an important role in Mallarme's evolving conception of language, which necessarily impacts both his later poetry and his prose. From 1885 through his death in 1898, music became more important as a subject in his critical writings. In part, this was a response to the popularity of Wagner infin-de-siecleParis, but it was also a reaction against several poetic movements that attempted to imitate musical effects in verse. Although Mallarme's aesthetic is based in large part on the idea of music, his 1  conception of this art is intensely personal, having little to do with conventional notions of music. His mature writings advocate disregarding the traditional sonorities of instruments and voices in favour of a mental music that he found in language and explored through poetry. In the early part of his career (1865-1869), Mallarme's desire to write a new and original kind of poetry coincided with a series of spiritual and epistemological crises. As letters from the period testify, he became intensely interested in the relationship between the sounds of words and their meaning, which in turn inspired him to consider the formation and evolution of language. Mallarme undertook a series of informal linguistic and aesthetic studies that were crucial to the development of his mature poetic style. The idea of music plays an important role ' For example, the instrumentiste poetry of Rene Ghil and his circle, who arbitrarily assigned both colours and musical timbres to French phonemes in the hope of imitating music and painting.  50 in the crisis period, one that has received relatively little critical attention. From letters and the notes that remain from abandoned literary projects, as well as the poetry produced during that time, the role of music in Mallarme's early works can be studied. Likewise his notion of a musical language based on the coincidence of phonetic and semantic relationships between words can be reconstructed. The conclusions that can be drawn from this evidence are confirmed by Les mots anglais, a pedagogical philology of the English language that applies many of these ideas about language to a foreign idiom. During the crisis, Mallarme came to believe that poetry—which is the highest expression of the musical possibilities of language—was uniquely suited to explore the mysteries of the universe. He saw a model for a new epistemology based on the way language functions: on the way that sound and sense interact to produce and refine meaning, rather than the strictly denotative system of contemporary language in which words have their meanings through a tacit conventional agreement among its users. Furthermore, he came to see that the 2  universe is also constructed this way, as a series of relationships between things that does not necessarily point beyond itself. The reality of objects in the world (or words in language) was important insofar as it allowed him to examine the relationships they generated. Between the end of the crisis period in 1870 and his sudden rise to prominence in 1884, Mallarme published very little poetry. Instead, he dedicated himself to exploring the consequences of the conclusions he had drawn about language and the universe. During this time, Mallarme began serious work on Le Livre, his vision of a great work that would summarize the history and development of the universe through poetry. This work was never finished. In the notes that remain for Le Livre, music takes on a slightly different role than in his earlier poetry. Here, music becomes a symbol for abstract patterns of thought that identify Mallarme often ridiculed the commercial use of language, where the meaning of words was as arbitrarily defined and conventional as the value assigned to money. See Mallarme, "Crise de vers," OC II, 212. 2  51 relationships between apparently disparate things. Le Livre is dedicated to the notion that there is a unity underlying the universe that is hidden from common perception. The relationships suggested by language, be they semantic, phonetic, graphic or etymological trace relationships between the actual objects and ideas that language evokes. The music of language that Mallarme had discovered in the 1860s now became a model for the essential relationships that make up all of existence. Mallarme's notes for Le Livre provide evidence for the expanded role that music now plays in his thought. In 1885, Mallarme rose to prominence and his output of both poems and critical prose increased significantly. Many of the essays deal explicitly with music. Some are dedicated to the subject of Wagner, and to outlining the essential differences between the Gesamtkunstwerk and Mallarme's own ideal synthesis of the arts in poetry. Others are more abstract and theoretical, describing the contributions of music to contemporary writing and to the Great Work. In these essays, Mallarme espouses his belief that poetry itself is music par excellence. Actual instrumental sound, as heard at a concert for example, was for him merely a sketch of some unwritten poem awaiting its fullest expression through language. Mallarme had often attended the Lamoureux concerts from the middle of the 1880s, and could often be seen there furiously scribbling in a notebook as the music was playing, as if trying to transcribe in language the poem latent underneath the orchestra's sonic surface. In letters to fellow poets, he often spoke of repatriating the rhythms and techniques of music to language, which was their ultimate source, and of the challenge that Wagner's music dramas presented to a new generation of French poets. The late poetry puts into practice many of the ideas contained in the late essays, which are in essence simply further developments of Mallarme's ideas about musical language and verse in general.  52 This chapter traces the evolution of Mallarme's musical thinking as it relates to the sound-sense relationship in language. It draws out specific ideas about the musical nature of language from his correspondence and critical writings. The most important of these are letters that describe the spiritual and linguistic crisis that Mallarme suffered in the 1860s, notes from an aborted dissertation on Words that he planned in 1869, and the book  Les mots anglais,  which applies many of these ideas to the English language. I then show how these ideas are transformed into poetic practice in selected poems that are all obsessed with a relatively small group of letters and phonemes shared by a large number of words in the text. With this background established, I turn to Mallarme's concept of musical thought as revealed in letters and the notes for  Le Livre.  Finally, I show how Mallarme's ideas about music are further  developed in some of the late critical essays and in his later poetry.  M u s i c IN M A L L A R M E ' S E A R L Y W R I T I N G S  Mallarme's first significant writing about music is found in the critical essay "Heresies artistiques: L'art pour tous," which dates from 1862. Here, Mallarme regrets the easy access 3  that the public has to literary works of art. He claims that a degree of mystery is necessary to protect artworks from the profaning gaze of the general population. His ideal example is musical notation, a macabre procession of "severe, chaste and unknown signs." These signs, 4  meaningful only to people with specialized musical training, protect music from the intrusion of ignorant readers who might otherwise stumble through masterpieces and feel that they understand the composer's message. Poetry has no such defense from idle curiosity, since it is printed in the normal characters of the alphabet. This gives rise to the practice of teaching poetry in the schools, and 3 4  Mallarme, "Heresies artistiques: L'art pour tous" OC II, 360-64. Ibid., 360: "signes severes, chastes, inconnus."  53 promotes the notion, abhorrent to Mallarme, that an "educated" person—one who would not give a second thought to his inability to understand a page of musical score—feels compelled to offer judgment on poetry. To remedy this situation, Mallarme dreamed of inventing "an 5  immaculate language,—some hieratic forms whose arid study blinds the profane and encourages the predestined person who has patience." This new language would be inherently 6  musical in its function; it would speak only to those initiated in the secrets of poetry, while keeping the casual onlooker at arm's length from the work. As the polemical work of a twenty-year-old poet, "Heresies artistiques" provides a sketch of certain aspects of Mallarme's mature aesthetic. Certainly, there are derivative elements in the essay, particularly Mallarme's valorization of Wagner, which at this point in time is largely borrowed from Baudelaire's essay on Tannhauser, published the previous year. But from this point forward in Mallarme's career, the connection between music and poetry and the need to make poetic language musical become recurring themes in both his poetry and his critical essays. The search for a new language, one that could answer the ambition of "Heresies artistiques," became an obsession for Mallarme in the following years. Letters from the period between 1864 and 1866 indicate that this search was transformed into a study of words, and a careful consideration of the relationship between sound and sense in them. In a long letter to Henri Cazalis, Mallarme speaks of the hours of research spent on each word for the poem "L'Azur" so that "the first word, which contains the first idea, not only helps to create the  It does not occur to Mallarme to mention that that although the casual onlooker may not be able to read musical notation, anyone who can hear has access to music as sound, and virtually anybody who hears a piece of music feels entitled to have an opinion of it, exactly the same situation that he decries in poetry. Mallarme, "Heresies artistiques," OC II, 361: "une langue immaculee, —des formes hieratiques dont l'etude aride aveugle le profane et aiguillonne le patient fatal." Mallarme did not attend the 1861 premiere of Tannhauser, and performances of Wagner were rare until the 1880s, when Wagner excerpts were played at the concerts spirituels led by Charles Lamoureux. There is no evidence that Mallarme ever made a pilgrimage to Bayreuth. 5  6  7  54  general effect of the poem, but also serves to prepare the last word. The effect produced, without a dissonance, without a fioritura, no matter how appealing, for that distracts—this is what I'm seeking." Mallarme continues: For those who, like yourself, seek in a poem something other than the music of the verses, they will find a real drama there. And it has been a terrible difficulty to combine, in proper harmony, the dramatic element, hostile to the idea of pure and subjective Poetry, with the serenity and the calmness of the lines essential for Beauty. 9  Like many of Mallarme's poems, the ostensible subject of "L'Azur" is the poet's inability to write poetry. Here, the poet is oppressed by the very Azur that inspires him.  10  Mallarme, Corr. I, 103-5. Letter to Cazalis, January 1864: "le premier mot, qui revet la premiere idee, outre qu'il tend par lui-meme a l'effet general du poeme, sert encore a preparer le dernier. Ceffet produit, sans une dissonance, sans une fioriture, meme adorable, qui distraie - voila ce que je cherche." Translation after Lloyd, 5  Selected Letters, 26.  Ibid., 104: "[P]our ceux qui, comme toi, cherchent dans un poeme autre chose que la musique du vers, il y a la un vrai drame. Et c'a ete une terrible difficulte de combiner, dans une juste harmonie, l'element dramatique, hostile a l'idee de Poesie pure et subjective, avec la serenite et le calme de [sic] lignes necessaires a la Beaute." Bertrand Marchal has given a compelling account of Mallarme's fascination with the Azur. See La religion de Mallarme: Poesie, mythologie et religion (Paris: Jose Corti, 1988). 9  10  55  L'Azur"  Blue Sky  De l'eternel azur la sereine ironie Accable, belle indolemment comme les fleurs Le poete impuissant qui maudit son genie A travers un desert sterile de Douleurs.  The serene contradiction of the eternal Azure Overwhelms, beautifully indolent like the flowers The impotent poet who curses his genius Across a sterile desert of Sorrows.  Fuyant, les yeux fermes, je le sens qui regarde Avec l'intensite d'un remords atterrant, Mon ame vide. Ou fuir? Et quelle nuit hagarde Jeter, lambeaux, jeter sur ce mepris navrant?  Fleeing, eyes closed, I feel it watching With the intensity of a dismal remorse, My empty soul. Where to flee? And what haggard night To cast, shreds, to cast on this unfortunate scorn?  Brouillards, montez! versez vos cendres monotones Avec de longs haillons de brume dans les cieux Que noiera le marais livide des automnes Et batissez un grand plafond silencieux!  Mists, arise! pour out your colorless cinders With long rags of mist in the skies Which will drown the livid marsh of autumns And build a great silent ceiling!  Et toi, sors des etangs leth6ens et ramasse En t'en venant la vase et les pales roseaux Cher Ennui, pour boucher d'une main jamais lasse Les grands trous bleus que font mechamment les oiseaux.  And you, come out from Lethean pools and gather In your coming the slime and the pale reeds Dear Boredom, to block with a tireless hand The great blue holes that the birds spitefully dig.  Encor! que sans repit les tristes cheminees Fument, et que de suie une errante prison Eteigne dans I'horreur de ses noires trainees Le soleil se mourant jaunatre a l'horizon!  Even though without respite the sad chimneys Smoke, and though the soot, a wandering prison Chokes in the horror of its black streaks The sun dying yellow on the horizon!  —Le Ciel est mort. —Vers toi, j'accours! donne, 6 matiere L'oubli de l'ldeal cruel et du Peche A ce martyr qui vient partager la litiere Ou le betail heureux des hommes est couche.  —The Sky is dead. —Towards you I run! give, O matter Forgetfulness of the cruel Ideal and of Sin To this martyr who comes to share the stable Where the happy herd of man is laid down.  Car j'y veux, puisque enfin ma cervelle, videe Comme le pot de fard gisant au pied du mur N'a plus 1'art d'attifer la sanglotante idee Lugubrement bailler vers un trepas obscur ...  For there I want, because finally my brain, empty Like ajar of makeup lying up against the wall No longer has the ability to dress up the sobbing idea To yawn lugubriously towards an obscure death ...  En vain! L'Azur triomphe, et je l'entends qui chante Dans les cloches. Mon ame, il se fait voix pour plus Nous faire peur avec sa victoire mechante, Et du metal vivant sort en bleus angelus!  In vain! The Azure triumphs, and 1 hear it singing In the bells. M y soul, it takes voice to make us More afraid with its spiteful victory, And from the living metal emerges an angelus in blue!  II roule par la brume, ancien et traverse Ta native agonie ainsi qu'un glaive sur Oil fuir dans la revoke inutile et perverse? Je suis hante. L'Azur! L'Azur! L'Azur! L'Azur!  It rolls through the mist, ancient and across Your native agony like a sure sword; Where to flee in the useless and perverse revolt? / am haunted. L'Azur! L'Azur! L'Azur! L'Azur!  Mallarme's desire to have the "music of the verses" exist in a proper harmony with the "real drama" of the poem is worked out in its prosody. Specifically, Mallarme's end-rhymes 11  Mallarme, OC I, 14-15.  56  enact the semantic message of the poem, in a technique that will come to characterize his mature style. If we take Mallarme at his word, then the word "Azur"—the "first word" of the poem, serves both to prepare the last line of the poem phonetically and to focus the reader's attention on the treatment of this word in particular. In this light, it is not surprising that Mallarme uses "Azur" as an end-rhyme in the final quatrain of the poem. What is surprising is the relative poverty of the rime suffisante "sur / L'Azur", which shares only two phonemes: lyl 12  and / R / .  13  Most of the rhymes here, as with Mallarme's poetry in general, are rimes riches.  For example: regarde / hagarde  cieux / silencieux  chante / mechante  traverse / perverse  This observation becomes more meaningful in light of the fact that Mallarme had several other words at his disposal, ones that would have produced a rime riche with Azur and could have easily been worked into the poetic fabric (of the options, "brisure" and "croisure" are especially appealing). However, each of these words would have required Mallarme to use two different letters (s and z) to produce the Izl phoneme, a process that he generally avoids elsewhere in the poem. In fact, if one insists on maintaining the spelling of the phonemes, then Azur becomes a rhyme-less word. Read this way, the "serene contradiction" of "Azur" is not merely that the Azur-as-Ideal is unattainable metaphysically, but poetically or prosodically as well. Not content to abandon the struggle to find a rhyme for "azur," Mallarme simply switches from end rhyme to assonance, opening up a whole new set of possibilities. The  Throughout this chapter, IPA phonetic characters are used to represent the sound of specific phonemes. For those unfamiliar with these characters, see Appendix B for a complete listing of French phonemes and a guide to pronunciation. ' A rime suffisante shares 2 phonemes, a consonant and a vowel with the same voicing. A rime riche has at least three shared phonemes. 12  K  57  proliferation and concentration of /a/ phonemes in certain lines of the poem suggests that Mallarme was consciously exploiting this sound: A ce martyr qui vient partager la litiere N'a plus Fart d'attifer la sanglotante idee II roule par la brume, ancien et traverse Ta native agonie ainsi qu'un glaive sur The tight control of phonemes that these lines exhibit, and the way that they connect with the key word Azur is likely what Mallarme had in mind when he wrote of the "effect produced" and the "music" of the lines. Yet the haunting four-fold reiteration of "L'Azur" that closes the poem suggests that the word continues to torment the poet long after the poem itself reaches its conclusion. Mallarme's struggle with language reaches a kind of critical mass with "Herodiade," which Mallarme began in 1864. The desire to focus on the "effect produced" in "L'Azur" had 14  revealed the inadequacy of everyday words, as they were commonly used, to notate the kinds of effects Mallarme had in mind. He returns therefore to the idea set forth in "Heresies artistiques" of inventing a new poetic language. However, now the concern is not primarily with protecting poetry from the casual glance. Instead, Mallarme writes of inventing a language that must o f necessity burst forth from a very new poetics, which I could define in these few words: paint, not the object, but the effect it produces. Therefore, the lines in such a D o e m must not be composed of words, but of intentions, and all the words must fade before the sensation. 5  The language that Mallarme wanted to produce was one in which there would be an accord between the intentions of the poem and the words actually used to fashion it. He wanted to eliminate, as much as possible, the tendency of individual words to stick out of the poetic texture, to live lives independent of their contextual situation. Increasingly, his thoughts turned Despite working on the poem for more than 30 years, "Herodiade" remained unfinished at Mallarme's death. Mallarme, Corr. I, 137. Letter to Cazalis, late October 1864: "une langue qui doit necessairement jaillir d'une poetique tres nouvelle, que je pourrais definir en ces deux mots: Peindre, non la chose, mais I'effet qu'elle produit. Le vers ne doit done pas, la, se composer de mots, mais d'intentions, et toutes les paroles s'effacer devant la sensation." Translation after Lloyd, Selected Letters, 39. 14  1 5  58 to music as a model for this kind of language and the considerable use of musical metaphor to describe his struggles in the composition of "Herodiade" testify to this. I have, moreover, there [in "Herodiade"] found a singular and intimate means of painting and notating very fleeting impressions. Add, for more terror, that all of these impressions follow one another as in a symphony, and that I often spend entire days asking myself whether this impression can accompany that one, what is their relationship and their effect. 16  However, Mallarme found work on "Herodiade" exceptionally difficult. In a letter to Cazalis, he complains that "I was throwing myself like a desperate maniac on the elusive overture of my poem ["Herodiade"] that sings within me, but which I can not write down." The 17  following day, in a letter to Theodor Aubanel, Mallarme makes a similar complaint, adding "I need the most silent solitude of the soul, and an unknown forgetfulness, to hear singing inside myself certain mysterious notes."  It seems that Mallarme struggled not with the idea of the  poem per se, but in deciding on how that idea would be best expressed in verse. The musical metaphors here are telling, for Mallarme's search for the correct words with which to express his ideas are centered on the notion of phonetic harmony and balance. This is easy to see in an 1865 letter to Eugene Lefebure on "Herodiade": "The most beautiful page of my work will be that which contains only the divine word Herodiade. What inspiration I've had I owe to this name, and I believe that if my heroine had been called Salome, I would have invented this dark word, red as an open pomegranate,  HerodiadeZ  9  Whatever inspiration Mallarme may have drawn from the story of Herodiade, he also drew significant poetic inspiration from the phonemes of word itself. As Eric Gamier notes, this Mallarme, Corr. I, 161. Letter to Cazalis, January 1865: "J'ai, du reste, la [in Herodiade] trouve une facon intime et singuliere de peindre et de noter les impressions tres fugitives. Ajoute, pour plus de terreur, que toutes ces impressions se suivent comme dans une symphonie, et que je suis souvent des journees entieres, a me demander si celle-ci peut accompagner celle-la, quelle est leur parente et leur effet." Mallarme, Corr. I, 179-80. Letter to Cazalis, 5 December 1865: "je me jetais en maniaque desespere sur une insaisissable ouverture de mon poeme [Herodiade] qui chante en moi, mais que je ne puis noter." Mallarme, Corr. I, 181. Letter to Aubanel, 6 December 1865: "j'ai besoin de la plus silencieuse solitude de 1'ame, et d'un oubli inconnu, pour entendre chanter en moi certaines notes mysterieuses." Mallarme, Corr. I, 154. Letter to Lefebure, 18 February, 1865: "La plus belle page de mon oeuvre sera celle qui ne contiendra que ce nom divin Herodiade. Le peu d'inspiration que j ' a i eu, je le dois a ce nom, et je crois que si mon heroine s'etait appelee Salome, j'eusse invente ce mot sombre, et rouge comme une grenade ouverte, Herodiade." Translation from Lloyd, Selected Letters, 47. lfa  17  18  19  59 technique—the "singular and intimate means of notating fleeting impressions" quoted above— can be clearly seen in line 116 of the poem, where the phonemes of the word Herodiade (/eRodiade/) are reordered and scattered throughout the line:  20  Herodiade au clair regard de diamant /er//o//dio//d3/ lol /er/ /ds//dia/ This phonetic game is more than just an isolated display of poetic virtuosity. It also responds to the semantic surroundings of the text. In a section that celebrates Herodiade's purity and that, by extension, celebrates the purity of the poetry with which Mallarme describes it, the image of Herodiade gazing into a mirror finds its echo in the prosodic relationship between the word and its own phonetic surroundings: Et ta soeur solitaire, 6 ma soeur eternelle Mon reve montera vers toi: telle Rare limpidite d'un cceur qui le songea, Je me crois seule en ma monotone patrie Et tout, autour de moi, vit dans 1'idolatrie D'un miroir qui reflete en son calme dormant Herodiade au clair regard de diamant... O charme dernier, oui! je le sens, je suis seule.  And your solitary sister, O my eternal sister, M y dream shall rise toward you: such Rare clarity of a heart which dreams of it, I believe I am alone in my monotonous country, And everything, around me, lives in the idolatry Of a mirror that reflects in its sleeping calm Herodiade with the clear gaze of a diamond ... O final charm, yes! I feel it, 1 am alone.  Several similar examples can be adduced. It is significant, however, that Mallarme never allows the complete phonetic makeup of the word Herodiade to appear in any other line. Instead, the lines typically lack one phoneme or fragment the individual syllables of her name. In this way, Herodiade is constantly suggested by the play of phonemes on the sounding surface of the poem while at the same time highlighting the singular appearance of the entire phonemic set shown above:  Eric Gamier, La Musicalite de tout (Chamber/: Editions Comp'acte, 2007), 39.  60  Une Aurore a, plumage heraldique, choisi ("Ouverture," line 4) HI  lal / e R / / d i /  /R//R//Q/  Par le diamant pur de quelque etoile, mais ("Ouverture," line 15) / R / /dia/  lnJIdal  Id  De ses espoirs, pour voir les diamants elus ("Ouverture," line 95) Id/Id  /R/  / R / /R// //dia/ e  Id  Empruntent leur clarte melodieuse, et vous ("Scene," line 92) IKI /aRe/ /e//o//di/ Id Et tout, autour de moi, vit dans l'idolatrie ("Scene," line 114) Id  lol iRlIdd  Midi  /id/ /ml  The same phonetic techniques that Mallarme developed for "Herodiade" can be seen at work in other poems explicitly connected with music. One such poem is "Brise marine."  T H E S O N G OF T H E SAILORS: M A L L A R M E ' S "BRISE MARINE'  Brise marine  21  La chair est triste, helas! et j ' a i lu tous les livres. Fuir! la-bas fuir! Je sens que des oiseaux sont ivres D'etre parmi I'ecume inconnue et les cieux! Rien, ni les vieux jardins refletes par les yeux Ne retiendra ce cceur qui dans La mer se trempe 6 nuits! ni la clarte deserte de ma lampe Sur le vide papier que la blancheur defend, Et ni la jeune femme allaitant son enfant. Je partirai! Steamer balancant ta mature, Leve l'ancre pour une exotique nature! Un Ennui, desole par les cruels espoirs, Croit encore a l'adieu supreme des mouchoirs! Et, peut-etre, les mats, invitant les orages Sont-ils de ceux qu'un vent penche sur les naufrages Perdus, sans mats, sans mats, ni fertiles ilots ... Mais, 6 mon cceur, entends le chant des matelots!  21  Mallarme, O C I , 15.  Sea Breeze The flesh is sad, alas! and I have read all the books. To flee! to flee down there! I feel that the birds are drunk To be amid the unknown foam and the skies! Nothing, not the old gardens reflected in my eyes Can restrain this heart that soaks in the sea 0 nights! nor the deserted light of my lamp On the empty page that the whiteness defends, And not the young woman nursing her child. 1 will depart! Steamer balancing your masts, Weigh anchor for an exotic nature! A Boredom, left desolate by cruel hopes, Still believes in the supreme farewell of handkerchiefs! And, perhaps, the masts, inviting storms Are those that a wind tips toward the shipwrecks Lost, without masts, without masts, nor fertile isles ... But, 0 my heart, hear the song of the sailors!  61  Composed in 1865 and published the following year, "Brise marine" is, like "L'Azur," a poem motivated by the inability to write a poem. The poet sits at his desk, facing again the empty page whose very whiteness offers it a defense from the profaning stroke of his pen. The weight of literary tradition stifles his creativity, as each tentative beginning seems to be merely an imitation of another writer and another writing ("j'ai lu tous les livres"). He dreams of a voyage to uncharted regions where the poet, drunk to be in unfamiliar lands, can create something truly original. The obstacles of his quotidian life—his young family, familiar surroundings, even his notorious writer's block—will not stop him this time. He is determined to leave all this behind in search for new and exotic lands. Yet his naive enthusiasm is tempered by a deep-seated fear that this voyage, like others before, will end in disaster, his poor poetic vessel no match for the storms that it will face. Nevertheless, he seems resolved to set forth, spurred on by the song of the sailors. Interpreted this way, "Brise marine" picks up on a common Mallarmean theme of failure: failure to write the great poems that his imagination has sketched (especially "Herodiade"). Poems like this one are then seen as studies for larger works, like  Le Livre.  Mallarme himself often refers to his shorter poems in this way. However, this interpretation 22  is somewhat superficial, since it focuses on the denotative meanings of the words and the analogies they generate. It takes no account of the actual poetic processes that Mallarme uses in its creation and is therefore only one half of the counterpoint that is essential to understanding the poem. The voyage that Mallarme dreams of is a voyage of language, and through language. It enacts its own success even as it speaks of its own peril.  Mallarme, Corr. II, 299. Letter to Verlaine, 16 November 1885.  62 In " B r i s e m a r i n e , " the poet i s c o m p e l l e d b y the s o n g o fthe sailors. W h a t , exactly, is  t h i s s o n g ? I n o r d e r t o a n s w e r t h e q u e s t i o n , w e m u s t l o o k at M a l l a r m e ' s p r o s o d y . A q u i c k  at M a l l a r m e ' s r h y m e s h e r e r e v e a l s a p r e p o n d e r a n c e o f  livres / ivres  /IVR/  /ivR/  mature / nature  /atyR/  (riche)  trempe / lampe  /waR/  (riche)  / w a R / (riche)  orages / naufrages  (suffisante)  defend / enfant /fa/  /atyR/'  espoirs / mouchoirs  /le sjo/ /lezjo/  /ap/  rimes riches:  (riche)  les c i e u x / les y e u x  /ap/  look  {suffisante)  /fa/  Of the three instances of rimes  suffisantes  /Ra3/  /Ra3/  ilots ... / matelots /lo/ l\ol  (riche)  (suffisante)  in this poem, the first two coincide with the poem's  description of writer's block and the challenges posed by his family commitments. One may 23  interpret these  rimes suffisantes  as enactments of the poet's struggles to write the kind of  poetry to which he aspires. The third rime 24  suffisante  in the final couplet is harder to explain  as an enactment of the semantic message of the text. The song of the sailors is supposed to be the main source of the poet's inspiration, the reason why he will risk the storms on his rickety boat. Yet the relative poverty of this rhyme calls the whole endeavor into question: why risk everything for such an uninspiring rhyme?  The visual "dissonance" in orthography reinforces the trouble here - Mallarme's letters from this period often complain that his daughter Genvieve (born in 1864) cried often and loudly. I do not mean to imply that rimes riches are somehow more desirable than rimes siffusantes in general, but merely that in this poem the rimes suffisantes are made to coincide with the poem's description of the things that stand in the poet's way.  63 The answer, I think, is found in the exhortation to listen to the song of the sailors, to hear the way that it winds through the text and transforms the final word of the poem. Stated simply, Mallarme's poetic technique encourages the reader to hear: "mats-ilots" /matilo/ as a homonym lurking behind the actual word : matelots /mat(3)lo/  in order to provide the rime riche that the listener has come to expect (the six previous lines all end with rimes riches). This fuses at a phonetic level three words that already share both an orthographic and a semantic connection: "mats", "ilots" and "matelots." This is the "music of language" that Mallarme will identify later on in Les mots anglais, and an important part of his poetic technique. The surrounding lines are densely packed with visual and acoustic echoes of this "new" word: Et, peut-etre, les mats, invitant les orages Sont-ils de ceux qu'un vent penche sur les naufrages Perdus, sans mats, sans mats, ni fertiles ilots ... Mais, 6 mon coeur, entends le chant des matelots! The three-fold repetition of "mats," the two instances of "lie" (island) hidden in "sont-ils" and "fertiles," and the two Ms in the last line demonstrate a conscious effort by Mallarme to reinforce the individual letters of this group of words. It may also be that the striking number 26  Graham Robb has detailed similar cases of mispronunciation in Mallarme's poetry. See Unlocking Mallarme, 52-54. Compare this with the original 1865 version of the poem, whose last four lines read: Et serais-tu de ceux, steamer, dans les orages And would you be one of those, steamer, in the storms Que le Destin charmant reserve a des naufrages That charming Fate has destined for shipwrecks Perdus, sans mats ni planche, a l'abri des Hots... Lost, without masts or plank, in the shelter of isles... Mais, 6 mon cceur, entends le chant des matelots! But, O my heart, hear the song of the sailors! Z i  2 6  64  of circumflexes (6) are significant, providing a kind of visual connection between "mats" and 97  "ilots."  This makes the alternative pronunciation of "matelots" seem motivated by the text  itself—not just a chance fragment of language or a poetic conceit, but a reconstruction of language itself through poetry. Mallarme's linguistic games thus invite an alternative reading of, and an alternative reading strategy for, the poem. In addition to the usual dialogue of rhythm, meaning and rhyme, here the reader is encouraged to scan the text for common phonemes and letters that call out to one another independently of their usual disposition in words. Certain key letters and sounds emerge from various words and invite the reader to read them as integral to the poem's significance. Much later, Mallarme describes the process thus: "Words, spontaneously, are exalted by several recognized facets [letters] the more rare or valid for the mind, the centre of vibratory suspense; that perceives them independently of the ordinary sequence."  29  Mallarme himself compares this kind of reading strategy to the reading of 30  constellations.  Reading a constellation is essentially an act of imagination that creates  meaning through the recognition of celestial patterns, connecting certain stars and not others The conflation of "mats" and "ilots" in "matelots" is evident here, but the surrounding lines do relatively little to reinforce the phonetic similarity of these words. See Mallarme, OC I, 122. Cohn suggests that the circumflex represents "the effect of waves" in Mallarme's poetry. Though 1 am not convinced of the universality of this claim, it may be relevant to this passage. See Cohn, Toward the Poems, 279. Mallarme discusses this practice in "Crise de vers," an essay concerned with the increasing "musicalization" of poetry in France: "Le vers qui de plusieurs vocables refait un mot total, neuf, etranger a la langue et comme incantatoire, acheve cet isolement de la parole : niant, d'un trait souverain, le hasard demeure aux termes malgre I'artifice de leur retrempe alternee en le sens et la sonorite, et vous cause cette surprise de n'avoir ouT jamais tel fragment ordinaire d'elocution, en meme temps que la reminiscence de l'objet nomme baigne dans une neuve atmosphere." "The verse from which several words remakes a total word, new, foreign to the language and as if incantory, achieves this isolation of the word: denying, by a sovereign stroke, the chance remaining in terms despite the artifice of their alternative re-tempering in sense and sonority, and causes you this surprise of having never heard such an ordinary fragment of speech, at the same time that the memory of the named object bathes in a new atmosphere." See Mallarme, "Crise de vers," OC 11, 213. Mallarme, "Le Mystere dans les lettres," OC II, 233: "Les mots, d'eux-memes, s'exaltent a mainte facette reconnue la plus rare ou valant pour l'esprit, centre de suspens vibratoire; qui les percoit independamment de la suite ordinaire." See, for example, a letter to Cazalis in which Mallarme contrasts his own poetic practice with that of mutual friend Emmanuel des Essarts, who "prend une poignee d'etoiles... et les laisser se former au hasard en constellations imprevues" ("takes a handful of stars ... and allows them to form themselves at random into unexpected constellations"). Mallarme, Corr. I, 104. Letter to Cazalis, January 1864. 2 7  2 8  2 9  3 0  65  because of their shared participation in a particular image. A poetic constellation applies the same kind of thinking to language, except here the stars are replaced by letters. In order to read a poetic constellation, the reader must be willing to read both in a traditional, linear, syntactic manner, and in a new, essentially geometric way, drawing connections between words with shared phonemes (or actual letters in the present case) that are significant in themselves, in addition to their grammatical or syntactic function. These constellations are engraved across the surface of the poem, adding to and refining the total significance of the text. This play of phonemes and letters in Mallarme's text is, literally, the song of the sailors ("le chant des matelots"), a song which winds through the final lines of "Brise marine" as a positive counterpoint to its semantic message of longing and impending failure. On a boat made of words and sounds, the poet sails through the storms of chance, navigating by the constellations of shared sounds and letters that reveal the music of language.  M A L L A R M E ' S SILENT M U S E : "SAINTE C E C I L E "  Around the same time as "Brise marine," perhaps as a way to work out some of the finer points of the new language with which he was struggling, perhaps as a pleasant diversion from his frustrating work on "Herodiade," Mallarme penned a short poem that he described as "melodic and created above all with music in mind."  1  31  Mallarme, Corr. I, 181. Letter to Aubanel, 6 December 1865: "melodique et fait surtout en vue de lamusique."  66  Sainte Cecile jouant sur l'aile d'un cherubin  32  (Chanson et image anciennes)  Saint Cecilia playing on the wing of an angel (Ancient song and image)  A la fenetre recelant Le santal vieux qui se dedore De la Viole etincelant Jadis parmi flute et mandore,  In the window concealing The old sandalwood that loses its gilt Of the Viol sparkling Once amid flute and lute,  Est une Sainte recelant • Le livre vieux qui se deplie Du Magnificat ruisselant Jadis a vepree et complie,  Is a Saint concealing The old book that unfolds The Magnificat flowing Once from vespers and compline:  Sainte a vitrage d'ostensoir Pour clore la harpe par l'ange Offerte avec son vol du soir A la delicate phalange  Saint at the window of a monstrance To close the harp offered by An angel with its evening flight To the delicate tip  Du doigt que, sans le vieux santal N i le vieux livre, elle balance Sur le plumage instrumental, Musicienne du silence!  Of a finger which, without the old sandalwood Nor the old book, she balances On the instrumental plumage, Musician of silence.  The work stands out in Mallarme's oeuvre in many ways. Formally, it is a departure from Mallarme's self-admitted "mania" for sonnets in his early lyric works. It is also written in octosyllabics, the first instance of this meter in his mature poetry. It is, moreover, the first of 33  the lyric poems to have a musician as its subject. The musical elements of the poem go far beyond the musical subject of the work, however, right down into the fabric of the words themselves. The subject of the poem is a stained glass window containing a portrayal of the patron saint of music, adorned with the customary viol and hymnal and surrounded by other winged angels. In the setting sun, the window reveals the Saint's finger resting lightly, or so it appears, on the wing of one of the surrounding cherubim. Mallarme sees a kind of harp in the angel's  Mallarme, OC I, 114. Like many nineteenth-century poets, Mallarme organized many of his youthful poems according to form and meter, suggesting that he saw these elements as significant in themselves, and not wholly external to the poem's content. 3 2  3 3  67  wing, which the Saint seems to prefer to the faded viol and the ancient missal. It is not 34  difficult, given the frequency with which Mallarme equates the poetic act with wings and flights, to understand that by choosing the poetic wing rather than the traditional musical instrument, Mallarme is challenging music by suggesting that her patron saint actually prefers his silent, poetic music. Guy Michaud has noted that this poem is arranged like a diptych in which the two panels of the window (one showing the Saint with the instruments, the other with her playing the wing/harp, represented by the two halves of the poem) are superimposed onto one 35  another.  Thus, the images of the Saint as musician and as poet are simultaneously present in  what Michaud calls a polyvalent image. Michaud also suggests a sort of mimetic connection between certain phonemes in the poem and its key images: the labial consonants represent the flight of the angel, the dentals and combined liquids imitate the sound of the viol and harpsichord, the sibilants and " i " sounds of the final stanza evoking silence.  37  Michaud's polyvalent image also works as an apt description of rhyme itself: the superimposed image of two words whose individual meanings are simultaneously present in an oscillation of similarity and difference. In light of this, notice that the masculine rhymes of the poem's first half recelant / etincelant / recelant / ruisselant /sslaV /sslaV /sslaV /ssla/ find their own image in the rhymes on /la/of the second half of the poem:  Cohn, Toward the Poems, 91-95. Cohn points to several other places in Mallarme's writings where angel's wings are compared with harps. Michaud, Mallarme, 42. A polyvalent image is similar to an analogy or metaphor in that it involves two ostensibly unrelated images brought into a comparison by the poet. Yet in a metaphor or analogy, the flow of meaning is hierarchical - the subsidiary image is designed to elucidate the meaning of the primary one. In a polyvalent image, the distinction between primary and subsidiary images is blurred, and the images reflect meaning onto one another. Thus, the flow of meaning is circular rather than linear. See Michaud, Mallarme, 42. Michaud, Mallarme, 42. j 4  3 5  3 0  3 7  68 l'ange / phalange /la/ /la/  balance / silence /las/ /las/  These Isl and /Id/ phonemes are also present, though reordered, in "santal" and "instrumental", thus uniting each of the poem's four stanzas. The play of phonemes in "Sainte Cecile" is not restricted to end-rhyme alone, but spreads out in constellations across the sonorous surface of the poem. Contrary to Michaud, who argues for an essentially mimetic impulse, the phonetic concentrations here are less concerned with the imitation of physical objects or musical instruments per se than with a sort of phonetic free association in which the common phonemic elements of words seem to insist upon analogies or relationships that transcend representation and suggest more distant, elemental relationships. The concentration of sibilants in the poem, primarily Isl but also, in a subsidiary way If/, is striking. Far from being an evocation of silence, as Michaud claims, the Isl sounds flow through the poem, from the first phoneme of "fenetre" through the closing "silence" as a kind of sibilant tonic whose presence is felt in the background of the poem regardless of the actual 38  sounding phoneme.  The Isl phonemes connect some of the central images of the poem in a  constellation: Sainte - Cecile - chanson - anciennes - santal - ostensoir - soir - instrumental Musicienne - silence Most of these words are directly related to music, either of their own semantic accord or by association in this particular poem. The others have more distant relationships that can be reconstructed with a relatively small dose of poetic imagination. For example, "anciennes" (ancient), is an approximate homophone of "antienne" (antiphon or hymn), which is expressly connected with both the "chanson" of the subtitle and the Magnificat that appears later in the This opens up the possibility of the /s/ phoneme being prolonged by other, subsidiary phonemes in a manner analogous to a tonal prolongation in music. 3 8  69 poem.  39  A n d it is the angel's "vol du soir" that produces the harp that the Saint appears to be  playing and the "ostensoir" (monstrance-like appearance) o f this image in general. These phonemic connections, like Michaud's polyvalent images, bring words into circular relationships. It becomes increasingly difficult to know whether the  semantic  connection between words is primary, supplemented by these phonetic links, or whether the basis for the poetic comparison begins with the phonetic similarities. Ultimately, it matters little whether  Sainte Cecile is  the  musicienne du silence  because she actually abandons music  in the semantic sense (i.e. she prefers poetic music to that produced by instruments, resulting in actual silence) or whether the play o f Is/ phonemes casts itself ("s'y lance") across the surface of the poem in a shower o f sibilant sound. Mallarme plays similar phonemic games with the NI, producing the constellation: vieux - Viole - livre - vepree - vitrage - vol Unlike the earlier Isl constellation, where a semantic link to music was either present or easily discernible, this one lacks a definite semantic center. The most important o f these images is the vol, a common Mallarmean metaphor for the poetic act. in the  Viole, much  stanza.  and  This vol is contained (as an anagram)  as the poem itself is interwoven with music. But the other  seem to exist in an uncertain relationship to the  vieux  40  santal vieux  vol.  NI  phonemes  Particularly striking is the reversal o f livre  from the poem's first half as  vieux livre and vieux santal  in its final  This kind of adjectival reversal may be motivated by Mallarme's study of English—  where the adjective normally precedes the noun—but also picks up on the chiastic reversal of the letters V and I in  vieux  and  livre.  41  This graphic mirroring creates another kind o f  Roger Pearson chronicles the role of homophony in Mallarme's poetic practice in Unfolding Mallarme. He gives an identical reading of "ancienne" as "antienne" in conjunction with the Ouverture to "Herodiade" on page 93, but does not invoke this homophone in his discussion of "Sainte" (pages 62-64.) See, for example, "Les fenetres," where Mallarme uses the image of featherless wings to represent his inability to write his ideal poetry, and "Le jour"—later "Don du poeme,"—which also compares a poem to a bird. This is similar to the phonemic reversal "Ivre, il vit" in "Les fenetres." See Mallarme, OC I, 9-10. 3 9  4 0  41  70 polyvalent image, superimposing the two words as if folded around the blank space that i  42  separates them. This act of folding is suggested by the graphic shape of the V itself, which also resembles a wing. Now, it would be foolish indeed to suggest that all Vs are wings in 43  Mallarme's poetry, but in this particular case, the treatment of the Vs in "Sainte Cecile" leaves little doubt that this kind of relationship—in which the individual words "fade before the intention" of the line—is consciously exploited here. It is also possible that V, itself a graphic mirror image, is the motivation behind the phonemic reversal: a kind of sonorous reflection that represents the visual one. To some extent, then, Mallarme's choice of words in "Saint Cecile" was motivated by his desire to invent a language that enacts the semantic content of his poem. The proliferation 44  of Isl phonemes could then be understood as flowing from Cecile (the dedicatee of the poem) out into the phonetic texture of the poem, reinforcing the semantic and analogical relationships from which it is constructed. However, it is also possible that Mallarme's primary motivation here was an obsession with the relationship between sound and sense in language itself. This is not a simple reversal, where the phonetic sounds of the poem become primary and its sense secondary. Instead, the two are held in a supplementary relationship, where each is the image of the other, a polyvalent image made of both semantic and phonetic images. This makes their relationship here particularly unstable, and gives the poem its uniquely radiant quality. The unusual syntax of "Sainte Cecile" plays an important role in allowing these elements to become more prominent. The poem is composed of a single sentence whose syntax is full of digressive clauses that threaten its intelligibility. This both unites the two halves of the Jacques Derrida has made much of the notion of a structural blank in Mallarme's poetry. See "The Double Session," 201. ' Cohn, Toward the Poems, 114. See Genette's brilliant analysis of Mallarme's approach to mimetic language in Mimologics, 201-218.  4 2  4  4 4  71  poem and opens up a kind of semantic gap between the stanzas. In these semantic gaps, where the reader casts around to make sense of the syntax and grammar, the common phonemes and letters between individual words become more significant. The syntax of the first two stanzas of the poem could be radically simplified thus: A la fenetre est une Sainte. From this semantic kernel, Mallarme's sentence blossoms with a series of digressions that transform this mundane sentence into something far more poetic. Mallarme begins with a parallel series of descriptive digressions, both introduced with recelant. In the following examples, the parentheses have been added for clarification: A la fenetre (recelant le santal vieux de la Viole) est une Sainte (recelant le livre vieux) Here, the digressions are set up so as to apply equally to the window or the Saint: the syntax and lack of punctuation in the poem itself makes both interpretations possible and productive. Mallarme then further disrupts the syntax by drawing the reader's focus further from either the window or the Saint: A la fenetre (recelant le santal vieux (qui se dedore) de la Viole (etincelant jadis parmi flute et mandore)) est une Sainte (recelant le livre vieux (qui se deplie du Magnificat ruisselant jadis a vepree et complie) This kind of syntactical reordering is present to some degree in virtually all poetry, but already here Mallarme has taken it to an unusual degree. By withholding the main verb ("est") until the beginning of the second stanza, Mallarme encourages the reader to cast around the poem in order to create meaning. Since there are multiple possible syntactic paths to chart in the poem, the non-semantic elements of the poem—form, meter, phonetic content—become increasingly significant.  72  Mallarme's concern with the sound-sense relationship in his poetry continued to grow in the months following "Saint Cecile" and "Brise marine." He writes of the "study of the sounds and colors of words, the music and painting through which your thought must pass, however beautiful it may be, if it is to become poetry." What he desires, above all, is that the 45  individual words in a line of verse work together in support of the underlying idea. This is the basis of a criticism that he levels at Francois Coppee. After praising some of Coppee's poems for their unity and purity, he continues: Chance doesn't enter into any line and that is the great thing. Several of us have achieved this, and I believe that when lines are so perfectly delimited, what we should aim for above all, in a poem, is that the words—which are already sufficiently individual not to receive external impressions—reflect upon each other to the point of appearing not to have their own colour anymore, but to be merely transitions within an entire scale. Although there is no space between them, and although they touch each other wonderfully, I feel that sometimes your words live a little too much as individuals, like the stones in a mosaic of jewels. 46  Mallarme's use of a musical metaphor (the scale) to describe the kind of poetry he values is neither fortuitous nor rare. For him—at this time—a musical poem is one in which the formal elements of the poem are in harmony with its semantic elements. However, his faith in the ability of language to live up to these expectations was about to be shattered.  Music  A N D T H E CRISIS  Between 1866 and 1868, Mallarme underwent a series of aesthetic and spiritual crises. In a series of letters to close friends, he writes of his struggle to come to grips with the consequences of his growing atheism, both in philosophical and poetic terms. The religious aspects of Mallarme's crisis have been explored in detail by Bertrand Marchal, and cannot Mallarme, Corr. \, 168. Letter to Cazalis, July 1865. Translation from Lloyd, Selected Letters, 52-53. Mallarme, Corr. I, 234. Letter to Coppee, 5 December 1866: "Le hasard n'entame pas un vers, c'est la grande chose. Nous avons, plusieurs, atteint cela, et je crois que, les lignes si parfaitement delimitees, ce a quoi nous devons viser surtout est que, dans le poeme, les mots - qui deja sont assez eux pour ne plus recevoir d'impression du dehors - se refletent les uns sur les autres jusqu 'a paraltre ne plus avoir leur couleur propre, mais n 'etre que les transitions d'une gamme. Sans qu'il y ait d'espace entre eux, et quoiqu'ils se touchent a merveille, je crois que quelquefois vos mots vivent un peu trop de leur propre vie comme les pierreries d'une mosai'que de joyaux." Translation from Lloyd, Selected Letters, 69. Emphasis mine. 4 5  4 6  73  come under consideration here. The critical feature of the crisis for the present study is that 47  these realizations are triggered through an exploration of poetic language rather than abstract philosophical contemplation. Frustrated with his work on "Herodiade," Mallarme describes the first crisis thus: Unfortunately, while mining the lines [of "Herodiade"] to this extent, I have encountered two abysses that fill me with despair. One is the Void, which I have reached without any knowledge of Buddhism, and I am still too distraught to be able to believe even in my poetry and get back to work, that this crushing thought has made me abandon. Yes, / know, we are merely empty forms of matter—but truly sublime for having invented God and our soul. So sublime [...] that 1 want to gaze upon matter, becoming conscious of it, and, nevertheless, launching itself madly into Dream, which it knows does not exist, singing of the Soul and all the divine impressions of that kind that have collected within us from the earliest times and proclaiming, in the face of the Void which is truth, these glorious lies! 48  For him, "the Void" is the realization that there is nothing beyond this world—not only an afterlife for the soul, but no higher meaning or purpose. Mallarme sees that "reality" is in fact just a construction of the mind that seeks connections between the actual physical world and metaphysical worlds of its own creation. This realization is the natural outgrowth of his earlier recognition of the conventional nature of language, fueled by his nai've yet passionate belief that poetry is (or at least ought to be) capable of explaining the universe. Mallarme's conception of reality as a construction of the mind threatens to sterilize the poet in him. For if there is nothing "out there" that is real, no referent to which his words can point, then how can poetry (or language for that matter) be anything but nonsensical sound on which we arbitrarily project meaning? Yet in this deficiency of language, Mallarme sees productive possibilities as well. For matter can launch itself into Verse as easily as into Dream  Bertrand Marchal has documented this period in detail in La religion de Mallarme. Mallarme, Corr. I, 207. Letter to Cazalis, 28? April 1866: "Malheureusement, en creusant le vers a ce point, j ' a i rencontre deux abimes, qui me desesperent. L'un est le Neant, auquel je suis arrive sans connaitre le bouddhisme, et je suis encore trop desole pour pouvoir croire meme a ma poesie et me remettre au travail, que cette pensee ecrasante m'a fait abandonner. Oui,y'e le sais, nous ne sommes que de vaines formes de la matiere mais bien sublimes pour avoir invente Dieu et notre ame. Si sublimes, [...] que je veux me donner ce spectacle de la matiere, ayant conscience d'elle, et, cependant, s'elancant forcenement [sic] dans le R i v e qu'elle sait n'etre pas, chantant 1'Ame et toutes les divines impressions pareilles qui se sont amassees en nous depuis les premiers ages, et proclamant, devant le Rien qui est la verite, ces glorieux mensonges !" Translation from Lloyd, Selected Letters, 60. Emphasis mine. 4 /  4 8  74 (and the homophonic palindrome "Vers / Reve" is often productive in Mallarme, as we shall see in "Soupir," treated in Chapter Four), and the poet can in fact try to reconstitute the missing order of a Godless world. The process is straightforward, but not easy. First, the poet must always be aware of the constructed nature of language and the universe. Second, the poet must explore his own history, and the history of his race, by exploring his language, collecting the "divine impressions" that the centuries have deposited there "from the earliest times." Lastly, the poet must produce poems that reconstruct language, and in so doing, the universe, as if there were a reality beyond this one at which they could point. This is the "glorious lie" to which Mallarme refers. These ideas are developed in a series of letters to Theodor Aubanel from the summer of 1866. There, Mallarme claims to have laid the foundations for his life's work, a solitary pursuit that demanded the exclusion of all outside influence. His project was to reconstruct language 49  from the inside out, to penetrate into the collective linguistic consciousness of the French mind and uncover an essential unity between sound and sense in language that had been lost with the proliferation of languages and the passage of time. He thought that it would take him twenty years of patient work to complete this task. As it turned out, the work would never be 50  finished, and it is impossible to know, from the drafts that remain, how close he came to realizing this dream.  M u s i c A N D EPISTEMOLOGY  Mallarme's meditations on language led him to more abstract reflections on the nature of thought and knowledge itself. The epistemological aspect of this crisis was probably influenced by his exposure to the ideas of Hegel, who saw an essential unity behind the 4 9 5 0  Mallarme, Corr. I, 222. Letter to Aubanel, 16 July 1866. Translation from Lloyd, Selected Letters, 66. Ibid.  75 contradiction and negation of the exterior world. Hegelian terms like Idea, Being, Time, 51  Synthesis, etc. become regular features of Mallarme's correspondence, and these same letters often dwell on the essential interconnectedness of all things. In his later critical works, as we shall see, Mallarme will consistently describe this interconnectedness as musical. For now, let us begin with some of the basic patterns of Mallarme's thought. Shortly after his discovery of the Void—the realization that "reality" is purely subjective and internal, rather than something to be discovered externally—Mallarme found Beauty, a new aesthetic that he claimed would form the cornerstone of his work. This Beauty 32  was nothing more or less than the desire to explore the various relationships inherent in language, and through them to better understand the universe. His new poetics became a model for a new epistemology that conceives of the world as a series of relationships—at various distances and through various media—rather than as an objective reality. Mallarme believed that understanding the fundamental relationships of the universe was best served by exploring these relationships through language. The ultimate goal of this was to uncover a divining thread that would bind the various illusions of the physical world together. Mallarme called this goal the Idea, and described it as a series of interconnected lines between objects or notions. I wanted to tell you [in a previous letter] simply that I had just cast the plan for my entire Work, after having found the key to myself—keystone, or center, if you will, so as not to mix metaphors—, the center of myself where I dwell like a sacred spider, on the principal threads that have already been cast from my mind, and with the assistance of these threads, 1 will weave at the meeting places marvelous lace, which 1 divine, and which exists already in the core of Beauty. 53  It is unclear exactly how Mallarme might have gained his exposure to Hegel's ideas. Two of his close friends— Eugene Lefebure and Auguste Villiers de 1'Isle-Adam—were self-admitted Hegelians. See Janine D. Langan, Hegel and Mallarme (Lanham, M D : University Press of America, 1986) for a comprehensive treatment of this question. Mallarme, Corr, I, 220. Letter to Cazalis, 13 July 1866. ' Mallarme, Corr. 1, 224. Letter to Aubanel, 28 July 1866: "J'ai voulu te dire simplement [in a previous letter] que je venais jeter le plan de mon CEuvre entier, apres avoir trouve la clef de moi-meme—clef de voute, ou centre, si tu veux, pour ne pas nous brouiller de metaphores—, centre de moi-meme, ou je me tiens comme une araignee sacree, sur les principaux fils deja sortis de mon esprit, et a l'aide desquels je tisserai aux points de rencontre de merveilleuses dentelles, que je devine, et qui existent deja dans le sein de la Beaute." Translation after Lloyd, Selected Letters, 67. 51  52  5  76  Here, Mallarme lays out his poetic project with remarkable concision. The "principal threads" of his thought are basic patterns of similarity and difference, while the individual poems are woven at the intersection of these lines, at the places where similarities (of image, metaphor, phoneme, etc.) converge momentarily. Therefore, each individual poem is a synecdoche of the whole, a fragment of the Idea that demonstrates one of its facets. The Idea as a totality is inaccessible (though Mallarme attempted this in Le Livre), but is rather a product of poetry itself in its totality. By exploring individual intersections of thought in particular poems, Mallarme hopes to sketch the essential outlines of the Idea. The thought process that Mallarme outlines above is not coldly intellectual, though its consequences are extremely challenging. Mallarme is seeking a total integration in the mind of all things simultaneously. Poetry provides a model for this in the way in which it superimposes words, images and ideas, but even poetry is not thought in its most abstract form. Instead, thought must be musical if it is to encompass the totality of the Idea. Mallarme describes his process thus: I think that to be truly a man, to be nature capable of thought, one must think with one's entire body, which creates a full, harmonious thought, like those violin strings vibrating directly with their hollow wooden box. As thoughts are produced by the brain alone [...], they now appear to me like airs played on the high part of the E string whose sound is not reinforced in the box,—that pass through and disappear without creating themselves, without leaving a trace of themselves. 54  The musical metaphor of the violin here is no accident. Mallarme uses the same image in the prose-poem "Le demon de l'analogie" to represent the syllable nul and the way that its sound and signification haunt the narrator, to which we shall turn now our attention. At its core,  Mallarme, Corr. I, 244. Letter to Eugene Lefebure, 27 May 1867: "Je crois que pour etre bien l'homme, la nature se pensant, il faut penser de tout son corps—ce qui donne une pensee pleine et a l'unisson comme ces cordes du violon vibrant immediatement avec sa boTte de bois creux. Les pensees partant seul du cerveau [...] me font maintenant l'effet d'airs joues sur la partie aigiie de la chanterelle dont le son ne reconforte pas dans la boite — qui passent et s'en vont sans se creer, sans laisser de trace d'elles." Translation from Lloyd, Selected Letters, 79-80. 5 4  77 Mallarme demands a polyvalent sensibility to sound, sense, image and analogy, a network of relationships that will come to dominate his oeuvre despite the fact that they are radically unstable, and ultimately illustrate the process of thought.  55  MALLARME'S DEMONS  Mallarme's struggle to uncover the relationship between individual letters and the meaning of the words that they compose, and how this play of language explains objective reality, are central themes of his prose poem "Le demon de l'analogie." Mallarme slowly 56  realized that poems made of words that "exist merely as transitions within a scale" are difficult to make because of the essentially unstable relationship between sound and meaning in language. The epigram that he chooses for the poem summarizes this struggle aptly: Des paroles inconnues chanterent-elles sur vos levres, lambeaux maudits d'une phrase absurde ? (Have unknown words ever sung on your lips, cursed fragments of an absurd phrase?)  In the body of the poem, the poet is haunted by a sound which at first seems to be a violin melody. This melody is transformed into a voice uttering the phrase "la penuliteme est morte" in such a fashion that the violin's tone can still be discerned in the sound nul. This meaningless phrase becomes an obsession, as the poet turns it over and over in his mouth and his mind, trying vainly to understand either the semantic significance of this phrase or its transformation into the violin melody. As he wanders the streets of Paris, a sense of panic Jean-Pierre Richard's thematic criticism of Mallarme, and his ability to trace images from poem to poem and analogy to analogy must stand as a landmark in Mallarme criticism. Yet, as both Derrida and Cohn have shown, Mallarme's capacity for analogy and hidden relationships exposes any thematic criticism as a tautology, and the whole signifying chain of images collapses. In the final analysis, there is no way to fix the boundaries of a group of images. What is left, then, in a poem is not a meaning per se but the illustration of a thought process, a way of seeing the relationship between things. See Richard, L 'Univers imaginaire de Mallarme. ; Derrida, "The Double Session," 174-285. ; Cohn, "Appendix: on Derrida's Mimique" in Mallarme's Divagations, 167-71. In the 1945 edition of Mallarme's CEuvres Completes, Henri Mondor and Georges Jean-Aubry place the date for this prose poem in 1864 (it was only published in 1874). However, Bertrand Marchal suggests a date of 1867, which would put this work right in the middle of Mallarme's spiritual crisis. Given the subject of the poem and its flirtation with insanity which closely parallels Mallarme's descriptions of the crisis, Marchal's date seems credible. The text and my translation appear in Appendix C. See Mallarme, OC I, 1335-6. 55  5 6  78  overtakes him as he struggles to convince himself that the nul can be contained by the semantic meaning of "penultieme." Seeing that this is not possible, he attempts to bury his concern by singing the words as if they had no meaning at all ("le secret espoir de l'ensevelir en l'amplification de la psalmodie"). But this too fails to satisfy the feeling that this "cursed fragment" of language is meaningful. Engrossed in thought, the poet hardly notices his surroundings until he catches sight of his reflection in a shop window. With a sense of horror, he sees his reflected hand passing over the strings of a violin in a luthier's shop window, and realizes that this was the source of the original sound that haunted him, and the voice pronouncing the absurd phrase was his own. 57  Certainly, the tone here owes something to Poe's The Imp of the Perverse. However, the central issue for us here is the poet's obsession with the syllable nul and the way in which it fluctuates between pure music (as a representation of the violin tone) and pure representation (either as an independent term signifying "nothing"—which can hardly be coincidental—or as a component of "penultieme"). The sound emerges from and returns to music, as if enacting Mallarme's earlier proclamation in the letter to Coppee about words which must fade before the sensation or intention of the line. The search for the "meaning" of nul motivates the action of the poem, but is also a consequence of it. The poem's speaker can not decide whether his obsession with nul comes from his struggle to write poetry—finding an apt rhyme for "penultieme" for instance—or from the violin hanging in the luthier's window that he "plays" with the reflection of his hand at the end of the poem. Priority and anteriority are blurred, as are cause and effect. The drama of the poem itself is wrapped up in its obsession with nothing, both literally and figuratively. In this way, nul is at once a negative space (the Void) signifying lack and absence, and a positive space, both through its presence - through - absence (in Cohn, Mallarme's Prose Poems, 2-19. As Cohn points out, the title was probably influenced by Baudelaire's translation of Poe's The Imp of the Perverse, which Baudelaire rendered as Le Demon de la Perversite. 5 7  79 Derrida's sense of the term) and in that nul is a real sound, even if it points to nothing beyond itself. So the poem is at once inherently self-referential and at the same time reaches out to a potentially infinite number of metaphorical or analogical readings. It is a masterpiece of undecidability, and in that way approaches Mallarme's basic condition of Music as a series of mental relationships that stretch over a yawning chasm of meaninglessness. In the year after "Le demon" was written, and at the height of the crisis, Mallarme wrote the "Sonnet allegorique de lui-meme" in 1868. When he first sent it to Henri Cazalis, Mallarme claimed that it was extracted from a projected study on "the Word" on which he had CO  been working. Like "Le demon," the sonnet is created from obsession with a particular sound that may in fact be meaningless. Sonnet allegorique de lui-meme  Sonnet allegorical of itself  La nuit approbatrice allume les onyx De ses ongles au pur Crime lampadophore, Du Soir aboli par le vesperal Phoenix De qui la cendre n'a de cineraire amphore  The approving night lights the onyx O f her nails in the pure flame-bearing Crime, From Evening abolished by the vesperal Phoenix From whom the ash has no funeral amphora  Sur des consoles, en le noir Salon: nul ptyx, Insolite vaisseau d'inanite sonore, Car le MaTtre est alle puiser l'eau du Styx Avec tous ses objets dont le reve s'honore.  On the credenzas, in the black Room: no ptyx, Strange vessel of sonorous inanity, For the Master has gone to draw water from the Styx With all the objects with which the dream honors itself.  Et selon la croisee au nord vacante, un or Nefaste incite pour son beau cadre une rixe Faite d'un dieu que croit emporter une nixe  And according to the vacant northern casement, a gold Ill-fated incites for its beautiful frame a brawl Made by a god that believes he is carrying a nixie  En l'obscurcissement de la glace, Decor De 1'absence, sinon que sur la glace encor De scintillation le septuor se fixe.  In the obscurity of the mirror, Decor O f absence, except that in the mirror still The scintillation of the septet fixes itself.  This early version of the "Sonnet en yx" shows the same obsession with rhymes on "yx" as its more famous published version (1887). In fact, this early version contains seven rhymes on "yx" (compared with only six in the "Sonnet en yx"). The details of this sonnet are  Mallarme, Corr. I, 277. This study—elsewhere identified as a dissertation—was never in fact completed. The notes which remain will be examined later on in this chapter.  80  well-known to Mallarme enthusiasts: Mallarme set a task for himself to create a sonnet using the rare syllable "yx" as one of its rhymes. Having exhausted the French language of such rhymes, Mallarme invented the word "ptyx," believing that it existed in no language whatever. Thus, the poem seems to call forth this word as an incantation, a word created by 59  the text in order to complete itself. Mallarme claimed that the sonnet's meaning "is evoked by an internal mirage of the words themselves." Robert Greer Cohn has noted the double image created here, both by the 60  repetition of the syllable "yx" and of the palindromic shape of the letter x itself. The rhyme 61  on "yx" here is more than Mallarme's attempt at poetic virtuosity. It is also intimately related to his notion of the Void and to a growing poetics of absence. The poem is set in a vacant room on some abolished evening.  The room is wholly  defined by absence—the Master is out, the "ptyx" missing—except for the mirror, which reflects the lights of a distant constellation out of the window. Since this presence is merely a 63  reflection, or illusion, it serves only to confirm the emptiness of the room. From this point of view, the "sonorous inanity" of "ptyx" becomes the perfect phonemic symbol for the emptiness of the room itself: a word that is pure play of sound, devoid of semantic meaning.  It has since been identified as a possible derivative of a Greek word referring to a shell, but Mallarme's letters suggest that he was unaware of this at the time. Mallarme, Corr. I, 278. Letter to Cazalis, 18 July 1868. Cohn, Toward the Poems, 139. Cohn further connects this idea to a basic pattern in Mallarme's thought that he calls "tetrapolarity." I will return to this idea, and its musical basis, later on. In the letter to Cazalis, Mallarme offers his own explication: " A nocturnal window is open, its two shutters fastened; a room with no one inside, despite the stable appearance that the fastened shutters present, and in a night made of absence and questioning, without furniture, except for the plausible outline of vague credenzas, a frame, warlike and dying, of a mirror hung up at the back, with its reflection, stellar and incomprehensible, of [the constellation] Ursa Major, which links to heaven alone this lodging abandoned by the world." ("[U]ne fenetre nocturne ouverte, les deux volets attaches ; une chambre avec personne dedans, malgre l'air stable que presentent les volets attaches, et dans une nuit faite d'absence et d'interrogation, sans meubles, sinon l'ebauche plausible de vagues consoles, un cadre, belliqueux et agonisant, de miroir appendu au fond, avec sa reflexion, stellaire et incomprehensible, de la grande Ourse, qui relie au ciel seul ce logis abandonne du monde") Mallarme, Corr. I, 278. Letter to Cazalis, 18 July 1868. Mallarme identifies this constellation as Ursa Major, but given the reference to the "septuor," he may well have had in mind the Big Dipper (itself a fragment of the larger Bear). y )  6 0  61  6 2  6 3  81  However, Mallarme is not content to invoke an empty and meaningless space. By invoking the constellation glimpsed through the mirror, he suggests an alternative reading that allows the richness of rhyme to remunerate the semantic deficiency of "ptyx" in particular, and the sonnet in general. The seven rhymes on "yx" here can be aligned with the constellation glimpsed in the mirror, likely the seven stars of the Big Dipper. Like the celestial constellation, whose individual stars have no meaning, but onto which, as a whole, human imagination imposes a form (and, in the case of the constellations, an entire mythology), Mallarme's rhymes transform an otherwise meaningless sound into something meaningful. In this way, 64  "ptyx" becomes a significant nothing, a physical emptiness that alludes to a poetic fullness, and potential mastery. Before leaving the "Sonnet allegorique de lui-meme," we would be well advised to consider its other rhyme, on /OR/. If "ptyx" is the site of semantic poverty, and its rhymes serve to compensate for this lack, then "sonore" is its opposite, a site of semantic richness ("un or") that acts as a counterbalance. The homophonic rhyme "sonore / s'honore" is predicated on the notion that individual phonemes can carry multiple meanings, and that this too is the province of the poet. This homophonic richness finds a further echo in "[glajce encor" and "septuor." 65  The difficulty posed here is in uncovering the hidden relationships between these words and the others that would justify their appearance in the poem, if in fact Mallarme's claim that "chance does not enter into any line" is valid. It may be that the seven letters of the word "septuor" are called forth by the seven rhymes on /OR/, the seven on /iks/ and the seven star points of the constellation that fills the otherwise empty room. If so, then in the face of absence (of meaning, of furniture, of the Master) that is represented by the rhymes on /iks/, the  6 4  6 5  Given the myriad possible referrents given for ptyx in the critical literature, Mallarme seems to have succeded. For a comprehensive treatment of homophony in Mallarme's poetry, see Pearson, Unfolding Mallarme.  82 phonetic richness of the rhymes on /OR/ suggest its opposite, the poet's defense against a language that does not always behave as it should.  A N A B O R T E D DISSERTATION  The "projected study on the Word" from which Mallarme claimed the "Sonnet allegorique de lui-meme" was detached had evolved in Mallarme's mind into a doctoral thesis by 1869. Like so many of his projects, this was ultimately unrealized. However, a collection 66  of notes published posthumously as "Notes sur le langage" gives some insight into the scope of the project, and its conclusions, and provides perhaps the most profound insight into the linguistic dimension of Mallarme's spiritual crisis.  67  For Mallarme, the word is, or rather it ought to be, tied to its meaning in a stable and predictable way. Its component phonemes should reinforce the semantic meaning of the words they comprise, such that "certain sounds are equivalent to a certain idea, modified in such a fashion that a certain sound signifies this particular thing." The signification of the phoneme 68  ought not to be restricted to one language in particular, but should be a universal property of sound, so that "finding a neutral language, if there is one, that some sound means this thing absolutely, that this sound has a certain value." However, this is not the case in contemporary 69  language, since words with diverse phonetic properties can have very similar meanings (synonyms, for example), while words with virtually identical phonemes can have widely divergent meanings (homonyms, etc.). Yet in other cases, there are (or at least there seem to be) words in which sound and sense are in perfect accord, displaying both the inherent  Mallarme, Corr. I, 315. Letter to Cazalis, 31 December 1869: "Petude n'est que l'humble servante de l'Imagination ... et, si j ' e n ai la force, me laisse entrevoir les resultats possibles de la licence et du doctorat." In the new collected edition of Mallarme's works, these are collected under the title "Notes sur le langage," OC I, 501-512. Ibid., 510: "tels sons equivalent a telle idee, modifiee de telle et telle facon que tels son[s] signifient ceci." Ibid.: "trouvant une langue neutre, s'il en est une, que tel son par excellence signifie ceci, a telle valeur."  6 6  6 7  6 8  6 9  83  potential of language, and its history. In other words, the relationship between the meaning of a word and its phonetic structure has fallen victim to chance, its elements scattered throughout contemporary language. Minimizing the role of chance through poetic language will become Mallarme's chief poetic project. For the time being, he formulates it thus: The Word is a principle that develops through the negation of all principle, chance, like the Idea, and reconstitutes itself forming (like Thought, aroused by Anachronism), itself, Speech, with the assistance of Time which permits these scattered elements to meet up and to join together following the laws aroused through these diversions. 70  In Mallarme's conception of language, words are in constant danger of losing their semantic capacity and degenerating into a nonsensical collection of sounds. This was the essential drama of "Le demon de l'analogie." Yet words, recognizing their own deficiency, reach out to one another to form Speech, and in this way negate chance. In other words, the conversations through which people interact in their daily lives proceed as if all words have stable meanings and therefore as if certain sounds actually signify certain things. This negation of chance is not absolute, of course, but provisional. In the course of speech, the mind compensates for the failing of words and gives them a temporary cohesion. In this, semantics, syntax, tone, custom and grammar all play a role, but none of these is absolute. Therefore, language works in the same way as the universe, which—like language—proceeds from day to day as if there were some "reality" beyond itself, despite the fact that (for Mallarme), this is an illusion.  71  Mallarme, "Notes sur le langage," OC I, 505: "Le Verbe est un principe qui se developpe a travers la negation de tout principe, le hasard, comme I'idee, et se retrouve formant (comme elle la Pensee, suscitee par l'Anachronisme), lui, la Parole, a l'aide du Temps qui permet a ses elements epars de se retrouver et de se raccorder suivant ses lois suscitees par ces diversions." It is no accident that "Verbe" can mean not only "word" (mot), but also "word of God" (as in the English notion of "the Word"). Another aspect of Mallarme's crisis, which cannot come under consideration here, is Mallarme's struggle with his faith. For Mallarme, as for the apostle John, the beginning of the world begins with the word. There is some similarity between Mallarme's ideas and those of Hegel, Schopenhauer, Plato, etc. Perhaps the best philosophical consideration of Mallarme's aesthetic is Jean-Paul Sartre's Mallarme, or the Poet of Nothingness, trans. Ernest Sturm (University Park, PA: Pennsylvania State University Press, 1988). 7 0  71  84 Language reconstitutes some of the meaning which words, on their own, are lacking. Language, then, perpetuates a fiction, a deception, a "glorious lie" in which words which ought to be related by sound analogies are brought together through grammar and syntax instead. Language creates secondary analogies to supplement the inconsistent nature of the word, brings terms together as temporary mental acts which appear to draw their meaning from the lexicon but in fact are also relational: The spoken word, across Idea and Time which are "the identical negation of the essence" of Becoming becomes Language. Language is the development of the Word, its idea, in Being, Time became its mode: this across the phases of the Idea and of Time in Being, that is to say according to Life and the Mind. From whence come these two manifestations of Language, Speech and Writing, destined [...] to reunite both things through the analogies of sound—Writing by marking the gestures of the Idea manifesting itself through speech, and offering them their reflection, in order to perfect them, in the annals of successive effort of speech and of its filiations and to give it ancestry so that one day, their analogies noted, the Word would appear behind its means of language, returned to the physical and the physiological, as a principle, freed, suitable to Time and to the Idea. 72  So for Mallarme, language is the social development of the word over time which serves to reunite speech and writing through the "analogies of sound" while at the same time allowing the word itself to remain independent of its context, both at the heart of and beyond 73  language.  Therefore, the history of the world is wrapped up (for Mallarme) in the history of  language, the way that it has transformed over time. If one were able to strip away the layers of history from language, then words would show themselves to be analogous to the Idea. Their Mallarme, "Notes sur le langage," OC I, 505-6: "Le Verbe, a travers l'Idee et le Temps qui sont "la negation identique a l'essence" du Devenir devient le Langage. I Le Langage est le developpement du Verbe, son idee, dans l'Etre, le Temps devenu son mode: cela a travers les phases de l'Idee et du Temps en l'Etre, c'est-a-dire selon la Vie et I'Esprit. D'ou les deux manifestations du Langage, la Parole et l'Ecriture, destinees [...] a se reunir toutes deux choses par les analogies des sons - L'Ecriture en marquant les gestes de l'Idee se manifestant par la parole, et leur offrant leur reflexion, de facon a les parfaire, dans le annales de 1'effort successif de la parole et de sa filiation et a en donner la parente de facon a ce qu'un jour, leurs analogies constatees, le Verbe apparaisse derriere son moyen de langage, rendu a la physique et a la physiologie, comme un principe, degage, adequat au Temps et a l'Idee." 11  The division that Mallarme makes here between "Parole" and "Ecriture" is also important to Mallarme's thought but cannot be explored in detail here. For Mallarme, writing is more than a transcription of the phonetic qualities of sound (although it is that, too). The letters themselves, in their graphology, are like a transcription of the Idea, as a kind of written ballet of signs which temporarily fix meaning. See Mallarme's essays on Ballet in Crayonne au theatre, OC II, 170-8 for an expansion of this idea. It has been treated by a host of literary critics, including Derrida and Cohn. See Derrida, Of Grammalology, and Cohn, Mallarme's Divagations, 145-64. 7 j  85 connections would accurately and consistently point to the significant and essential relationships that explain the universe. At the same time, the Word is a physical and physiological entity. It exists as an object in sound (and in writing), and it is this material existence of the word that, freed from the semantic constraints of language, forms the poet's basic materials. Mallarme's formulation of language as a fiction is based on the paradox of the word. On one hand, the word is abstract, like thought itself. When language draws words together, it creates a fiction because it refuses to acknowledge the arbitrary nature of the individual word. This fiction is "the same process as the human mind," which itself draws things (objects, ideas, etc.) together momentarily, as if they had real existence. This is why language itself, and 74  literature in particular, is uniquely suited to explorations of abstract thought, since they both operate on the same foundation. However, and at the same time, words do have meanings, and these are not (or at least are not always) merely conventional: Finally words have several meanings, otherwise people would always understand one another—we will profit from this—and for their principal meaning, we seek what effect they would produce in us if pronounced by the interior voice of our mind, deposited through exposure to books of the past [...], if this effect distances itself from the one that it makes on us these days. 75  Mallarme has not entirely abandoned the notion that there is a bond between the sound of a word and its meaning. Rather, the plurality of meanings that accrues around a word is the result of the word's use in common speech and the passage of time which transforms and deforms an original meaning. If there is access to this original meaning, then it is through literary language ("la frequentation des livres du passe") that these relationships may be reconstituted.  Mallarme, "Notes sur le langage," OC I, 504: "le procede meme de l'esprit humain." Robert Greer Cohn has identified this formulation of Mallarmean fiction as the keystone of his poetic vision. See. Cohn, Toward the Poems, 299. Ibid., 508-9: "Enfin les mots ont plusieurs sens, sinon on s'entendrait toujours—nous en profiterons—et pour leur sens principal, nous cherchons quel effet ils nous produiraient prononces par la voix interieure de notre esprit, deposee par la frequentation des livres du passe [...] si cet effet s'eloigne de celui qu'il nous fait de nos jours." 7 5  86 In "Notes sur le langage," Mallarme claims that language is a manifestation of the Idea, connecting individual words as the Idea connects individual thoughts. Since language is merely a "play of relationships in the Mind", an understanding of language is itself an approximation of the idea. Mallarme continues by claiming that writing is a kind of transcription of the Idea, 76  and not merely a phonetic transcription of language, since the very shape of the letters themselves is significant.  77  For Mallarme, the Idea is both eternal and ephemeral. On one hand, the Idea as the essential interconnectedness of all things is a kind of universal truth, a universal constant. However, the idea as the individual act of making connections between specific thoughts is merely an enactment of the Idea, "a temporary act of the mind responding to the need of a 78  notion." Language models this duality in its capacity for analogy and its propensity to rhyme. Both of these make connections which are essentially temporary—they may be undone or refocused in another poem—and are suggestive of more stable and universal relationships between things.  79  This leads Mallarme to the paradoxical conclusion that "[t]he moment of the Notion of an object is therefore the moment of the reflection of sound purely present in itself or its OA  present purity." The idea-as-temporary-act is a mental representation of the Idea, a flash of the pure thing-in-itself which is triggered by its reflection in sound (rhyme in its various forms). This is, in summary form, Mallarme's epistemology.  Mallarme, "Notes sur le langage," OC I, 505. Ibid., 506. Ibid., 507: "un acte momentane de Fesprit repondant au besoin de notion." Mallarme's continual use of certain words and images - aile, vierge, ecume, grimoire, etc. is related to this position. Mallarme, "Notes sur le langage," OC I, 509: "Le moment de la Notion d'un objet est done le moment de la reflexion de son present pur en lui-meme ou sa purete presente." 7 7  7 8  7 9  8 0  87 M U S I C AFTER THE CRISIS  Many of the ideas which Mallarme sketched out for his aborted doctoral dissertation surface again in the 1870s in Les mots anglais, a pedagogical work designed to help French students to gain an understanding of the English language more quickly. In it, Mallarme 81  again, and in more systematic fashion, lays out his views on language and on the relationship between the sound of words and their meaning. In Les mots anglais, Mallarme takes the position that modern languages are merely "a transformation, corrupt or elegant, of earlier speech."  82  Every contemporary idiom, therefore,  can be traced back to a common source, a single generative language from which all others developed over time. This generative source, this common origin no longer exists in the world, but fragments of it can still be discovered in contemporary language.  Mallarme's stated task  in Les mots anglais is to uncover these fragments as they exist in English in order to help students to recognize the semantic relationships between words that share common letters and phonemes. However, in his constant digressions, a much more ambitious aim becomes clear: to reconstruct this generative language by a double application of intellect and imagination. This project is related to his desire, expressed in "Notes sur le langage," to discover a "neutral language" where the semantic values of phonemes are both constant and evident. Mallarme contrasts this with contemporary language in which historical layers of language  The practical value of Les mots anglais is certainly debatable. Mallarme dismissed the work (along with his Dieiix antiques) in the autobiography which he provided for Paul Verlaine's Les hommes d'aujourd'hui {OC I 789), calling them "des besognes propres [...] dont il sied de ne pas parler" ("my own labours [...] about which it is not fitting to speak"). Cohn questions their explanitory value for Mallarme's poetry, since the conclusions that he draws about the significance of English phonemes do not transfer readily to French. See Cohn, Toward the Poems, 265-80. However, my interest here is not in the pedagogical or truth value of the work, but in what it inevitably reveals about how Mallarme thinks about language. In this, I follow Genette, whose commentary on Les mots anglais places Mallarme in the context of mimetic theories of language. See Genette, Mimologics, 2 0 1 219. 81  8 2 83  Mallarme, Les mots anglais, OC II, 949: "une transformation, corrompue ou elegante, de parlers anterieurs." Ibid., 969.  88  (whether literary or spoken) have created a vast repertoire or words, while at the same time fracturing the sound-sense relationship that would bring it meaning. Important themes from the Notes sur le langage resurface here in a more polished form. Mallarme identifies words as a "jumble of phonemes" whose present arrangement in the columns of a dictionary is not merely the product of chance. Instead, the dictionary presents 84  layers of linguistic history, as words develop themselves according to discernable principles. New, however, is the notion that this process may be uncovered through the careful, inductive study of a single language (rather than an etymological exploration of many or  languages).  Such a study would reveal the development of language as "[s]o many acts, 86  complex and long forgotten, gently beginning again, for you alone, attentive to their history." Also new is the conjecture that acts of understanding (in language and more generally in the 87  world) can be achieved these days by "seizing a few relationships between many things." This posits an underlying unity that sits behind an apparent diversity, which is the strategy by which Mallarme sees connections between words themselves. A large portion of Les mots anglais is devoted to grouping the English lexicon. Mallarme seeks out words with a common set of letters that also share a semantic meaning, as a means of reconstructing the original signification of these letters from inside the language. For the most part, Mallarme assumes that common letters imply also common phonemes, but his examples show that this need not always be the case. In Les mots anglais, Mallarme gives preference to the consonants of words, which he likens to a skeletal structure (where the vowels and diphthongs act as a kind of flesh to be stripped away), since the vowels of many  Mallarme, Les mots anglais, OC 11, 948: "un pareil fouillis de vocables." Genette, Mimologics, 202. Mallarme, Les mots anglais, OC II, 948: "Tant d'actes, complexes et bien oublies, recommencant avec docilite, pour vous seul, attentif a leur histoire." Ibid.: "dans l'ere presente on n'apprend un peu qu'a force de comprendre, ou en saisissant quelques relations entre beaucoup de choses." In the 1880s and 90s, and with remarkable consistency, Mallarme equates this process of finding underlying relationships with music. 8 4  85  8 6  8 7  89  English words change depending on case and tense. Mallarme begins by noting several words whose common phoneme(s) and letters seem to reinforce their semantic connection: What more charming work, for example, and done even to compensate for many deceptions, than the bond recognized between words like house, la maison, and husband, who is its head; between loaf, un pain and lord, un seigneur, whose function was to distribute it; between spur, eperon, and to spurn, mepriser; to glow, briller, and blood, le sang; well, bien, and wealth, la richesse or even thrash, Voire to beat the grain, and threshold, le seuil, packed or close like pavement? 8 9  With the exception of glow and blood, which share the interior phoneme III and the letters L and O, all of Mallarme's examples here share an initial consonant group, reinforced by a common consonant later in the word. Mallarme's desire to see a semantic link between these words sometimes leads him to fanciful lengths, such as his link between loaf and lord. Yet he is also aware that common letters and phonemes do not always point to common meanings: [C]ertain words do not show this conformity of impression; but then are like a dissonance. The reversal in signification may become absolute to the point, nevertheless, where it is as interesting as a true analogy: it is thus that heavy seems to rid itself all of a sudden of the sense of heaviness that it denotes, to furnish heaven, le ciel, high and subtle, considered as a spiritual journey. 90  The challenge then, in Les mots anglais, is to give an account of those letters (and the phonemes they represent) that do seem to point towards a common signification in the language while at the same time avoiding false relations between words. Mallarme explains: Meaning, certainly, and sound, skilfully tried one against the other, here you have the double index guiding the Philologist in the familial classification. Without going so far as to unite dry and thirst, offering a relationship in ideas, but nearly none in form, one must not relate to mow, faucher, to mow, meule, where there is an accord in spirit and letter, because this one comes from a word that means pile and that one from a word that indicates cutting: but the affiliation of words which display only an exterior analogy is first and foremost to be avoided. The Word, in its personality so difficult to recognize,  Mallarme, Les mots anglais, 949. Ibid., 966. "Quelle plus charmante trouvaille, par exemple, et faite meme pour compenser mainte deception, que ce lien reconnu entre des mots comme house, la maison, et husband, le mari qui en est le chef; entre loaf, un pain, et lord, un seigneur, sa fonction etant de le distribuer ; entre spur eperon, et to spurn, mepriser ; to glow, briller, et blood, le sang; well!, bien, et wealth, la richesse ou encore thrash, Vaire a battre le grain, et threshold, le seuil, tasse ou uni comme un dallage ?" Ibid.: "[Cjertains vocables ne montrent pas cette conformite d'impression ; mais alors comme une dissonance. Le revirement dans la signification peut devenir absolu au point, cependant, d'interesser a l'egal d'une analogie veritable : c'est ainsi que heavy semble se debarrasser tout a coup du sens de lourdeur qu'il marque, pour fournir heaven, le ciel, haut et subtil, considere en tant que sejour spirituel." s s  8 9  9 0  90 it is necessary to come back to it: although the Family encompasses great diversity, but which, all, gravitate around something held in common. 91  In Les mots anglais, Mallarme provides tables for each letter that enumerate various constellations. These tables are organized around a common initial consonant, with key words appearing in the left column. The right column contains less fundamental members of the constellation. The example that follows is drawn from the tables in Les mots anglais; the English translation in the lower right column is my own.  w  to wave, ondoyer, et to —er, vaciller, puis to weave, tisser.  to waft, flotter. to waddle, se dandiner, Lat[in] veho. web, toile. weft, trame. weed, vetement de veuvage. woof, etoffe.  to warp, ourdir.  to wrap, envelopper, et —, enveloppe. to swerve, errer.  to writhe, tordre.  wry et awry, tors. wrist, poing. to wrest, torturer. —tie, lutter. wrath et wroth, colere. to wreath, entrelacer.  Tres frequent en tant que lettre initiale, W s'appuie atoutes les voyelles et peut-etre moins aux diphtongues, dont il risque aussi d'etre separe par un h inscrit; on ne le rencontre que devant une seule consonne, r, et il reste muet alors. Les sens d'osciller (celui-ci semblerait du au dedoublement vague de la lettre, puis de flotter, etc.; d'eau et d'humidite ; d'evanouissement et de caprice ; alors, de faiblesse, de charme et d'imagination) se fondent en une etonnante diversite : peut-on, par exemple, dire que wr, authentiquement, designe la torsion, a  Very frequently as an initial letter, W is followed by all of the vowels, perhaps less by diphthongs, from which it may also be separated by a written h; one finds it (W) before only one consonant, r, and it remains mute in that case. The sense of oscillation (this one here seeming to be the result of the vague division into two halves of the letter, then of floating, etc. of water and humidity; of disappearing and of caprice; then, of weakness, of charm and imagination) mixes in a surprising diversity: may one, for example, say that  Mallarme, Les mots anglais,OC II, 966-67. "Le sens, certes, et le son, habilement essayes 1'un a l'autre, voila le double indice guidant le Philologue dans le classement familial. Sans s'aventurer jusqu'a reunir dry et thirst, offrant une relation dans l'idee, mais presque pas en la forme, on doit ne pas rapporter to mow, faucher, a mow, meule, ou s'accordent l'esprit et la lettre, parce que celui-ci vient d'un mot qui veut dire tas et celui-la d'un mot qui indique couper : mais la filiation de vocables n'offrant qu'une analogie tout exterieure est principalement a eviter. Le Mot, dans sa personnalite si difficile a reconnaitre, il faut en revenir a cela : quoique la Famille en compte de tres divers, mais qui, tous, gravitent autour de quelque chose de commun." 91  91 cause de toute une famille nombreuse ou regne ce digramme ? L'appreciation la plus judicieuse a offrir de cette lettre, c'est que : parfois traduisant regulierement Pinitiale g ou v d'une serie entiere de mots qui appartiennent a d'autres langues, elle se trouve, toute grammaticale alors, denuee de vitalite.  wr, authentically, designates twisting, because of the whole of a numerous family reigned by this di-gram? The most judicious appreciation to offer this letter, is that: sometimes translating regularly the initial g or v from an entire series of words which relate to other languages, it finds itself, all grammaticality aside, stripped of its vitality. 92  FIGURE  1.1. Excerpt from Mallarme's Letter Tables (W) in Les mots anglais.  Upon reading the tables, it is immediately striking that Mallarme is unwilling or unable to give a primary signification to practically any letter. Instead, he identifies a few of the principal meanings that can be deduced from the examples, occasionally giving priority to one. Of some letters he has nothing to say whatsoever. This suggests that, far from having a workable scheme in mind for the absolute signification of individual letters, he is still identifying trends and tendencies in language, refining his ideas as he examines individual word families more closely. It is interesting, however, that in the tables there are several examples of words with a strong phonetic connection but only the most tenuous semantic connection. Likewise, there are words grouped together which do not share even an initial consonant (Mallarme's guiding principle) because the semantic link between the two words justifies the weak phonetic connection. Mallarme is aware of the limits of philology to fully explain the signification of letters. That task, if it can indeed be realized, is left only for the writers and poets, those who craft the language daily, seeking out its hidden symmetries and relationships: "It is up to the poet or even to the skilful writer of prose, through a superior and free instinct, to relate some terms whose unity contributes all the more to the charm and to the music of language."  911  Mallarme, Les mots anglais, OC II, 980. Ibid., 967: " A u poete ou meme au prosateur savant, il appartiendra, par un instinct superieur et libre, de rapprocher des termes unis avec d'autant plus de bonheur pour concourir au charme et a la musique du langage."  92  The "music of language" is therefore found in those instances when sound and sense in language are in accord ("whose unity..."), in other words, in places where rhyme suggests a "secret kinship" between words. Mallarme's notion of rhyme here is expanded to include 94  alliteration and, from the tables in Les mots anglais, it is also evident that a shared initial consonant is not essential for a relationship to exist, despite his statements to the contrary. The poet's effort to reconstruct the common origin of language touches on one of its perilous mysteries; and that it will be prudent to analyze only on the day when Science, possessing the vast repertoire of idioms ever spoken on the earth, will write the history of the letters of the alphabet across the ages and what was nearly their absolute signification, sometimes divined, sometimes misunderstood by mankind, creator of words. 95  So the poet may not indeed by able to go all the way back into the history of language to uncover the "absolute signification" of the letters of the alphabet, but he can move in this direction by being attentive to the music of language as it expresses itself through sonic similarities between words.  MALLARME'S GREAT WORK  After Les mots anglais, Mallarme never again attempted to uncover the absolute signification of letters, either in French or English, from a theoretical point of view. However, he continued to manipulate phonemes in his poetry along the lines that were explored earlier. He began to realize that, even if it were possible to assign an absolute value to letters, his aesthetic, which depends so heavily on allusion and suggestion, would be severely limited by such a denotative phonetic system. However, his interest in sound-sense relationships was not abandoned, but merely refocused. After Les mots anglais, Mallarme began to write more Mallarme, Les mots anglais, OC II, 1014. Ibid., 968: "touche a l'un des mysteres sacres ou perilleux de Langage ; et qu'il sera prudent d'analyser seulement le jour ou la Science, poss&iant le vaste repertoire des idiomes jamais paries sur la terre, ecrira I 'histoire des lettres de I 'alphabet a travers tous les ages et quelle etait presque leur absolue signification, tantot devinee, tantot meconnue par les hommes, createurs des mots." Emphasis mine. 9 4  9 5  93  consistently of the Great Work, "a book which would truly be a book, architectural and premeditated, one which would represent the sum total of all literary effort throughout the ages, and would be nothing less than the Orphic explanation of the World." Le Livre was 96  never completed, and the sketches which remain for the work have not yet been reconstructed into a coherent whole. However, Mallarme's descriptions of the work make it clear that the sounds of words would play a prominent role. He claims that Le Livre would be a "total expansion of the letter" that would draw correspondences from its component letters, and institute a literary game confirming the fiction of language.  97  From the fragmentary notes which Mallarme left for Le Livre, we gain further insight into the connection between phonetic relationships in language and abstract categories of 98  thought.  This is done through a series of geometrical figures that represent the relationship of  terms which are likely intended to stand for larger categories of art. In Figure 1.2 the primary relationships are shown with lines that connect individual terms, while secondary relationships are indirectly connected by two lines. Myst[ere] Soi  foi  Th[eatre]  Idee  moi  Dr[ame]  Hymne toi  loi  Heros roi  FIGURE  1.2. Diagram from Mallarme's "Notes en vue du Livre," OC I, 567.  This description of the Great Work comes from a letter addressed to Verlaine, 16 November 1885. See Mallarme, Corr. II, 301: "une livre qui soit un livre, architectural et premedite." Mallarme, "Le Livre, instrument spirituel," OC II, 226. See Jacques Scherer, Le "Livre" de Mallarme (Paris: Gallimard, 1978) and Mallarme OC I, 1372-83.  9 7  9 8  94  As the figure shows, Mallarme envisions primary relationships between Theater and Idea through the intermediary of the Hero, while Drama and Music (here represented by the Hymn) come together in Mystery." Elsewhere in the Notes, he writes "Drama is caused by Mystery of that which follows—the Identity of the Theater and of the Hero through the Hymn—the Hero—the (maternal) Hymn that creates him, and returns to the Theater that which was—from the Mystery where the Hymn is buried."  100  There is much work yet to be done to definitively reconstruct Le Livre from the remaining notes, and it may ultimately prove to be impossible. However, the diagrams found in the notes can help us to understand the relational patterns of Mallarme's thought, and how these patterns find expression in his poetry. In Figure 1.2, Mallarme includes six rhymes on /wa/, one paired with each of the six larger categories.  101  It is tempting to speculate about the  relationship between these larger categories and the rhymes that seem to bind them all together. For example, if the Hero is found at the convergence of Theater and Idea, then perhaps the king (roi) is located at the intersection of faith (foi) and law (loi). The king is indeed the law-giver who rules according to an act of faith by the people, and by Divine Right. Pursuing this idea, if Mystery is found where Drama meets Music (Hymne), then are we to understand that the true self is found only in the relationship between persons ("you" and "I")? In the case of the /wa/ rhymes, it is a phonetic similarity that drives the mind to find appropriate semantic links between the words, even if these sometimes teeter on the brink of the irrational. Similar diagrams are found throughout the notes. They differ somewhat in physical disposition, and a comprehensive treatment of the role of music in Le Livre must one day be Mallarme's fascination with the letter v—treated in earlier in "Sainte Cecile"—is probably related to these vshaped diagrams. Mallarme, Notes en vue du "Livre", OC I, 948: "Le Dr. est cause par le Myst. de ce qui suit - 1'Identite du Theatre et du Heros a travers l'Hymne - Le Heros - L'Hymne (maternel) qui le cree, et se restitue au Th. que c'etait - du Mystere ou elle est enfoui." ' This is the only diagram in the Notes in which every term is connected by rhyme. 9 9  1 0 0  I 0  95 made. For the present purposes, two more examples should suffice to establish the relational pattern of Mallarme's thought. Idee Th[eatre] \ ^  Poes[ie]  Symbole  y ^ ^ ^ \  Heros Myst[ere]  Dr[ame] FIGURE  Hymne  1.3. Diagram from Mallarme's "Notes en vue du Livre" OC I, 572.  Here, Mallarme refocuses several of the terms from the first diagram while adding new ones. Now it is the Symbol which meets Music to become the Idea (or the Idea which splinters into Symbol and Hymn—this reversal is completely typical of Mallarme's mature thought), while Poetry and Theater find common ground in Drama.  Theatre Figure  Mystere Vers Drfame] Livre  heros homme  hymne chant  Poesie ou Orch[estre] FIGURE  1.4.  Diagram from Mallarme's "Notes en vue du Livre" OC I, 578.  In this diagram, the lines are absent, but can certainly be inferred from the physical layout of the terms themselves. Here, Drama (not contemporary theater, but the drama that Mallarme intends from the reading and/or performance of Le Livre) emerges from the intersection of Theater/Figure (actor), Mystery/Verse, hero/man and music. A l l of this is summed up in  96  Poetry, an alternative to the Wagnerian Gesamtkunstwerk which truly operates a fusion of the arts.  102  Robert Greer Cohn has suggested that Mallarme's thought is a modification of a Hegelian dialectic. Essentially, he argues that to the traditional dialectic triad: Synthesis  Becoming  FIGURE  Thesis  Antithesis  Being  Non-Being  1.5. Schematic of a Hegelian Triad  Mallarme added a fourth pole, which Cohn terms "antisynthesis":  Synthesis  Becoming Thesis  Antithesis  Being  Non-Being Antisynthesis  Non-Becoming  FIGURE  1.6. Schematic of Cohn's tetrapolar representation of Mallarme's epistemology  This new fourth pole acts as a counterbalance to the teleology of triadic thinking. In this new scheme, every synthesis is provisional; antisynthesis constantly threatens to undermine any apparently stable relationship. For Cohn, this explains both Mallarme's linguistic practice (since there is no longer a univocal connection between words and meaning) and his more This becomes especially apparent in the term Orchestre which can mean "orchestra" in the instrumental sense, or refer to "the portion of Greek theater consecrated to dance and to the evolutions of the choir." See Littre, s.v. "orchestre." I am indebted to Dr. Richard Semmens for this observation.  97 abstract thought patterns.  103  With these thought patterns in place, I now turn to their application  in Mallarme's late poetry.  Music  IN T H E M A T U R E POETRY  While Mallarme worked, patiently and secretively, on Le Livre, he continued to apply these ideas to shorter poems, which he began to see as studies for this more ambitious w o r k .  104  Individual poems began to explore provisional meaning for particular sounds, discovered in a series or constellation of words and deployed, shaped and reinforced in a particular poem. A significant portion of the constellation shown in Figure 1.2 above is featured also in "Quand l'Ombre  mena9a."  105  Quand l'Ombre menaca de la fatale loi Tel vieux Reve, desir et mal de mes vertebres, Afflige de perir sous les plafonds funebres II a ploye son aile indubitable en moi.  When the Shadow menaced with the fatal law Some old Dream, desire and affliction of my vertebrae, Tormented to perish under the funereal ceilings It folded its unshakable wing in me.  Luxe, 6 salle d'ebene ou, pour seduire un roi Se tordent dans leur mort des guirlandes celebres, Vous n'etes qu'un orgueil menti par les tenebres Aux yeux du solitaire ebloui de sa foi  Luxury, o ebony hall where, to seduce a king Celebrated garlands twist in their death, You are but a pride fooled by the shadows In the eyes of the solitary dazzled by his faith.  Oui, je sais qu'au lointain de cette nuit, la Terre Jette d'un grand eclat l'insolite mystere Sous les siecles hideux qui l'obscurcissent moins.  Yes, I know that far off in this night, the Earth Throws a great flash of unprecedented mystery Under the hideous centuries which obscure it less.  L'espace a soi pareil qu'il s'accroisse ou se nie Roule dans cet ennui des feux vils pour temoins Que s'est d'un astre en fete allume le genie.  Space, its own image whether it grows or denies itself Rolls in this boredom, vile fires as witnesses That the genius of a festive star has illuminated.  Cohn, Mallarme's Un coup de des, 14-18. Cohn has coined the terms "tetrapolar" and "polypolar" to describe Mallarme's new epistemology. His initial argument was made several years before the notes for Le Livre were published by Scherer, though they seem to confirm his ideas. Cohn does not explore the musical nature of thought that I have outlined here, and his conclusions are somewhat different than mine—he argues, for instance, that despite this epistemological pattern, each of Mallarme's poems still has a principal and discoverable meaning. Derrida, using Cohn's theories, argues the opposite: that Mallarme's poetry is not reducible to a singular, stable collection of meanings. Mallarme, Corr. II, 301. Letter to Verlaine, 16 November 1885. Mallarme, OC I, 36. I O j  1 0 4  1 0 5  98 The topic of this poem—the poet's struggle to write—is by now a familiar one. The "vieux Reve," whether understood as the Great Work or more generally as his effort to remake language, is once again threatened by doubt ('T Ombre") with the "fatale loi" (the constraints of grammar and syntax, the limits of meaning, etc.). In the face of this opposition, the Dream lays down its pen ("il a ploye son aile"), its flight denied once again by the darkening skies. The ebony hall with its twisting garlands calls to mind the empty room of the "Sonnet en yx," but can also be seen as the night sky, the garlands representing stars enchained into constellations.  106  Their seduction, suggesting alternative paths of meaning and analogy, is a  conceit of the faithful man (the poet who continues to believe in their potential despite his inability to grasp it in its totality: he is blinded by his faith). Far off into the night, the Earth casts its own light (like the stars which illuminate the night sky) which will not darken with the passage of time. Space (unchanging and eternal) wanders in boredom, while the stars (now vile fires, taunting the poet) are witnesses to the poetic inspiration which one bright star inspires. Five of the six rhymes on /wa/ from the sketches for Le Livre appear also in "Quand l'Ombre menaca": loi - moi - roi - foi - soi  107  The first four of these are particularly emphasized as end-rhymes in the first and second stanzas. From these rhymes, the phoneme /a/ is detached, and comes to particular prominence in the first line of the poem, where six /a/s in close proximity suggest the sterility of the situation: menaca de la fatale l o i  108  Cohn, Toward the Poems, 121. The omission of "toi" is curious. The idea is certainly picked up in the Vous of line 7. Perhaps Mallarme could not find a convincing way to work it into the poem, which would certainly fit in with the pessimistic tone of the opening. Cohn, Toward the Poems, 120. 106  107  108  99  From this constellation, the phoneme reaches across the poem to other key words: Afflige - plafonds - moi - roi - foi - soi - accroisse - astre - allume It cannot be mere coincidence that these sounds, themselves the product of a poetic constellation, here connect words which mean "sky," "star," and "light." It also helps to explain Mallarme's choice of plafond rather than del for sky, and the metaphorical reading that it necessitates. Written only a few years after "Quand l'Ombre menaca," "Le vierge, le vivace, et le bel aujourd'hui" represents the most abstract stage of development in Mallarme's exploration of the sound-sense relationship in poetry.  110  Le vierge, le vivace et le bel aujourd'hui Va-t-il nous dechirer avec un coup d'aile ivre Ce lac dur oublie que hante sous le givre Le transparent glacier des vols qui n'ont pas fui!  The virgin, enduring and beautiful today Will it tear for us with a drunken wingbeat This hard, forgotten lake that haunts beneath the frost The transparent glacier of flights not flown!  Un cygne d'autrefois se souvient que c'est lui Magnifique mais qui sans espoir se delivre Pour n'avoir pas chante la region oil vivre Quand du sterile hiver a resplendi l'ennui.  A swan of former times remembers that it is he Magnificent but who without hope of saving himself For not having sung the region in which to live When sterile winter shone its boredom.  Tout son col secouera cette blanche agonie Par I'espace infligee a l'oiseau qui le nie, Mais non I'horreur du sol oil le plumage est pris.  His whole neck will shake off this white agony By space inflicted on the bird who denies it, But not the horror of the ground where the plumage is caught.  Fantome qu'a ce lieu son pur eclat assigne, S'immobilise au songe froid de mepris Que vet parmi l'exil inutile le Cygne.  Phantom whom to this place his pure brilliance assigns him, Immobile in the cold dream of scorn Which is worn amid the useless exile by the Swan.  A swan is trapped in the ice of a frozen lake, fixed, unable to take poetic flight or dive into the water. He is a victim of his own inaction; he did not take flight with the other birds in autumn ("n'avoir pas chante la region ou vivre") and his impotence has frozen around him as has the lake. He hopes that the rays of a new dawn might melt the ice, free him from his earthly O f course, "plafond" resonantes with "ploye" and "perir." The implication of enclosure in "plafond" is also relevant here. Mallarme, OC I, 36-37. It should be noted that Pierre Boulez set this poem, along with "Une dentelle s'abolit" and " A la nue accablante tu" in his Improvisations sur Mallarme. 109  110  100 prison and let him spread his wings skyward once more. He shakes the frost from his neck but his feathers remain caught; there will be no salvation today. Instead, the swan remains in the prison of his own making, his proud plumage stilled, awaiting the day when freedom will come. A conventional reading of this poem, which is no less plausible for that fact, would equate the poet with the swan, and to read the poem autobiographically as another instance of Mallarme turning his poetic impotence (in the face of the Great Work) into poetry, as he did with countless other poems. The plumage which is caught would then be a metaphor for the poet's immobile pen, the feather or wing which awaits the promise of a new dawn. When this new day brings no new flight, the poet covers himself in scorn (for conventional uses of language) and reconciles himself to wait for another day and another time. An alternative reading would take the "Cygne" as a homonym for "signe" (and it is no coincidence that Mallarme expressly rhymes these two sounds in the final tercet of the poem).  111  It would then be the sign—the word—and not the poet who is hoping that the new  day brings a poetic flight, that it would enter the poetic Jeu supreme and find liberation from *  112  its use in the language of commerce. The high concentration of I'll phonemes in this sonnet has led some to refer to it as the "Sonnet en i," a counterpart to the "Sonnet en^x."  113  The inclusion of this phoneme in each of  the sonnet's fourteen rhymes, and the extreme concentration of /i/s in the final tercet of this poem leaves little doubt that Mallarme was consciously exploiting this sound. The brightness of this sound is related to the dawn of the new day and the promise that it brings. It also  " ' Mallarme often uses a component of one or several words to create unexpected rhyme. See for example "tristement dort une mandore" from "Une dentelle s'abolit" - where "-ment dort" and "mandore" form an unexpected rhyme. OC I, 42-43. 1 see no reason why these two analogical interpretations (and others) can not coexist. After all, it is the simultaneous presence of two meanings that drives analogy and metaphor in the first place. See page 80 of the present study. 112  113  101 represents the "pur eclat" of the swan, a good example of Mallarme reinforcing an individual l\l phoneme in "Cygne" (itself a mot juste) by surrounding it with other words that contain the  same phoneme, even at the expense of clear semantic meaning and some twisted syntax. As in "Saint Cecile," there is also a particular concentration of NI phonemes, here signifying both the swan's now immobile wing and the poetic flights ("vols"), both those formerly unflown and those promised by the new day. The NI constellation in this poem is significantly longer than that in "Sainte Cecile," and even if allowances are made for fortuitous coincidences (such as va and avec), this constellation includes many of the poem's most important images: vierge - vivace - ivre - givre - vols - souvient - delivre - vivre - hiver - vet It is not merely the number of Vs in the piece, but their distribution that is significant here. In the octave, where the possibility of the new day is still uncertain, there are 13 NI phonemes, reinforcing the semantic suggestion of flight. In the sestet, where it is evident that today will not bring the desired flight, the V virtually disappears, with only a single instance in the final line of the poem. As in "Sainte Cecile," many of the key words in the poem are united by the unordered pair of phonemes NI I'll. If the NI is suggestive of flight and the HI of both the promise of the new day and the brightness of the swan in this poem, it may well be motivated by their convergence in the word "Cygne" itself, where the HI phoneme is represented graphically by the v shape of the letter^. The four end-rhymes on HVRI—ivre  / givre / delivre / vivre—are particularly striking  here, as much for what they contain as for what they omit." In his perceptive commentary, 4  Robert Greer Cohn has noted that these four rhymes are reinforced in the words "vierge" and  1 1 4  Mallarme uses the same four rhymes in the unpublished "Eventail de Mery Laurent," OC 1, 68.  102 "hiver" through the unordered quartet of letters v, /', e, r.  1 1 5  Following the customs of French  prosody, there are only eight rhymes available on / I V R / : cuivre - delivre - givre - guivre -ivre - livre - vivre - suivre. Some of these are semantically unsuitable for this poem, like "guivre," a heraldic term meaning serpent." But the absence of "livre," a word so close to the heart of 6  Mallarmean poetics cannot be simply by chance. Semantically, this absence is fitting, since the swan's earthly prison is fashioned precisely by the fact that he has not been able to sing of the Great Work. The absence of "livre" in the text of the poem is thus made present, a common theme in Mallarme's works. This presence-through-absence is reinforced through phonetic echoes of "livre" scattered through the text: d'aile ivre /de 1 i v R /  delivre /deliVR/  le givre ih glVR/  sterile hiver / l lVeR/  Several of these phonetic echoes are accompanied by anagrams of "livre" (present also in "Le vierge") in the text that are also richly suggestive. Like the swan, these letters are trapped inside their constituent words, awaiting a "coup de livre" ("coup d'aile ivre") to free them.  117  T H E C A S E OF W A G N E R  In the summer of 1885, responding to a request from Edouard Dujardin, Mallarme wrote his first and most important essay on Wagner.  118  It is significant, given the frequency  with which Mallarme's aesthetic is seen as a response to Wagner, that this essay comes relatively late in Mallarme's life and only in response to a request for an article. By his own Cohn, Toward the Poems, 125. Mallarme does in fact use this word in "Mes bouquins refermes," OC I, 45. Mallarme famously described Le Livre as the "expansion totale de la lettre, doit d'elle tirer, directement, une mobilite et spacieux, par correspondances, instituer un jeu, on ne sait, qui confirme la fiction" ("total expansion of the letter, must draw from it, directly, a mobility and spacious, through correspondences, to institute a game, one does not know, that confirms the fiction"). See "Le Livre, instrument spirituel," OC II, 226. Dujardin was a founding editor of the Revue wagnernienne and an important figure in the spread of French 115  1 1 6  117  118  wagnerisme.  103 admission, Mallarme had never seen any of Wagner's works in performance.  119  heard excerpts of Wagner at the  concerts spirituels,  However, he  and counted some prominent Wagnerians  as close friends, notably Villiers de l'lsle Adam and Eugene Lefebure. In the spring of 1885, he read  Quatre poemes d'operas tradults en prose francais, precedes d'une lettre sur la musique  par Richard Wagner  (1861), and the numerous critical commentaries and translated extracts of  Wagner's theoretical works in contemporary Parisian journals likely exposed him to the dramatic ideas of  Opera and Drama  in summary form.  120  Half article, half prose-poem,  Mallarme's "Richard Wagner, reverie d'un poete francais" is at once appreciation of Wagner's music dramas and a warning against using Wagner as a model for French poets to follow. Written at the beginning of the most fervent period of French  wagnerisme (1885-1900),  Mallarme's essay anticipates many of the criticisms raised by later French writers, notably Debussy. Throughout the essay, Mallarme modulates between genuine admiration of Wagner and a deep-seated suspicion of the  Gesamtkunsrwerk  as a viable way to achieve the union of the  arts to which both men aspire. Ultimately, and predictably, the foundations of Mallarme's critique of Wagner are based on the idea that Wagner does not go far enough in his effort to reconcile music and language. So while praising Wagner's effort as one of the best in contemporary theater, Mallarme is also eager to show that Wagner's music dramas do not in fact achieve the original fusion of the arts that they intend. Mallarme sees Wagner as the "singular challenge of poets, whose rights he has usurped with the most candid and splendid bravery."  122  This is really the crux of the case against  Wagner, and the foundations of Mallarme's position vis-a-vis contemporary music in general:  1 1 9  1 2 0  121  1 2 2  Mallarme, Corr., II, 290. Letter to Dujardin, 5 July 1885. Mallarme, Corr., II, 289. Letter to Kahn, 17 April 1885. Mallarme, "Richard Wagner, reverie d'un poete francais," OC II, 153-59. Ibid., 154: "Singulier defi qu'aux poetes dont il usurpe le devoir avec la plus candide et splendide bravoure."  104 that music has infringed upon poetry. Music presents itself as if it were the originator of mystery, when in fact it is merely a derivative (no matter how compelling) of language. The challenge to poets, therefore, is not to make their poetry more musical, not to make a copy of a copy, but instead to carry the union of the arts back further than Wagner, and in a somewhat different direction. In contemporary theater, Mallarme sees Wagner as the one whose effort goes furthest towards reuniting the arts under the direction of Poetry. For him, Wagner understands that in the process of reunification, traditional elements of the individual arts will be transformed from their current status, and put to a different purpose. The essay "Solennite"—written in 1887 after the failed revival of Lohengrin—clarifies this position: In Wagner, even, [...] I do not find, strictly speaking, theater (surely one will find more, from the dramatic point of view, in ancient Greek theater or Shakespeare), but legendary vision which suffices under the veil of sonorities and mixes there; nor are his scores, compared to Beethoven or Bach, simply, music. Something special and complex results: situated at the convergence of the other arts, issuing from them and governing them, Fiction or Poetry. 123  This is the source of Mallarme's genuine and abiding respect for Wagner's genius. In the "Reverie," Mallarme praises Wagner for revitalizing the theater, and credits his music for liberating it from a strictly representational function. After Wagner, the dramatic action of theater was more abstract, and therefore more likely to be read as independent allegory. This paves the way for art to replace religion, a goal common to both Wagner and Mallarme. He also lauds Wagner for representing the onstage dramatic action in the orchestra through the  Mallarme, "Solennite," OC II, 203: "Chez Wagner, meme, [...] je ne percois, dans l'acception stricte, le theatre (sans conteste on retrouvera plus, au point de vue dramatique, dans la Grece ou Shakespeare), mais la vision legendaire qui suffit sous le voile des sonorites et s'y mele ; ni sa partition du reste, comparee a du Beethoven ou du Bach, n'est, seulement, la musique. Quelque chose de special et complexe resulte : aux convergences des autres arts situee, issue d'eux et les gouvernant, la Fiction ou Poesie."  105  leitmotive. For him, Wagner's orchestra is the source of the actor's movements and emotions on stage, and this degree of integration is an important development in contemporary drama.  124  Yet Mallarme also has serious, philosophical reservations about the Wagnerian project. He admits that "of the two elements of beauty that exclude each other and, at least, ignore one 125  another, personal drama and ideal music, [Wagner] created a hymen."  While this seems to  be a valorization of Wagner's achievement, Jacques Derrida has argued persuasively that Mallarme's use of the word "hymen"—signifying both virginal membrane and the Greek god of marriage in French as in English—is an intentional syllepsis that speaks of admiration and doubt in the same sentence.  126  In Wagner's Gesamtkunstwerk, there is not, as many would  claim, a union of the arts, but merely a juxtaposition of music and drama. In Wagner, Mallarme argues that "everything has been made, rather than irradiated, by a direct game, from the literary principle itself."  127  Like literature (and particularly poetry),  Wagner's music dramas attempt to reconstruct the essential unity of the arts. However, instead of tracing this unity all the way back to a common original language as Mallarme outlined in Les mots anglais,  Wagner is contented with an assemblage of the various arts on stage. This  juxtaposition of the arts is a constructed simulation of literature insofar as it attempts to bring the arts together. But it does not emanate from a central source (language) and therefore is a simulacrum of poetry. The true union of the arts is only possible, for Mallarme, when united under the rubric of language, their unique source. This is the ultimate duty and the sole province of the poet.  It is unlikely that Mallarme would have understood Wagner's more subtle symphonic techniques. His criticism is largely based on the one-to-one relationship that he sees between the leitmotive and the characters onstage. Mallarme, "Richard Wagner," OC II, 155: "des deux elements de beaute qui s'excluent et, tout au moins, l'un l'autre, s'ignorent, le drame personnel et la musique ideale, il effectua l'hymen." Derrida, "The Double Session," 209-222. Mallarme, "Richard Wagner," OC II, 154: "que tout soit fait, autrement qu'en irradiant, par un jeu direct, du principe litteraire meme." 124  1 2 5  1 2 6  1 2 7  106 As such, Wagner never actually changes anything fundamental about the theater; he merely fuses its disparate elements into a simultaneous, multi-media presentation. From this philosophical position, Mallarme argues, ironically, that despite his monumental effort, Wagner fails to unite the arts because the very principle on which his effort is based fundamentally misunderstands Music.  128  This why Mallarme sees Wagner's effort as returning "to the waters of the primitive stream: not all the way to the source," and the temple of Wagnerian art as merely "halfway up the saintly mountain." For him, the Wagnerian project cannot be completed on the stage, but 129  only through poetry. In place of Wagnerian legend, he proposes: Fable, virgin of everything, place, time and person also, [which] reveals itself borrowed from the latent meaning in the concourse of everything, the one inscribed on the page of the Heavens and of which History itself is only the interpretation, vain, that is to say a Poem. 130  In order to reach the kind of abstraction that Mallarme sees as necessary, the physical trappings of Wagner's music dramas must be cast aside ("virgin of everything, place, time and person also"). The Poem which springs forth from "the latent meaning in the concourse of everything" is nothing other than the Music which Wagner misunderstands. This Music is the "hymn, harmony and joy [...] of the relations between all things" that Mallarme would later insist on as a fundamental characteristic of Le Livre, a kind of personal and abstract drama that would rise to the challenge of Wagner's pseudo-Greek spirituality and embody Music par excellence. Shortly after publishing Mallarme's "Reverie," Dujardin solicited another contribution from the poet, this one a poem. Mallarme was not particularly enthusiastic about the project, desiring above all not to add a banal epilogue to "the many suggestive things written on  Mallarme, "Richard Wagner," OC II, 156. Ibid., 157; 158. Ibid., 157: " A moins que la Fable, vierge de tout, lieu, temps et personne sus, ne se devoile empruntee au sens latent en le concours de tous, celle inscrite sur la page des Cieux et dont l'Histoire meme n'est que Interpretation, vaine, c'est-a-dire un Poeme." 1 2 8 1 2 9  1 3 0  107  Wagner."  131  Despite his reservations, the sonnet was published, along with several others, in  the edition of the Revue dated 8 January, 1886.  Hommage Le silence deja funebre d'une moire Dispose plus qu'un pli seul sur le mobilier Que doit un tassement du principal pilier Precipiter avec le manque de memoire.  The already funereal silence of a moire Lays out more than a single fold on the furniture Which a crushing of the principal pillar must Cast down with the lack of memory.  Notre si vieil ebat triomphal du grimoire, Hieroglyphes dont s'exalte le millier A propager de l'aile un frisson familier ! Enfouissez-le-moi plutot dans une armoire.  Our age-old triumphant struggle with the grimoire, Hieroglyphs which exalt the masses To spread from the wing a familiar shiver! Rather hide it for me in an armoire.  Du souriant fracas originel haT Entre elles de clartes maitresses a jailli Jusque vers un parvis ne pour leur simulacre,  From the smiling original fracas hated Among them master lights has flashed Up to a parvis born for their simulation,  Trompettes tout haut d'or pame sur les velins, Le dieu Richard Wagner irradiant un sacre Mal tu par l'encre meme en sanglots sibyllins.  Trumpets aloud of gold swooning on the velum, The god Richard Wagner radiating a sacrament Poorly silenced by the ink itself in sibylline sobs.  132  1  134  The sonnet opens in an abandoned room where the silks covering the furniture act as a funeral shroud for a theater mourning the death of the Master. The folds of the cloth itself, and the hollows created by its hanging on the furniture, suggest-the yawning chasm of a tomb. The "principal pilier"—presumably Wagner himself—is in the grave, and yet the weight of his  Mallarme, Corr. II, 290. Letter to Dujardin, 10 September 1885: "tant de choses suggestives ecrites sur Wagner." A variety of cloth that "communique un eclat changeant, une apparence ondee et chatoyante," so that "la beaute de la moire residant dans la variete des dessins changeant avec la position du spectateur." Surely this refers also to the variety of French perspectives on Wagner mentioned in the "Reverie." See Emile Littre, Dictonnaire de la langue frangaise (Paris: Redon, 1998) C D - R O M , s.v. "moire." This work was Mallarme's favorite dictionary, and is one of the best sources of information on contemporary meaning and pronunciation. Hereafter, all references to this work shall use "Littre" and the relevant entry. Littre gives two relevant meanings here: a sorcerer's book used to conjure demons and, figuratively, any difficult or obscure writing. See Littre, s.v. "grimoire." According to Littre, a courtyard in front of the main door of a church, particularly a catherdral. s.v. "parvis" 131  132  133  1 3 4  108  influence is still felt in contemporary theater, as he casts down the trappings of an outmoded 135  theater "le mobilier" into oblivion. In the second quatrain, Mallarme contrasts this with his own going concern: poetry, and his ongoing struggle with the Great Work. References to the grimoire and its hieroglyphic writing designed to uplift the masses are meant to contrast with the quasi-religious demagoguery of the Wagnerian cult. The "frisson familier" of the wing is probably the movement of the pen, which Mallarme frequently equates with bird imagery, owing to the dual meaning of "plume" as both "feather" and "pen." The closing request to stow the long-awaited book in an armoire speaks to Mallarme's patient approach to his own work. The negative connotations of "enfouissez-le-moi" are tempered somewhat by the fact that an armoire is not merely another piece of furniture but, etymologically, an armory, and the book that it contains a weapon as powerful as a bomb, but without the collateral damage.  136  Untwisting the syntax of the tercets, we may read the following: "From the smiling original fracas of master lights, among these one [Wagner] has flashed towards a parvis born for their simulation."  137  Here we find the same mixture of admiration and criticism that  characterized the "Reverie." If Wagner is indeed the god of a renewed religion, then his pulpit stands outside the temple itself, a kind of sideshow along the road. Yet Wagner is one of the supreme lights (or truths) of the "original fracas," the generative musical language that furnished the world's idioms and created instrumental music in its splitting apart. Wagnerian trumpets swoon on the pages of his scores, their gold referring at once to their sound, their physical colour and the illuminated manuscripts on which Mallarme imagines all music to be  The identity of the "principal pilier" is a matter of some debate. Lloyd James Austin suggests Victor Hugo, noting that Hugo died on Wagner's birthday, 22 May. Cited in OC I, 1198. The reading is tantalizing, given Hugo's importance to French theater and Mallarme's respect for his poetry. However, the main sense of the passage—renewal of the theater—is not really affected by the identity of the one doing the renewing. See Mallarme, " L a Musique et les Lettres" OC II, 72. 1 3 5  1 3 6  1 3 7  See Cohn, Toward the Poems, 181.  109  written.  138  These trumpets—and Wagner's music in general—are badly silenced both by the  ink with which they are inscribed on his scores, and by the ink of the poet's pen whose efforts may be overwhelmed by the Wagnerian tide. The semantic opposition that Mallarme sets up between Wagnerian art and the Great Work—represented in "Hommage" as the grimoire—is essentially  an opposition between  music-as-instrumental-sound and the music of language. It is hardly a coincidence, therefore, that "grimoire" derives etymologically from the Latin gramma, meaning "letter." For i f the music of language is played out in the relationships between its constituent letters, then Mallarme's "Hommage" to Wagner enacts its own criticism of his music dramas through the very fabric o f its own diction. A s Graham Robb has shown, much o f Mallarme's vers de cirConstance (which by definition includes all the tombeaux and hommages) draws its inspiration from the names o f the persons to whom they are dedicated.  139  In the quatrain which decorates the copy of  "L'Apres-midi d'un faun" which Mallarme gave to Debussy, the unordered phoneme /si/, taken from the last phonemes o f his name, appears with surprising regularity:  Sylvain d'haleine premiere /si/  III  Si ta flute a reussi  /si/  lyl  /y//si/  Ou'is toute la lumiere /is//u/  lyl HI  Q u ' y soufflera Debussy /is//u/  /y//si/  See Mallarme's description of orchestral scores in "Heresies artistiques," OC II, 360. Robb, Unlocking Mallarme, 84.  110 If we consider the independent /i/s in "premiere" and "lumiere," then each line of verse both begins and ends with sounds drawn from Debussy's name. Likewise, the numerous lyl and /u/ phonemes pick up on the penultimate syllable of "Debussy."  140  Returning to "Hommage," all eight lines of the octave end with the unordered letter cluster /, e, r, which is drawn from the first and last letters of Wagner's name. The cluster here may well be motivated by the graphic shape of the W, which suggests a redoubling or folding in on itself.  141  However, unlike the Debussy example, these letters are connected to Wagner's  name by their graphology alone: there is no phonetic similarity here.  142  From Mallarme's point  of view, this gesture posits the superiority of poetry over music, since it has this additional dimension of meaning readily available to the reader. In his response to a survey on the topic 143  of graphology, Mallarme claimed that writing was an index, a transcription of the poet's thought set down in a permanent form, analogous to a gesture of the body.  144  Therefore, his  writing is amenable to a graphic analysis as much as a semantic or phonetic one. This play of letters is intimately connected to Mallarme's notion of the grimoire: the Great Work which will eventually rise to the challenge of the Gesamtkunstwerk. So the old struggle with the grimoire ("notre si vieil ebat du grimoire") is, at least in part, a struggle with letters and their relationships, hieroglyphs whose true meaning is revealed only to those who The lyl and lul phonemes are close enough on the French vowel scale to make the phonetic similarity plausible here. See Mallarme's commentary on the letter w in Les mots anglais, OC II, 980. He argues that the multiplicity of meanings which w seems to convey may be the result of the letter's unique graphology. Given Mallarme's expertise in English, it is possible that there is a phonetic echo of Wagner's last name (pronounced Ava/ rather than /va/, but evoking the English graphology. While this may seem fanciful, echoes of English are sometimes evoked in Mallarme: for example, the echo of "I'll be none" in the title "Igitur ou la folie d'Elbehnon." It does seem more than coincidental that a letter — w — which does not exist in French proper should be invoked through the French letters oire. It seems to me that Mallarme would have been struck by the fact that "this foreigner" Wagner is marked graphically as such. 1 4 0  141  1 4 2  O f course, musical notation offers a comparable procedure in the play of enharmonic spellings. However, the average person consumes music through the ear primarily (as in listening to a performance), and only a trained musician can understand the enharmonic respellings of Beethoven, etc. By contrast, nearly everyone consumes poetry through the eye first, and then through the mind's ear. Although this is a disadvantage, Mallarme turns this into an advantage through a careful sensitivity to graphology. Mallarme, "Sur la graphologie," OC II, 669. 143  1 4 4  Ill  are willing to look past the semantics of language for additional layers of meaning. That Mallarme describes this struggle as "triomphalement" may have less to do with his actual success in writing the book than in the letter cluster m, o, i, r which unites the grimoire with its three other rhymes, and with the ultimate triumph of language: moire  - memo/re - grimoire - armoire - tn'omphalement  145  These additional ways of meaning are, in Mallarme's mind, unique to poetry. They are the tools with which he will mount his ultimate challenge to Wagnerian music drama in Le Livre.  M u s i c IN T H E L A T E C R I T I C A L W O R K S  The sound-sense relationship continues to be a powerful one in Mallarme's later writings, especially in the essay "Crise de vers" (1887-1895). "Crise de vers" is an extended 146  reflection on the nature of contemporary poetry. It is here that Mallarme makes some of his most audacious claims about the primacy and ubiquity of poetry, and about its relationship to music. Here again, but in a more mature form, Mallarme insists on the supremacy of the soundsense relationship in poetry, and on the musical basis of this phenomenon. What is new here is the confrontational tone that Mallarme adopts towards actual symphonic music. This, no doubt, is another response to wcignerisme, and to the increasing valorization of music as the art to which all others aspire. But the ideas are not fundamentally different from those put forward in his earlier writings; they simply show an evolution born of maturity and experience.  Mallarme pairs these letters in the opening quatrain of "Prose pour des Esseintes," a work supremely concerned with the poetic techniques that will lead to the Great Work: Hyperbole! de ma memoire Tr/owphalement ne sais-tu Te lever, aujourd'hui grimoire Dans un livre de fer vetu: Mallarme, OC II, 204-213. The article is constructed from various fragments published between 1887 and 1895. A complete chronology of the independent parts is contained in OC II, 1643. 145  146  112  Mallarme writes of the fracturing of language into "a thousand simple elements" (letters and their corresponding phonemes) that are "similar to the numerous cries of an orchestra, which remain verbal."  147  The suggestion here is not, as it is sometimes taken, that  contemporary poetry should try to capture the sounds of contemporary music, but rather that poetic music—the play of letters and semantics—is comparable to symphonic music in its play between sound and meaning. Mallarme writes: Certainly, I never sit on the concert tiers, without perceiving amid the obscure sublimity a certain sketch of one of those poems immanent in humanity or their original state, all the more understandable since it is silent and that in order to produce the vast line the composer tested his ability to suspend himself on the edge of the temptation to explain himself. I imagine, no doubt through an ineradicable writer's prejudice, that nothing can endure without being uttered; that we are there, precisely, to seek out, before a fracturing of the great literary rhythms [...] and their scattering into articulated shivers near to instrumentation, an art of completing the transposition, to the Book, of the symphony or simply to take back our own for it is not from the elemental sonorities of brass, strings and woodwinds that Music results, but undeniably from the intellectual word at its apogee, with totality and evidence, as the ensemble of relationships existing in all, Music. * 14  Mallarme hears a kind of latent poem underneath symphonic music, one which seeks its own completion in a kind of transposition into language.  149  This transposition of the symphony to  the Book depends on the fragmentation of language into its phonemic components (its "articulated shivers near to instrumentation") in which sound and sense are reconciled in an effort to contain the "ensemble of relationships existing in all." The "intellectual word"— literature in its highest form—can accomplish this transposition since it can mean both phonetically and semantically. Symphonic music,, for Mallarme, has only a "phonetic"  Mallarme, "Crise de vers," OC 11, 205. Mallarme, "Crise de vers," OC II, 212: "Certainement, je ne m'assieds jamais aux gradins des concerts, sans percevoir parmi l'obscure sublimite telle ebauche de quelqu'un des poemes immanents a I'humanite ou leur originel etat, d'autant plus comprehensible que tu et que pour en determiner la vaste ligne le compositeur eprouva cette facilite de suspendre jusqu'a la tentation de s'expliquer. Je me figure par un inderacinable sans doute prejuge d'ecrivain, que rien ne demeurera sans etre profere ; que nous en sommes la, precisement, a rechercher, devant une brisure des grands rythmes litteraires [...] et leur eparpillement en frissons articules proches de 1'instrumentation, un art d'achever la transposition, au Livre, de la symphonie ou uniment de reprendre notre bien car, ce n 'est pas de sonorites elementaires par les cuivres, les cordes, les bois, indeniablemeni mais de I'intellectuelle parole a son apogee que doit avec plenitude et evidence, resulter, en tant que I'ensemble de rapports existant dans tout, la Musique. " Emphasis mine. Mallarme attended the Lamoureux concerts regularly beginning in 1885. He was often seen sketching in notebooks during the concerts, as if making a transliteration of the music that he heard in poetic terms. These sketch books have not yet been positively identified. l 4 /  148  1 4 9  113  meaning: the sonorities of its instruments suggest mood, emotion, basic meanings (light, dark, etc.). But because music lacks a semantic dimension it cannot play between the two levels as poetry can. This is why the word is, for Mallarme, the highest and purest form of music. If the word is music, then the voice must be its instrument. This idea, which we encountered earlier in "Le demon de l'analogie" resurfaces here: Languages, which are imperfect due to their multiplicity, lack the supreme language: thinking being to write the immortal word without accessories, nor whispering but still silent, the diversity, on earth, of languages prevents anyone from proffering words that, otherwise would find themselves, by a unique stroke, materially the truth. This prohibition rages expressly, in nature (one stumbles upon it with a smile) so that there is no worthy reason to consider oneself God; but, at present, turned towards aesthetics, my feeling regrets that discourse fails to express objects through keys that respond to them in colour or appearance, those existing in the instrument of the voice, among languages and sometimes inside one. Beside ombre, opaque, tenebres darkens only a little; what deception, before the perversity conferring to jour as to nuit, contradictorily, timbres dark for the former, the latter light. The wish for a term of brilliant splendor, or if it extinguishes itself, the inverse; as for simple luminous alternatives— Only, we know, verse would not exist: it, philosophically pays back the defect of languages, superior complement. 150  Here, in perhaps its most famous formulation, is Mallarme's mature position on music and poetic language. We see again the idea that the plurality of modern languages obscures an originary one in which sound and sense would be in accord ("words that ... would find themselves, by a unique stroke, materially the truth"). The old dream—of rediscovering this generative idiom—returns here, but now accompanied by the realization that if it did, poetry would be non-existent, for all language would be in itself inherently poetic, and the poet would be Mr. Everyman. It is precisely in the failings of modern French that the poet finds his 151  Mallarme, "Crise de vers," OC II, 208: "Les langues imparfaites en cela que plusieurs, manque la supreme : penser etant ecrire sans accessoires, ni chuchotement mais tacite encore Pimmortelle parole, la diversite, sur terre, des idiomes empeche personne de proferer les mots qui, sinon se trouveraient, par une frappe unique, elle-meme materiellement la verite. Cette prohibition sevit expresse, dans la nature (on s'y bute avec un sourire) que ne vaille de raison pour se considerer Dieu ; mais, sur l'heure, tourne a de l'esthetique, mon sens regrette que le discours defaille a exprimer les objets par des touches y repondant en coloris ou en allure, lesquelles existent dans I'instrument de la voix, parmi les langages et quelquefois chez un A cote d'ombre, opaque, tenebres se fonce peu ; quelle deception, devant la perversity conferant a jour comme a nuit. contradictoirement, des timbres obscur ici, la clair. Le souhait d'un terme de splendeur brillant, ou qu'il s'eteigne, inverse ; quant a des alternatives lumineuses simples—Seulement sachons n'existerait pas le vers : lui, philosophiquement remunere le defaut des langues, complement superieur." Emphasis mine (Mallarme's original emphasis is underlined.) Genette, Mimologics, 215-18. 1 5 0  L  151  114 calling, to find or manufacture "keys in the instrument of the voice" which correspond to the semantic meaning of the words it pronounces when chance and the passage of time have corrupted or fragmented them. Once the poet has found (or approximated through analogy) the mot juste, the poem will create itself as its component letters call forth other words which rhyme (in the most expanded sense of the word), bringing ever more unusual analogies to mind. The poet's task is then to notate these impressions, to weigh their symmetries, to find their common meaning: [T]he poetic act consists of a sudden vision that an idea splinters in a number of motives equal in value and in grouping them; they rhyme: as an external seal, their common measure that the final throw relates. 152  Mallarme's proud vision of a new language also endures here. Mallarme now understands that recovering a generative language is not possible, but that a poetic approximation of this can be its equal, and may indeed be superior. This language, removed from common usage, is a fiction, an artificial and modern representation of a language which, if it ever existed, has vanished. Mallarme describes it thus: The verse which from several vocables remakes a total Word, new, foreign to the language and as i f incantory, achieves this isolation of the word: denying, by a sovereign stroke, the chance remaining in terms despite the artifice of their alternative re-tempering in sense and sonority, and causes you this surprise of having never heard such an ordinary fragment of speech, at the same time that the memory of the named object bathes in a new atmosphere. 153  L A MUS1QUE ET LES LETTRES  Perhaps the most important of Mallarme's writings on the relationship between music and poetry was "La Musique et les Lettres," which occupied Mallarme between 1892 and °" Mall