On 22 December 1894, a thirty-five-year-old Jewish artillery captain in the French army, Alfred Dreyfus, was convicted of the treason of communicating French military secrets to the German Embassy in Paris, and sentenced to life imprisonment and solitary confinement on Devil’s Island off the coast of French Guiana. That same evening Debussy’s Prélude à l’après-midi d’un faune had its first performance in a concert under a young Swiss conductor, Gustave Doret, in the Salle d’Harcourt. The work, which had been publicly rehearsed in the afternoon, was encored in the evening, despite the fact that the performance, by some accounts, left a good deal to be desired.

This curious coincidence of events might be taken as emblematic of the relation between French art and French political life as the nineteenth century drew to a troubled close. Debussy’s faune, dreaming his sexual fantasies obscurely and wordily on a hot afternoon in, presumably, Attic Greece, seems unimaginably remote from the political fears and prejudices of 1890s Paris that lay behind Dreyfus’s wrongful conviction. Debussy himself seems not to have been particularly interested in the Dreyfus affair, which, then and for many years afterwards, divided French opinion, French society and even French families to an extent hard to recapture today. According to his friend the writer René Peter, his instincts were on the side of the nationalist anti-Dreyfusards, but Peter admits that he failed to provoke the composer into anything approaching anger on the subject either way. Edvard Grieg’s gesture of refusing to conduct in France in 1899 because of Dreyfus’s reconviction that year was beyond Debussy’s comprehension, and was implicitly ridiculed by him four years later in an article in Gil Blas that referred to Grieg as ‘that Scandinavian composer who was so disagreeable to France at the time of the Affair’.1 It’s true that Debussy, as we shall see, was from time to time involved in theatrical projects with a definably political angle. But none of this came to fruition, while his own finished work is entirely innocent of such nuances until, near the end of his life, war created mental conditions that could even blow L’Art pour l’art – Art for Art’s Sake – out of its still waters.

Debussy had recently formed a close friendship with a young poet by the name of Pierre Louÿs, already, at the age of twenty-three, the author of two collections of erotic verse – one entitled Astarte (published in the literary review, La Conque), and a much longer set, Chansons de Bilitis, which was published in book form in 1894 – as well as a number of short stories. The two had met, probably, at Mallarmé’s Tuesdays and at Bailly’s bookshop, in 1891 or early 1892, but they became intimate only in the latter part of 1893, when they even considered sharing an apartment together. When Debussy went to Brussels in December that year to talk to Ysaÿe about his string quartet, then on to Ghent to get Maeterlinck’s permission to set Pelléas, Louÿs went with him and, according to his own report, did most of the talking with Maeterlinck while Debussy hung back like a shy wallflower (though Debussy’s own account has Maeterlinck behaving like ‘a young girl to whom one is presenting a future husband’).2 Three months later, at the end of the Thérèse Roger affair, Louÿs wrote to Mme de Saint-Marceaux defending Debussy’s behaviour, admittedly on the somewhat fragile grounds that he could hardly be expected to send packing a mistress of two years’ standing as if she were a chambermaid. ‘As for the rumours that have reached you about his former life,’ he added, ‘I go bail that they are monstrous calumnies … I know personally that Debussy is incapable of having lived the way they say.’3

It was a brave guarantee on five or six months’ friendship, and even a somewhat risky stance in the light of Louÿs’s own candid attitude to sexual mores. But he and Debussy had rapidly become close, on the basis of shared tastes in art and poetry, in Symbolist theatre and literature, in the music of Wagner (absolute in Louÿs’s case, guarded in Debussy’s), and not least in women as the object of sensual fantasy and self-indulgence. When Louÿs took up André Gide’s suggestion in 1894 that he follow him to Algeria to indulge his taste for underage girls (Gide’s own preference being for boys), he invited Debussy to join him, with the encouragement of a graphic description of his sixteen-year-old mistress ‘who speaks French so well that, at a moment I can’t specify without indecency, she let go the following affirmation: “Tarrarraboum!! ça y est!!”’, adding, ‘It’s hot, the light is stunning, and the women are all like Bilitis, at least the little girls.’4 (Debussy declined the invitation.) Louÿs was an unashamed erotomane, addicted to nude photography (including of Gaby Dupont), and to literary pornography, of which the Chansons de Bilitis are a stylish example. But he was also genuinely erudite, a classical linguist and a sufficiently knowledgeable student of Greek and Roman literature to be able to fabricate plausible imitations that, at least in ‘translation’, could for a time pass as authentic. Above all, he was well off, and a generous source of funds to his perennially insolvent composer friend.

Debussy had been working on a set of piano pieces that to some extent embodied the harmonic language he had perfected in Pelléas. There were two slow pieces – a very slow waltz and a sarabande – and a quick bravura piece based, like ‘La Belle au bois dormant’, on the children’s song ‘Nous n’irons plus au bois’. He called them Images, as we know from a letter of December 1894 to Lerolle about the pieces’ dedication to his daughter Yvonne.5 But there is not much evidence in the music of any pictorial association. It was probably the first one, headed simply Lent, mélancolique et doux, that Debussy described to Lerolle as ‘a waltz for the use of people who only like such things in a good armchair’.6 The last piece, also untitled, bears the sardonic epigraph: ‘Some aspects of “Nous n’irons plus au bois” because the weather is unbearable.’ The middle piece is labelled ‘In the tempo of a “Sarabande”, that is, with a grave and slow elegance, even a bit old-portrait, souvenir of the Louvre, etc.’ It was the only one of the three to be published in Debussy’s lifetime; it was printed by Le Grand journal du lundi in February 1896, and it then reappeared in 1901, slightly revised, as the central panel of the cycle Pour le piano.

As a pianist, Debussy must have known his instrument inside out, and as a keen improviser, he must have explored its unique properties of sound production, sound combination, and resonance as a musical recreation. But these Images oubliées (Forgotten Images), as they were called when they were at last published as a set in 1977, are generally vague about such questions beyond the actual notes and basic indications of touch and phrasing. Above all, pedal markings are completely absent, even though some details cannot be performed as written without the sustaining pedal. This is an issue that will raise its head especially in connection with Debussy’s great piano works of the 1900s, but it’s worth mentioning here – since it relates to his harmonic idiom of the 1890s – that these slow sequences of sevenths, ninths and elevenths, and the strings of parallel sixth chords and triads, all demand refined pedalling to avoid a total fog of dissonant harmony while achieving what Marguerite Long called the ‘intense legato’ of Debussy’s own playing.7 In this respect the first two Images are to some extent versions of each other. They explore the same or very similar chords. The third piece is crisper and brighter, like some Couperin harpsichord piece for the modern age. At one point the keyboard imitates ‘harps pretending to be peacocks spreading their tails, or the peacocks imitate the harps (as you like it!) and the sky again becomes favourable to bright clothing’. But this is little more than a study for a later piece whose sky is duller and whose clothing is tighter, Jardins sous la pluie, the last of the Estampes of 1902. It’s easy to see why the pernickety Debussy preferred not to publish the outer two Images, though he announced all three on publishing the sarabande. He knew he could do them better. The sarabande was, almost, done.

Work on these somewhat inchoate pieces, as well as on the violin nocturnes for Ysaÿe, seems to have been interrupted spasmodically by a project for a collaboration with Louÿs on a Christmas opera called Cendrelune, about a little girl who longs to wander in the forest but is kept locked up by her stepmother, until one day she is enticed away by a group of enchanted girls who tell her she is the daughter of the Dame Verte (the Green Lady) and that she will be able to find her mother if she goes of her own free will. A pair of Catholic saints tries to prevent Cendrelune from yielding to the temptation of the pagan greenery, but all in vain. Debussy was quite drawn to this hotchpotch, but Louÿs, after various modifications, turned against it. ‘One might make a book of it,’ he wrote, perhaps worried by the absence of male characters, ‘but to put it on the stage in 1895, there is no way.’ ‘Write Cendrelune yourself,’ he suggested a week or two later.8 He himself had other proposals. There was a ballet called Daphnis et Khloé, and a ballet or pantomime based on Louÿs’s novel Aphrodite (some elements of which might have been problematic even for the Parisian stage). Somewhat later Debussy himself (or, as he put it, ‘that little neurasthenic Mélisande’) had the idea of a symphonic suite based on Louÿs’s story Les Aventures du roi Pausole. So far as anyone knows, not one of these projects resulted in a note of music, even though Debussy pestered Louÿs about Cendrelune and claimed at least once to be working on Daphnis.9

A setting of Louÿs’s partial (and very free) translation of Rossetti’s Willowwood was slightly, but only slightly, more fortunate. In October 1896, Debussy wrote to Ysaÿe refusing him permission to perform excerpts from Pelléas, and offering instead ‘a thing I’ve done on a poem of D. G. Rossetti: La Saulaie. Note that this is very important and written according to my latest experiments in musical chemistry.’10 In fact at this stage Debussy had probably not so much as uncorked a single chemical, but three years later La Saulaie was again in hand, and this time, in the winter of 1900, a few sketches were made. As usual with Debussy, they are mere outlines, fragmentary, hard to read, and ambiguously notated, which has not prevented the resourceful Professor Orledge from producing a performable, genuinely beautiful if not chemically all that Debussyan score of the ‘entire’ twelve-minute cantata. Only one work of Louÿs’s, his Chansons de Bilitis, produced anything substantial from Debussy’s own hand, and this inspired two works, a short cycle of three songs, and a somewhat slender, but finished, score of incidental music for a tableaux vivants presentation of another twelve of the poems.

The Trois Chansons de Bilitis was the first vocal work he had seriously embarked on since completing Pelléas, and the songs are closely allied to that opera in a number of ways. In particular the writing for voice is almost entirely close set, with much note repetition and stepwise motion, frequent implied silences and an almost complete absence of melisma (more than one note to the syllable). As usual with Debussy, the prevailing dynamic level is soft or very soft, so that the few moments of mezzo-forte or louder – none in the first song, a couple in the second, one in the third – draw attention to themselves, and invite interpretation. The girl of the poems, supposedly a young lesbian courtesan on the island of Cyprus in the time of Sappho (seventh–sixth century bc), recounts a series of brief episodes in her life, mostly of an intimate, sometimes overtly sexual character, involving men, women and even children. In the very first poem, under the cheery subheading ‘Bucoliques en Pamphylie’, she climbs naked into a tree and masturbates against a high branch. When he first published the collection, in 1894, Louÿs passed the poems off as prose translations from authentic Greek originals, supposedly found inscribed on the walls of a tomb in Cyprus, but as with most fakes, it is hard subsequently to believe that experts were genuinely taken in. The poems are so obviously the product of a voyeuristic male fantasy on the decadent fringe of 1890s Art Nouveau. Fortunately, this has little bearing on their quality, and even less on their musical potential, which Debussy was able to show was considerable.

Perhaps pointedly, he selected three poems from the first of Louÿs’s three parts, in which, in her native Pamphylia, Bilitis is loved by men. In the first song, ‘La Flûte de Pan’ (in Louÿs simply ‘La Flûte’, but Debussy was still in the grip of his faune), he gives her a syrinx, a reed pipe, which they play in turns, one of Louÿs’s more discreet metaphors for the sexual act. (‘My mother will never believe’, the poem ends, ‘that I stayed so long looking for my lost girdle.’) The song itself is a masterpiece of antique simplicity. The singer chants softly within a range no greater than would be required to recite the poem with a moderate degree of emotion, while the piano oscillates between plain triadic chords coloured by gentle dissonances of the usual Debussyan kind, and modest filigree patterns with the occasional flourish, evidently meant as flute music. The whole effect is rather like the soundtrack of a Nouvelle Vague film: cool and matter of fact about an extremely delicate sensual experience, told without excess or insistence, but with a quiet intensity that verges on the private.

In the second song, Mallarmé’s flute is replaced by Maeterlinck’s hair, ‘La Chevelure’. In fact, the poem, which was added later by Louÿs and sent to Debussy when as yet unpublished, reads like a conscious plagiarism of the tower scene in Pelléas. He tells her that he dreamed he had her hair round his neck and over his chest, that they were forever locked together, mouth to mouth, etc., etc. Debussy’s setting is more restrained than the opera scene, though it does rise to the cycle’s only fortissimo, at the image of their members becoming ‘so intermingled that I became you’ and ‘you entered into me like my dream’. But the real tone of the song is set by the surreal colouring of the figures that represent her hair, floating and waving across the ‘screen’, to harmonies that express the curious mixture of intensity and inconsequentiality we associate with dreams. Here, as before, Debussy achieves a daring simplicity of musical effect: an almost comatose quality of vocal narrative against gently nagging ostinato repetitions in the accompaniment.

Finally, in ‘Le Tombeau des Naïades’, we witness the death of heterosexual love, though this is not apparent from the song, which concentrates on the Schubertian image of the winter ice as a sign of the end of a particular affair. In the book, the poem is the last in the first part, before Bilitis takes ship to Mytilene and discovers herself as a lover of women. Debussy captures the poem’s haunting picture of the girl walking slowly through the frosted landscape, her hair glistening with icicles, the whole image encapsulated in the rolling semiquaver figures in the piano right hand, and the melancholy thumb melodies, right hand and left hand. These middle voices are, literally, a Debussy fingerprint; they are important in ‘La Chevelure’ as well. As for the general style of the three songs, it’s perhaps worth noting that Debussy had just been orchestrating two of his friend Erik Satie’s Gymnopédies, and may have been infected with the studied emotional neutrality of Satie’s repetitiousness. The Chansons de Bilitis are far from neutral, but by feeding that base idea back into the intensely experienced but untheatrical sensuality of Louÿs’s prose poems, they achieve something new and individual, a kind of concentrated motion that constantly suggests cinematic parallels, whether the Nouvelle Vague of Truffaut or the surrealism of Buñuel or Robbe-Grillet.

Soon after completing the Bilitis songs, Debussy performed them, with an unnamed singer, at a soirée given by a certain Mlle Worms de Romilly, after which ‘an old Argentinian gentleman who had for some time been containing himself with difficulty … marched up and down the room, shouting, “No, no, that’s not music! How do people come to write stuff like that?”’

The first public performance of the songs was given by the young Blanche Marot in March 1900, who later reported that Debussy had called on her mother and enquired:

‘Tell me, Madame, your daughter is not yet twenty? Good. It’s very important, because if she understands the second song, “La Chevelure”, she won’t sing it in the right way: she mustn’t grasp the true brazenness of Bilitis’s language.’ … My mother set Debussy’s anxieties at rest and everything went splendidly.12

Roger Nichols, the editor of Debussy Remembered, elsewhere casts doubt on this story, pointing out that Blanche was already the mistress of Debussy’s publisher Georges Hartmann, a fact of which the composer must have been aware.

A solution to the problem could be that Debussy went to see her mother some time before the performance and that she and Hartmann formed their liaison some time in the interim. But for our understanding of Debussy’s mentality, the crucial point is that she was obviously able to sing the song as though not understanding it (otherwise he would surely have cancelled the concert).13

The issue is one of ‘knowingness’, as opposed to the native, unselfconscious sensuality of Louÿs’s poems and Debussy’s songs. ‘There is, therefore, no requirement’, Nichols adds, ‘to label the songs “for virgins only”.’

The Chansons de Bilitis are almost the only solid product of the two years following the completion of Pelléas et Mélisande, and strictly speaking only the first two songs, since ‘Le Tombeau des Naïades’ was not composed till the spring of 1898. Debussy had spent so much time, or at least mental energy, on abortive plans such as the various Louÿs projects, as well as all sorts of other theatrical schemes, some of them – like the idea for an opera on Verlaine’s verse comedy Les uns et les autres – never actually begun, others, like René Peter’s Tragédie de la mort and Mme Jean-Louis Forain’s Rosicrucian pantomime Le Chevalier d’or, at least tinkered with if not seriously worked on, that his focus on new conceptions of his own seems to have been severely undermined.14 But there were personal and circumstantial reasons as well. There was Pelléas still lying on his desk, unperformed and with no immediate prospect of performance. He had no money and had frequently to go cap in hand to friends to whom he was already in debt. He resumed the Wagner sessions that had been curtailed as a result of the Thérèse Roger affair, no longer chez Mme Escudier, but now in the salon of a certain Mme Godard-Decrais. He took on the direction of a family choir started in 1894 by Arthur Fontaine’s brother Lucien, and was still directing it four years later when he composed for it beautiful four-part settings of a pair of poems (‘Dieu! Qu’il a fait bon regarder!’ and ‘Yver, vous n’estes qu’un villain’) by the fifteenth-century Duke of Orleans.15 He took the occasional pupil, including Mlle Worms de Romilly, who sang in the choir and persuaded him to give her singing and, more believably, piano lessons. But over and above all these distractions, his love life once again, early in 1897, threatened to come between him and the calm contemplation he desperately needed for his creative work.

The affair with Gaby had survived the Thérèse engagement, and there had been other dangerous episodes, such as when Debussy had convinced himself that he was in love with Catherine Stevens, the pretty daughter of the Belgian painter Alfred Stevens, and had proposed to her sometime early in 1896. ‘He gave me proof’, she told René Peter, ‘of such disinterested love during the worst moments of a family crisis, he whom people have called “grasping”, and he was so intelligent and he played me Pelléas, which he was working on. I would have married him, despite everything that was being said about him at the time … if I hadn’t met Henry!’16

Towards the end of that same year there had been another fling of some kind, possibly with Alice Peter, the estranged wife of René’s brother. Then, at some point in January 1897, Gaby had found a letter in his pocket that, as he wrote to Louÿs, ‘left no doubt about the development, even somewhat advanced, of a love affair with all the most romantic details needed to move the hardest of hearts’.

Thereupon! … drama … tears … a real revolver and the Petit Journal to tell the tale … Ah! my dear fellow, I would have needed you to be there to help me recognise myself in that bad literature. All this is barbarous, pointless, and changes absolutely nothing; you can’t erase a mouth’s kisses or a body’s caresses with an india rubber. That would be a nice invention, a Rubber to rub out Adultery.17

Somehow it was all patched up without such help, and Gaby was probably still around when, just over a year later, Debussy met a couturier’s model by the name of Rosalie Texier, an attractive, elegant blonde from a bourgeois background in the Yonne. Lilly Texier came on the scene just when, at the age of 35, he was in search, not of sexual adventure, but of some kind of stability that he could regard as permanent. ‘I’ve been very unhappy since you left,’ he wrote to Louÿs (who had gone to Cairo),

unhappy in the most passionate sense, and have wept a great deal, since that simple act, in which all humanity meets, was the one thing that remained for me in so much anguish … What’s more, I’ve never been able to get anything done whenever things have happened in my life; I think that’s what makes for the superiority of memory: you can draw valuable emotions from it, but people who weep while writing masterpieces are incorrigible humbugs.18

Perhaps all the same he needed to feel passionately attached to the stable object. Peter’s later reminiscence that at this time, ‘imperceptibly, the preoccupation with women was relegated to a secondary tier in his mind’19 has to be moderated in the light of his eventually overwhelming physical passion for Lilly, which nevertheless did not preclude at least one infidelity that nearly brought the relationship to an end, and which, after four years of marriage, faded as quickly as it had blossomed, with results even more drastic than in the Gaby case. One might suggest that, as for other great artists, Debussy’s work was always paramount but that it created a parallel need for emotional and sexual expression, a sort of reassurance on the human level. Creative work demands dullness in life outside, but life outside is not always willing to stay dull. Hence Debussy’s penetrating remark about the superiority of memory, through which creativity can be fed by excitement without being blocked by it.

Another remark of Peter’s, that Lilly ‘was remote from the dreamer Bilitis whose songs Claude was then singing’, is beyond denial.20 One wouldn’t imagine this daughter of a railway telegraph manager in the provincial town of Montereau who, according to Marcel Dietschy, ‘prepared tea surpassingly well’, echoing Bilitis’s ‘I enjoy life only when naked. My lover, take me as I am … Take me as my mother made me in a night of love long past.’21 But Debussy could accommodate both types. A year after marrying Lilly in October 1899 he agreed to write incidental music for a stage performance of a selection of the Bilitis poems, consisting of readings illustrated by tableaux vivants, and accompanied by what Louÿs modestly described as ‘eight pages of violins, silences and brassy chords that give what one might call “an impression of art” at the Variétés’.22 Debussy’s eventual music for the private performance at the premises of Le Journal on 7 February 1901 was hardly more substantial than Louÿs’s prediction – not more than ten minutes of music as interludes accompanying the tableaux for a mere eleven or twelve poems. But since the composer was probably playing the celesta (with a pair of flutes and a pair of harps), he will have participated in the preparations as well as the performance itself, which included enough nudity and ‘erotic contact between women’ to provoke a threat of legal action by a certain Senator Béranger, a threat Louÿs seems simply to have ignored. As for Debussy’s music, it has silences but no violins and no brassy chords. All is refined and delicately sensuous, a fragmentary afternoon of another faune. Unfortunately the celesta part disappeared, perhaps into Debussy’s pocket, and has had to be reconstructed for modern performances.23

*

For the past seven or eight years, Debussy had had on his desk, or in his mind, or at least in his letters, a conception of an orchestral suite of an essentially nocturnal character, a conception that may or may not have remained broadly the same throughout those years, even if the actual notes on paper – such as they were – altered over time.

There is actually no hard evidence that he ever composed a note of the Scènes au crepuscule, which he told André Poniatowski in September 1892 were ‘practically finished’ and five months later had been ‘well reworked’.24 As for the Nocturnes for violin and orchestra that he described in some detail to Ysaÿe in September 1894 and reminded him about two years later, there were, as we saw, enough sketches for a speculative performing version to be made, but no resemblance in them to any of the three Nocturnes that Debussy eventually completed at the end of 1899. This doesn’t mean, though, that there was never a connection. During 1897 and 1898 Debussy kept up a running commentary with Georges Hartmann, his new publisher, about progress on the Nocturnes. Finally, in June 1898, he declared them finished, but then three weeks later, typically, modified the announcement: ‘As I told you, I’ve finished the Nocturnes, but the orchestration isn’t yet done, which shouldn’t hold things up too long …’25 Alas, the delay turned into a worry. ‘You will hear and possess the three Nocturnes,’ he assured Hartmann in September, adding that ‘the three of them have given me more grief than the five acts of Pelléas’.26 But at the New Year he was still sitting on them, and finally in April he admitted that the first piece was not good enough and he was starting it all over again. Worse, perhaps, was to follow. On 3 July 1899 he wrote again to his long-suffering publisher: ‘You ask me what I’m doing! … I can scarcely do more than tell you what I shall be doing … I shall finish La Saulaie, then three other Nocturnes and the Nuits blanches.’27

Three ‘other’ (‘autres’) Nocturnes, meaning that he had discarded all three of the originals? This is hard to believe, since less than three months later he was again announcing their completion, if not happily. ‘If I haven’t handed the Nocturnes over for engraving,’ he now told Hartmann, ‘it’s because of my terrible mania [fussiness], of which I am moreover the first victim; so I was leaving them for a while and working on La Saulaie, another nightmare.’28 Not till January 1900 did he finally hand the Nocturnes over for performance and publication.

Poor Hartmann! Already bankrupted once (in 1891) and forbidden to reincorporate as a publisher, he had nevertheless done so under the name Eugène Fromont, and had bought the Prélude à l’après-midi d’un faune in October 1894, before it had even been performed. Now he was embroiled with that work’s composer at his most evasive. The Nocturnes might have clinched their relationship; in fact it was to be his epitaph, since just over three months after finally receiving the manuscript he died suddenly at the age of fifty-six. He saw the score but may well have been unsure of its effect, so strange and original must it have seemed to eyes and ears used to the appearance and sound of the music of Massenet or Saint-Saëns.

Whether or not any musical connection existed between the different stages, it seems likely that there was a persistent background idea that survived from first to last. The word ‘scenes’ in connection with ‘twilight’ in the original title is a visual, not a theatrical, trope, and when he first wrote to Ysaÿe he described the violin and orchestra piece he was writing as ‘a research into the various arrangements that a single colour can yield, like for example what, in painting, would be a study in grey’.29 As for the eventual title, Nocturnes, this clearly has nothing to do with the piano music of Chopin or Fauré, but is a borrowing from the American painter James McNeill Whistler, whose own Nocturnes – a title he himself took from music – are studies in particular named colours: blue and silver, black and gold, blue and green, etc., and who had also painted a portrait (of his mother) called Arrangement in Grey and Black (No. 1). These are all figurative paintings, not abstracts, but the artist seems at least to want us to believe that the subject matter is merely a peg on which to hang the colour scheme, since he usually places the subject after the colours in the title.

Debussy’s own titles are precise but generic: Nuages (‘Clouds’), Fêtes (‘Festivities’) and Sirènes (‘Sirens’). The clouds idea looks as if it might be the study in grey he mentioned to Ysaÿe, except that the first piece was the one he discarded, so either the order was changed or Nuages as we have it was based on the same general idea as before but recomposed. These are not quite his first instrumental pieces based on visual images, but they are the first for orchestra – apart from Le Printemps, with its wordless chorus and lost or never executed orchestration – and the first in which the visual idea is worked into the bones of the music, rather than being merely a picturesque programme, as in En bateau in the Petite Suite or Clair de lune in the Suite bergamasque. It was Le Printemps that had prompted the deathless insult ‘vague impressionism’ from the Conservatoire assessors, and these Nocturnes have, in their turn, been taken as a prime example of Debussy’s Impressionism by an age that did not see it as an insult. The original application of the term had, of course, nothing to do with the techniques of the Impressionist painters, everything to do with the effect of their work on the viewer. Monet’s Impression, Sunrise and his Boulevard des Capucines gave a vague idea of the subjects, as long as you were willing to accept that the black splodges on the water were boats (what else could they be?) and that what M. Vincent called the ‘innumerable black tongue-lickings’ at the near end of the boulevard were people. In the same way, Debussy’s Nocturnes are ‘impressionistic’ because they sound blurred at the edges, just as the paintings look rather blurred. You don’t expect to identify objects in quite the way you do with a picture; after all, this is music. But somehow you feel the same way about Debussy as you do about Monet, and the descriptive term puts that feeling into words.

It is easy to make fun of this kind of thing, but as so often with Vox Pop there is a truth lurking in the pigeonholes. From a technical point of view Debussy is about as far from Monet or Pissarro or Sisley as it would be possible to get. The idea behind so-called Impressionist painting was to capture the image very quickly, painting typically in the open air in the presence of the subject, and the techniques involved related to physical matters such as the changing light and the drying speed of paint. Such issues have no meaning to a composer, working indoors in a medium that is not, like the painter’s canvas, the eventual object, but only a set of instructions for its production (that is, performance). This is an essentially meticulous, partly abstract process involving the thinking through of an aural idea that may or may not have been prompted – still in the composer’s mind – by a picture or a story or even a concept, like the one behind Liszt’s From the Cradle to the Grave or Strauss’s Ein Heldenleben. As composers go, Debussy was generally slow and painstaking, not because he couldn’t write quickly, but because he wouldn’t compromise with the smallest detail. With Monet, one supposes, compromise was avoided by preserving only what was deemed satisfactory. As a ‘quick’ painter, you could in theory cover many canvases in a single day (save for the considerable cost of canvas and paint, a problem composers are largely spared), and you could discard or paint over most or all of them. The idea of a composer writing, say, half a dozen symphonic movements in a day, then binning five of them, is absurd, surely, even to those who have never thought what might be involved in composing and writing out five minutes of music for a large orchestra.

So where shall we find the truth? Debussy, as we’ve seen many times, was in revolt against the musical establishment, with its rule-books and its thou-shalts and thou-shalt-nots, and he had been spending most of his social time in the company of outcasts from other academies: Symbolists, Impressionists, café artists and performers, theatre people of a more or less alternative kind. Since his student days he had taken an enthusiastic interest in art and theatre, more than in new music, which he would have agreed with Chabrier was self-satisfied and stick-in-the-mud. Then in the mid-1890s he found himself accepted, even lionised, by a circle that, while bourgeois in its lifestyle, was artistically open and progressive. Chausson and Lerolle were that comparative rarity, artists with money. In their houses, and that of their brother-in-law Arthur Fontaine, you were surrounded by paintings and murals and drawings by all the new artists, and you met the artists themselves and perhaps discussed their work. There was a saturation of artistic talk, a sharing of tastes and distastes, a continuous contact with the maker and the made and, surely, a swapping of notes about aims and processes.

Take drawing, for instance. At the Académie des Beaux-Arts the ability to draw well was still regarded as a sine qua non for a properly equipped artist. But drawing implies outlines, whereas in Monet’s and Pissarro’s pictures form was defined not by drawn lines but by areas of colour. In a famous essay on Impressionism, the poet and critic Jules Laforgue traced the preoccupation with drawing to the idea that our sense of form is derived from our sense of touch, and has no direct connection with the eye.30 A musician who is likewise eager to break down academic prejudices might suggest a parallel with form defined by harmony or, worse, presented as a set of schemata to which a properly trained composer is expected to adhere. Of course, this isn’t really a very good parallel. Drawing is a fundamental discipline that only needs to know its place. Harmony and sonata form might seem to be advanced, if ossified, studies that only academy students bother with. Singing might be closer to drawing as a basic requirement, but singing and melody were certainly not on Debussy’s hate list. For him it looks as if the example of Impressionism breaking the primary rules of draughtsmanship was suggestive as a general model for the diffusion of boundaries. But there was also a sense that, in blurring the detail of the subject and concentrating it into a set of motifs, the painters were making discoveries about its soul that were obscured in the more painstakingly accurate Academy landscapes. And this was a terrain on which the composer could meet them as an equal. Wassily Kandinsky, a painter with musical credentials, numbered Debussy among the modern composers who ‘reproduce spiritual impressions which they often borrow from nature and transform into spiritual images of a purely musical form’.31 But in implying that this was a property unavailable to visual art, he was being too modest. This is exactly what (apart from the kind of form) one finds in the greatest work of Monet, Cézanne, Gauguin, and Kandinsky himself, and it is this shared ‘interior content’ (Kandinsky again) that justifies the shared terminology at a level somewhat deeper than the mere blurring of objects and textures.

Exactly why Debussy’s drifting chords and fragmentary melodies in Nuages suggest clouds moving slowly across a sky of melancholy grey with occasional breaks is a question in perceptual psychology that is beyond my competence. Not many listeners would dispute that they do, though it can be argued that the trigger word is needed for the ear to hit the right image (as opposed, say, to water flowing past water lilies, or even the hallucinations of a troubled mind). To call this piece Fêtes would be patently absurd; to include it as a movement in a symphonic work called La Mer, much less so. But the choice of subject is itself significant. Clouds not only drift; like water they imperceptibly change shape, so that the skyscape is not like a landscape seen from a moving train, which changes only because of the changed perspective. Debussy’s chord sequences seem to flow into one another, and although this is obviously an illusion (not least because precise repetition is an aspect of musical form), it’s the illusion that counts.

More even than most orchestral works, Nuages is formed as much as anything by its scoring. The orchestra is not blended but layered, a feature of clouds more than water or, perhaps, hallucinations, though the opening chord sequence, for clarinets and bassoons, has often been identified with a sequence in Musorgsky’s ‘The busy, noisy day is over’, the third song in his Sunless cycle, where the poet is ‘inhaling the poison of passionate spring dreams’. There are other reasons for accepting this influence. In the summer of 1898 while at work on the Nocturnes, Debussy wrote two songs towards an intended cycle of five called Nuits blanches, to vers libre poems of his own, songs that resemble Sunless so closely in general texture and atmosphere as to virtually rule out coincidence.32 In Nuages the chords outline the main theme, but their scoring is almost always discriminatory: woodwind for the rather severe, open-textured version that starts the work off, divided strings for the altogether more sumptuous extended version that follows. Wind and strings hardly ever play these sequences together. In the same way, the lapidary cor anglais theme (bar 5, etc.) is never played by a stringed instrument. It hangs in the texture like some motionless object, always the same and always at the same pitch. To identify this motif with some actual conceivable object in the sky (a hovering kestrel, a drone, a flying saucer) would be infantile. It’s at this point that Kandinsky’s idea of interior content comes into play.

Nuages is motion without change, surface change without existential change. Things drift but do not mingle. ‘Plus ça change,’ the French say, ‘plus c’est la même chose’; the more it changes, the more it’s the same. This is a profoundly French, un-Germanic, un-British sentiment. All progress is apparent, not real. Just when we think we’ve moved on, we look up, and there is the same cor anglais theme in the same place, unaltered. Even when what seems to be a real change takes place and flute and harp play a much brighter, more hopeful theme in a pentatonic F sharp major, the cor anglais quickly reminds us that hope – like despair – is an illusion of temporality. By now Debussy’s harmonic language has developed to the point where chords no longer carry implications, but generate colours prismatically and in a series of fluctuating patterns. The beauty of this music lies not in dramatic development or the strong contrasting of themes, but in the refined juxtaposition of colours, melodic, harmonic and instrumental. Debussy’s ear for chord voicing – the exact placing of each note in terms of register, balance and timbre – was equalled in his lifetime only perhaps by Wagner, whose orchestration (in Parsifal) Debussy once memorably described as ‘seeming to be lit from behind’, though he had also on an earlier occasion dismissed Wagner’s scoring as ‘a kind of multi-coloured cement, almost uniformly spread, in which [M. Croche] told me he could no longer distinguish the sound of a violin from that of a trombone’.33 His own palette is unique, but a Russian influence might be detected in the desire to separate rather than blend the orchestral sonorities, a preference that goes back to Glinka and survives in Rimsky-Korsakov and to some extent in Tchaikovsky.

This separation is still more apparent in the brilliant second nocturne, Fêtes. It is even visible, and perhaps even more Russian. For much of the time the orchestra is laid out in blocks, or choruses, clearly grouped on the page, which has an antiphonal appearance not unlike certain Tchaikovsky scores, for instance, the Fourth Symphony, a work Debussy knew well. The clear spacing is the source of the music’s transparency, and perhaps also of that sense of illumination, of lights moving from hand to hand, which Debussy implied in a description to a lawyer friend, Paul Poujaud: ‘a memory’, he called it, ‘of popular festivities in the Bois de Boulogne in the old days, lit and invaded by the crowd’.34 The festivities evidently included dancing, of a somewhat bucolic variety. The programme note for the first performance, probably by Debussy, referred to

the movement, the dancing rhythm of the atmosphere with abrupt flashes of light, together with the intervention of a procession (a dazzling and chimerical vision) passing through the celebration, mingling with it: but the event remains intact, and there is still the festival and its mélange of music, of luminous dust participating in a total rhythm.35

There is in this account a fusion of visual inspiration and something more generalised and abstract. The visual parallel might be with a canvas such as Camille Pissarro’s Boulevard Montmartre at Night, painted at more or less the same time (1897). Very obviously done at speed, the painting is a partially abstracted vision of Paris metropolitan life on a wet evening, with a procession of carriage lights down the boulevard, and a chaos of restaurant and hotel lights down either side, the whole thing executed with swift brush-strokes and a deliberate approximation of detail. There is no question of approximation in Debussy’s wonderfully calculated orchestration, and instead the sense of an imprecise, flickering image comes from music’s natural tendency to generalise, so that the festivities might be what Debussy tells us they are, or they might be Pissarro’s, or they might be any one of a number of other possibilities.

A clue to this lies in Debussy’s title, which denotes a concept rather than a specific event or locale. In subsequent works he would sometimes be precise (La Soirée dans Grenade, Les Collines d’Anacapri), sometimes conceptual (Mouvement, Voiles). In odd cases he would change from one to the other. The first movement of La Mer was originally ‘Mer belle aux îles Sanguinaires’, but ended up as ‘De l’aube à midi sur la mer’. But whatever the title, there is always – because music is the way it is – a leaning inwards beyond the implications of a particular event or place or person. Just as festivities leave a certain platonic impression on our mental retina, aside from the specifics of the event itself, so music seems able to go directly to that retina without passing through the event, which may be attached to it afterwards as a kind of excuse or explanation. We know what the Symphonie fantastique is about, but we wouldn’t know if Berlioz hadn’t told us, though we might sense the presence of something beyond the mere notes. The Russian art historian Vladimir Stasov considered that all music was programme music, and he had explanations for the ‘Eroica’ Symphony and even for Schumann’s Piano Quintet, neither of which had the remotest authority from the composer. In his view, the finale of the ‘Eroica’ was itself a ‘fête’, ‘a crowd of people, a popular festival, in which diverse groups of people succeed one another: now ordinary people, now soldiers, now women, now children – and all against the background of some rural landscape’.36 Nothing in this description clashes with Debussy’s Fêtes, not even the soldiers, whom he represents with ‘a trio of muted trumpets [that] correspond to the memory of the music of the Republican Guard sounding the retreat’.37

At the time, Fêtes probably seemed an odd sort of middle panel for the triptych, and certainly Sirènes was an unusual finale, not least because (like Le Printemps) it required a female chorus singing without text. Probably Debussy did not think of the Nocturnes as inseparable, since when Camille Chevillard conducted the first performance in December 1900 he left Sirénes out, presumably because of difficulties with the female chorus; and as we’ve seen in the cases of the Fantaisie and Pelléas, Debussy was inclined to be awkward about the excerpting of works he thought of as integrated.38 All the same, Sirènes does balance Nuages in some ways. Between them, they provide a framing character of the unworldly, music that is slow moving, mysterious and above all intangible, offset by the brilliantly lit festivities of the centrepiece. In other ways, these outer movements could hardly be more different. Where Nuages presents a fairly straightforward musical simulacrum of natural phenomena in terms of clearly defined instrumental blocks, Sirènes is complicated by a hidden human narrative, not specified by the composer or the music, but implied by the title.

In Greek mythology the Sirens are a pair (or more) of winged maidens whose song is of such irresistible beauty that it lures mariners to shipwreck on the rocks. Odysseus, their most famous target, had himself roped to the mast of his ship in order to hear the song without being enticed to his death. Debussy’s sirens are of course the female choir, and if one imagines what a siren song might be like, the chances are one will be in the same mental parish as Debussy, even if one has never heard Sirènes. There are precedents. For instance, in Rimsky-Korsakov’s Sadko, first performed in St Petersburg in January 1898, the hero is lured (not fatally) to the ocean depths by a wordless chorus of sea-maidens that would not be wholly out of place in Debussy’s score. In the final act of Wagner’s Götterdämmerung the Rhinemaidens extend their lament for the gold in a wordless string of Weia-la-las. This is not to imply anything unduly derivative about Debussy’s siren song; nevertheless the mere presence of a wordless female chorus in the finale of an orchestral work of this kind is perhaps more original than the actual music he gives them to sing.

Debussy regarded Sirènes as a work about the sea, and the siren song is plainly heard as an emanation of the windy-watery music played by the orchestra, to the point where it is sometimes hard to distinguish the one from the other. The interior of this orchestral writing is more complex and varied than in either of the other nocturnes; in certain respects it looks forward tentatively to the intricacies of La Mer and even his Diaghilev ballet, Jeux. But once again the appearance of the music on the page reveals a layering similar to that in Nuages, if more fluid and with more sharing of material, as befits the action of wind on water. Occasionally the layering makes for a visual effect that already, amazingly, hints at the Stravinsky of The Rite of Spring, the only thing, perhaps, the two works have in common. Also sometimes visible on the page, even to the non-musician, is the unusual degree of repetition of two-bar phrases, conceivably a weakness, or possibly a deliberate factor in the hypnotic quality of the siren song.

What sets Sirènes firmly apart from its two companions is the character of its melodic lines. In Nuages melodies are fragmentary; texture is all. Fêtes is a (somewhat erratic) tarantella made up of fast dance tunes and brass fanfares. But Sirènes is all about ornamented melody, arabesque and what Nectoux calls fouetté, the ‘whipped’ flourishes that start in the clarinet in bar 1 and decorate the texture throughout.39 These flourishes set up regular patterns against the slow weave of the melodies, which sometimes take on a positively oriental colour, because of this or that interval or this or that arabesque pattern. One recalls that the curling arabesque line was an aspect of the Art Nouveau that it derived from (among other things) Japanese prints. Whether this makes Sirènes an example of musical Art Nouveau, as Nectoux suggests, is a question that takes us straight back to the whole issue of equivalences that I discussed in relation to Nuages and Impressionism. A Beardsley drawing or a Shekhtel balustrade reveals its debt to the East in a direct imitation or indirect derivation from a visual model, but all a composer can do is translate this mimicry into a metaphor of line, a line that strictly speaking doesn’t exist. Needless to add, if line is a legitimate description of melody, it can be found in every composer of the nineteenth century and before (Debussy heard it in Victoria and Palestrina). So the musical Art Nouveau turns out to be an extremely Vieil Art.

Diverse as they are, these three pieces make up a beautiful and highly original triptych. As with the Prélude à l’après-midi d’un faune, and bearing in mind that Debussy had not yet orchestrated, let alone heard, Pelléas, his refinement of orchestral touch in three such different contexts is little short of astonishing. The voicing and balancing of chords in Nuages, the deftness and athleticism of Fêtes, and the idiomatic instrumental detailing of Sirènes, show a range of mastery in a skill that remains elusive for some composers even after years of experience. And Debussy’s scoring is never merely functional, never simply a question of transferring to the orchestra music conceived at or for the piano. In every case, for instance, the contrast and layering of wind and strings is vital for the articulation of the form, just as the contrasting and sequencing of colour are the main formative elements in an Impressionist painting.

No less significant is the way he uses his harmonic language as part of an overarching design, while mostly avoiding its traditional, textbook procedures. His method actually varies from piece to piece. In Nuages certain chords are effectively motivic; though often dissonant, they no longer remotely call for resolution, but provide a colour motif for the piece as a whole. In Sirènes the harmony is mainly underpinned by tonic–dominant pedals: for instance, F sharp and C sharp in the key of F sharp major in which the piece begins, then moving, more or less arbitrarily, to A plus E in A major at bar 5, and C plus G in C major at bar 8, and so on. These open fifths give stability and depth to the decorative polyphonies of the upper parts, like the deep ocean waters below the turbulent surface. In Fêtes open fifths are a motif, often in rapid parallel sequences, with or without the missing third. All three pieces show that Debussy was not in rebellion against tonality, only against the rules by which it supposedly worked. There is always a key just round the corner, even if it never quite arrives, which is probably one reason why, though it disturbed academic musicians, his music never seems to have created serious problems for audiences.