For Jacques Durand it may have been bad news that, just when Debussy had contracted to write a whole series of piano and orchestral Images, it turned out that he was starting a quite different orchestral work about the sea. But Durand could see only half the picture. Either that summer or very soon afterwards Debussy seems also to have been hatching a completely new set of piano pieces under the title Suite bergamasque.

Exactly what was going on is still by no means entirely clear. Long ago he had written the suite of pieces that is today universally famous under that title, but it had never been published and was still lying idle, along with other short pieces, in his bottom drawer. Meanwhile he had the idea for a set of three more substantial pieces that seemed to justify a title of that kind. One of them, L’Isle joyeuse, was connected in his mind with Watteau’s painting L’Embarquement pour Cythère and perhaps from there with Verlaine and the ‘charmant masques et bergamasques’ of ‘Clair de lune’. Another was actually called Masques. A third piece, dividing these two allegros, was another sarabande. As early as June 1903, Viñes noted in his journal that Debussy had played him a piece called L’Isle joyeuse, presumably an early version of the piece we know; and the following January he played him all three pieces, again presumably versions of the intended Suite bergamasque pieces. How it came about that in 1905 Fromont printed the Suite bergamasque we now know – a completely different work that seems to have no particular call on that title – is a commercial mystery: in other words, no mystery at all. Fromont wanted to bring out some marketable piano music and Debussy, albeit under protest, needed cash. The Viñes pieces stayed separate, and to this day we don’t know absolutely for certain the identity of the sarabande, or even that the music Viñes heard was the music we know under the other two titles. Only the one title, L’Isle joyeuse, is certain; the music not.

How much does all of this matter, beyond the resistible entertainment of watching musicologists count the angels on the points of each others’ needles? The answer is that if it were to lead to the identification of the sarabande it could release musical information about the other two pieces. Roy Howat, himself a brilliant pianist as well as a leading scholar, came to the conclusion that the missing piece was in fact a perfectly well-known work called D’un cahier d’esquisses (‘From a book of sketches’), which was published in Paris illustré in February 1904. This was no idle speculation. The piece is indeed a sarabande of sorts – that is, a slow dance in triple time with an accent on the second beat of the three. Admittedly it is mostly barred in six-eight time, but it could quite reasonably be rebarred in three. It has significant thematic connections with both Masques and L’Isle joyeuse. Its sixth bar picks up a rocking minor-third figure from Masques, while L’Isle joyeuse begins pointedly on the keynote of the sarabande (D flat rewritten as C sharp) and clearly develops rhythmic and melodic ideas from the sarabande material as a whole.

Nobody knows precisely why Debussy decided after all to publish these three pieces separately. He may have become dissatisfied with them as a balanced set, but that can hardly have been for lack of integration between the three, which is a lot greater than between the three Estampes or, for that matter, the three pieces in his next set, the first book of Images. The fact that he allowed D’un cahier d’esquisses to come out in an illustrated magazine suggests, according to Howat, an intention to publicise the forthcoming triptych, rather than any doubts about the piece itself or the suite to which, on this interpretation, it belonged. Unfortunately Debussy then for some, no doubt financial, reason sold the piece to the Brussels branch of Schott & Co., so that when he struck a deal with Durand later that spring only Masques and L’Isle joyeuse remained available for him to publish. Debussy then proceeded to rewrite both pieces; Durand brought them out separately in the autumn, and by then the whole Fêtes galantes connection had leaked away into a new cycle of Verlaine songs under that title, also published by Durand in September.

None of this, of course, prevents the three pieces from being played in sequence, as a coherent set, and when done in this way they convince absolutely. Debussy had been working on his three-movement symphonic piece La Mer since the previous autumn and was clearly thinking in longer spans than usual. Thus Masques, the natural first piece of the suite and a dazzling, bravura dance in its own right, at the same time sets up thematic and stylistic links with the other two pieces that might not be noticed in separate performance but are unmistakable in sequence. For instance, its prominent open fifths, perhaps suggested by fiddle music, are immediately picked up by D’un cahier d’esquisses alongside the rocking minor-third figure (B flat–D flat), first heard at bar 22 of Masques and echoed in the first two bars of D’un cahier d’esquisses at the same pitch. There is instantly a sense of reflection on things heard. The sarabande also takes issue, so to speak, with the cross-rhythms, the so-called hemiolas of three against two, that energise Masques from its first bar but which the sarabande turns over very slowly, like some strange object picked up on the beach. Out of this process there then emerges a rising theme that anyone familiar with L’Isle joyeuse will recognise as a preliminary version of its spectacular second subject, slow sarabande transmogrified into lilting barcarolle. L’Isle joyeuse also has its own version of the rocking thirds figure, major rather than minor, and rhythmically reversed, but still a clearly audible reference.

This is not quite cyclic writing in the Liszt or Franck sense. There is no laborious reworking of the same old material, merely a succession of hints, combined with an effortless consistency of harmonic and rhythmic style. In all three pieces Debussy handles his resources with a virtuosity that seems to belie the essential novelty of the language. Masques, in a clear A major, nevertheless has little truck with tonality in any textbook sense; instead it speeds through different regions of harmony, sometimes uncomplicatedly diatonic, sometimes pentatonic, sometimes whole-tone, sometimes chromatic. For much of the time the music is anchored by ostinato and tonic pedals – held or repeated notes that are the roots (‘keynotes’) of the chords in question, allowing for the fact that Debussy uses all kinds of chords, whether traditionally consonant or dissonant, as if they were consonant: tonal music without process, but highly refined, a peasant dance for masked courtiers in smocks. As usual with him, the predominant dynamic is pianissimo, rising steeply here and there to fortissimo, and without much in between.

In Masques the continuity is rhythmic, that of a fast dance with a lot of repetition. D’un cahier d’esquisses, very slow and ‘sans rigueur’, reveals the essential discontinuity of Debussy’s harmony. Like the Soirée dans Grenade, and as its title suggests, it has the character of a montage of unused fragments left over from other work. The first page presents no fewer than five basic ideas, one after the other and without formal linkage; and though there is some show of development on the second page, the montage principle returns and survives to the end. On its own, the piece remains somewhat baffling, which is probably why it has been relatively little played. But as part of the triptych it acts as a meaningful pause, bound together in itself by the sarabande rhythms, and linking Masques to L’Isle joyeuse by the motivic allusions I’ve mentioned. It seems to enshrine an unanswered question (the title itself is actually preceded by an ellipsis, as if something before it has gone missing). But if so, an answer is immediately forthcoming.

Of all Debussy’s piano pieces before the Études, L’Isle joyeuse is the most confident and self-assured, the most superbly extrovert. Perhaps it describes an island, but above all it describes a feeling about an island, and that feeling is certainly one of exuberance and delight. Were the exuberance and delight Debussy’s own, or were they virtual, imagined, in the spirit of his remark to Louÿs that ‘people who weep while writing masterpieces are incorrigible humbugs’?1

Sometime in 1903 he had met Emma Bardac, the wife of a well-to-do banker and the mother of an aspiring composer, Raoul Bardac, who for the past four years had been an informal pupil of his. Emma was an attractive, well-kept forty-year-old, intelligent and cultivated, and Debussy was slowly beginning to weary of his sexy but intellectually limited model wife. Unlike Lilly, Emma was a good amateur musician, a singer, and a former lover of Gabriel Fauré, whose great Verlaine cycle, La Bonne chanson, she seems to have inspired back in the early 1890s. Photographs of Emma reveal a tiny, smart, somewhat bourgeois lady in expensive hats. But in such matters appearances are often deceptive, and she evidently had no qualms about making herself available, for the second time, to one of France’s leading composers.

They were probably lovers by the early months of 1904. That April they went away together for a few days to the Vallée de Chevreuse, beyond Versailles in the Île de France, and by June they surely knew in their hearts, if not in so many words, that his marriage to Lilly was at an end. He was finding excuses to get her out of the house, or to be out of it himself. At the end of June he told his new friend, the music critic Louis Laloy, that they were shortly off to Lilly’s parents in the country. But in fact only Lilly went, on 15 July, seen off at the station by her husband and his crocodile tears. ‘Don’t believe’, he wrote to her the next day,

that I felt very joyful helping you drily into the carriage. It was even hard! Only, for reasons I’ll tell you later, it was necessary … I have next to find new things, or it will be my downfall; I’ve been uneasy for some time at turning in the same circle of ideas, [and] it seems to me that I’ve found a new track, which is why I daren’t let it go, whatever the cost.2

Lilly’s reply is lost, but it must have expressed puzzlement and some measure of self-abasement, to judge from Debussy’s next letter.

Several letters and a flying visit to the Yonne later, he at last came clean, at least by his standards. ‘I have the very clear conviction,’ he said,

after these days spent far from you, when I’ve for the first time been able to reflect on our life coolly, that even while loving you so much, I had never made you happy in the way I should … And I also recalled those trying moments when you demanded that I give you your liberty.4

‘We’re no longer children,’ he went on. ‘Let’s try to extract ourselves from this episode, without noise and without involving other people.’ A very grown-up sentiment. But alas it was crushed by a juvenile lie. The letter was sent from Dieppe and claimed that its author was on his way to London with the painter Jacques-Émil Blanche, whereas he was in fact returning to Jersey, where he had arrived with Emma at the end of July and put up at the Grand Hotel in St Helier. There he was, on his ‘joyous isle’, for several days, before moving to the mainland seaside at Pourville, just west of Dieppe. Here he could contemplate his new life, in the charming person of Mme Emma Bardac and in the abstract form of his other main work-in-progress, La Mer.

As we know, Jersey was not in fact Debussy’s original ‘Isle joyeuse’, since the title, if not the music, had been attached to a work of some kind for more than a year.5 But Debussy certainly revised the piece, possibly even recomposed it, during these weeks with Emma, and perhaps for once it really does report on feelings as, or soon after, they were being felt. Quite apart from its brilliance as keyboard music, it has a kind of structural energy and range not to be found in any previous single movement of his, not even in Fêtes, but which we shall find again, still more thrillingly, in all three movements of La Mer. The energy is already contained, waiting to be released, in the single trilled C sharp that starts the piece off, for some reason all the more exciting when it follows the final D flat chord of D’un cahier d’esquisses. Then comes the dancing main theme, ‘light and rhythmed’, cleverly expanding the trill into a melody, which adds a B to the trilled notes, then gradually spreads upwards through a scale that approximates to the old Lydian mode, with a sharp fourth, but also occasionally a flattened seventh. One might almost hear this whole first subject as a glorified variation on the trill, anchored for most of the time by rhythmic ostinatos in the left hand. The one (brief) intruder is the ‘thirds’ motive from the two preceding pieces. Then suddenly the passionately expressive second subject bursts on the scene, still in Lydian A major, but lyrical where the first theme was puckish. This is the theme adumbrated in D’un cahier d’esquisses. Debussy next, rather uncharacteristically, supplies a true development section, harmonically vagrant as in a classical sonata, followed by a compressed recapitulation and a coda based on the opening trill music, rounding the seven-minute piece off, rather untypically, with several flourishes and a loud bang.

L’Isle joyeuse is a moderately bravura, rather than a toweringly virtuoso, piece of piano writing. Yet Debussy, having only just composed it, complained to Durand about its difficulties. Still in Dieppe, he had received the printed copy and approved it. ‘But, Lord!’ he exclaimed. ‘How hard it is to play … this piece seems to me to combine all the different forms of piano attack, for it combines force with grace … if I dare say so.’6 Not only technically, but also aesthetically, the description goes to the heart of this masterpiece. Unlike the Estampes, it is neither descriptive imagery (Soirée dans Grenade, Jardins sous la pluie) nor the focused exploration of a single idea (Pagodes), but a sonata movement in all but name, complete with working of contrasted themes, a certain measure of tonal contrast, and a decisive ending. In some ways it is less radical than those pieces or than the Images that were to follow. It represents the maturing of the radical. But the truly radical had plenty more up its sleeve.

*

Joyful these Jersey days may have been, but they had their shadows. Four years later, Debussy scribbled a note, perhaps to himself, perhaps to his divorce lawyer, about the Lilly episode:

The couple returned to Paris from Dieppe on about 12 October, and on the 13th poor Lilly took a revolver and shot herself in the stomach. Pierre Louÿs, whose own marriage had perhaps been a factor in precipitating Debussy’s, went to see her in hospital.

The bullet [Louÿs informed his brother] went through her stomach twice and hasn’t been retrieved. However, the operation went well, and the poor thing seems out of danger, but she’s still broke and homeless. The husband left with a Jewess of forty-something, Mme S. Bardac. I think you know Bardac … Quite accustomed to his wife’s escapades, he smiles in response to any request for news: ‘She’s simply having a good time with the latest musician à la mode, but I’m the one with the money, so she’ll be back. Nice race.’8

Lilly’s attempted suicide was widely reported, and there were even subsequent false stories that she had tried to kill herself a second time. So once again Debussy was faced with the opprobrium of a society at ease with adultery but unable to reconcile itself to (or perhaps too ready to enjoy) the marital failures of the famous and their sometimes tragic consequences. Once again many of his friends turned against him, and when on 6 November his Danses sacré et profane for harp and strings had their first performance at the Concerts Colonne, the critics either absented themselves or took the opportunity to twist the knife in the composer’s chest, even if they liked his music. ‘I did everything I could to dislike the piece,’ one critic informed his readers. ‘It wasn’t possible … An invincible attraction … tantalising whole-tone scales.’ And Fauré, Mme Bardac’s previous musician à la mode, wrote in Le Figaro that ‘one finds there once again in profusion the same harmonic singularities, sometimes curious and seductive, sometimes merely disagreeable’.9

These charming but harmless harp dances were an ironic comment on the painful events on which they set their seal. They had been commissioned a year earlier by Gustave Lyon, director of the instrument makers Pleyel and Wolff, for a harp of his own invention, a fully chromatic, cross-strung instrument that was intended to cope better than the standard pedal harp with the complicated demands of modern music. Debussy’s work, completed in May 1904, was a promotional exercise; but it fell flat, partly because his music – harmonically quite straightforward – was playable and just as effective on the pedal harp, so made no particular case for the chromatic instrument, and partly because the music itself, however attractive, was hardly calculated to set the doves fluttering in the musical dovecots. When Ravel composed his Introduction and Allegro the following year to a commission by the Érard firm for their pedal harp, the result was in every way more spectacular and persuasive. Debussy’s pieces are lightweight, but the instrument was heavy, and yet less powerful than its rival. Today the chromatic harp still has its enthusiasts, but orchestral players play the pedal harp, even when they play Debussy’s dances.

Soon after finishing these pieces, he had composed two more sets of songs, one a setting of Charles d’Orléans and the seventeenth-century poet and dramatist Tristan (François) l’Hermite, the other a new series of Fêtes galantes to poems by Verlaine. No wonder, one might think, that La Mer was hanging fire, despite the optimistic predictions of the previous autumn. The Trois Chansons de France – as Debussy called the Orléans–l’Hermite set – are admittedly very short, not much more than five minutes for the three songs, but they have an intensity that belies their brevity. The two Orléans settings are both rondels, in which the first line (or two) is repeated at the end of each subsequent verse, a device that Debussy respects musically, as he had in the rondels by Leconte de Lisle many years before; but a feature of both songs is that, Schumann-like, they are composed essentially as piano pieces with the voice, so to speak, ‘reciting’ the poem discreetly above while the piano proceeds on its way. In the second rondel, ‘Pour ce que Plaisance est morte’ there are echoes in the piano writing of the ‘sketches’ on which D’un cahier d’esquisses is supposedly based, just as in the central Tristan l’Hermite song, ‘Auprès de cette grotte sombre’, the image of the flow of water engaged in a struggle with the pebbles generates a dark, crab-like ostinato figure that looks forward to the footprint motive in Des pas sur la neige in the first book of Préludes. If one wants to argue Debussy as a Symbolist, this tiny, concentrated song would be a good place to start. The sleeping, dreaming pool, with its bright red flowers and dangling rushes and memories of the demigod Narcissus, is like a deep dark womb into which we are ineluctably drawn, with Debussy’s nagging, swelling motive urging us on and down.

Symbolist elements intrude also in the second book of Fêtes galantes, though here they are more overt, less oblique. The supposedly innocent girls in ‘Les ingénus’ probably put on their loose skirts and unbuttoned collars knowing that the wind would reveal more than was proper; and Debussy’s wind motive, soft and suggestive in its persistent whole-tone harmonies, has a definite glint in its eye. The second song, ‘Le Faune’, about a terracotta faun laughing on the bowling-green and ‘no doubt foretelling a bad outcome to these serene moments’, must have had a special significance for the composer, since he wrote its opening ‘flute-like’ flourish on a postcard to Emma on 19 June. The erotic link with Mallarmé’s faun hardly needs pointing out. But the song itself, with its unbroken, tabor-like piano ostinato, is one of Debussy’s least interesting. On an altogether higher plane is the final song, ‘Colloque sentimental’, in which two ghosts wander round a park disagreeing about the significance of their former love. In Verlaine’s Fêtes galantes series this is the final poem, immediately preceded by ‘En sourdine’, with its nightingale, ‘the voice of our despair’. In the poem itself the despair comes as a kind of Heine-esque surprise, but ‘Colloque sentimental’ might be its explanation, and Debussy duly quotes extensively from the earlier song in this later one. ‘Do you always see my soul in your dreams? – No!’ ‘Ah! the beautiful days of unspeakable happiness when our mouths joined! – It’s possible.’ ‘How blue the sky and great the hope! – Hope has fled, defeated, towards the black sky.’ Whether or not the complex emotional texture of Debussy’s song needs ‘En sourdine’ to make complete sense of it is an interesting question. But even as it stands it is one of his most darkly moving compositions.

All this time La Mer had been brewing. It was almost a year since he had written to Colonne tentatively offering him ‘some orchestral pieces’ he was working on, and to Durand a fortnight later listing the titles of the three movements: ‘Mer belle aux îles sanguinaires’, ‘Jeux de vagues’, and ‘Le vent fait danser la Mer’.10 But thereafter the trail goes dead until July 1904, when he writes to Lilly from Paris expressing the (no doubt hypocritical) hope that ‘La Mer will be so kind as to release me so as to be with you by the 15th August’, though by the end of the month it’s ‘this “Mer” that so disturbs me’ that is keeping him in the capital.11

Of course, these were evasions, but all the same it was probably true that La Mer was giving him trouble. A glance at the eventual score reveals an intricacy of texture and a thematic complexity well in advance of anything he had previously attempted for orchestra, and at the same time he was battling with a concept that was pictorial only up to a point, and beyond that distinctly abstract and imaginary. His personal relationship with the sea was not at all like that of Turner or Monet, for whom the sea in all its moods was one crucial aspect of their thinking about light and movement. Writing to André Messager from the Yonne in the early stages of composition, he explained that ‘I was promised for the fine career of a sailor, and only the chances of existence made me branch off. I’ve nevertheless preserved a sincere passion for Her [the sea].’ But he added:

You’ll tell me that the ocean doesn’t exactly wash the hillsides of Burgundy …! And that this could well turn out like some studio landscape! But I have countless memories; that’s better, to my mind, than a reality whose charm generally weighs too heavily on one’s thoughts.12

Perhaps predictably, the sea was at its kindest for him in Jersey with Emma. He told Durand that it ‘has been very good for me, and showed me all her dresses. I’m still quite giddy with it.’13 He always loved the sea as an object of contemplation, in its changing colours and moods. The sea at Dieppe, he wrote many years later, ‘is as blue as a waltz [presumably ‘The Blue Danube’]; grey as an unusable sheet of metal; most often: green as the absinthe the old captain does without. All the same it’s beautiful, more beautiful than La Mer of a certain C.D.’ But he came to dislike resorts like Pourville, with their ugly holidaymakers (‘thieves for sure’), crowded beaches and shabby hotels. He did not – could not – swim, and seldom if ever so much as dipped his toes or even exposed any more of his body than would normally have been acceptable in a Paris boulevard. Above all he relished the calm of the ocean (even, perhaps, when it was not calm), and the fact that it was ‘always in the same place’.14

None of these images seems reflected in ‘La Mer of a certain C.D.’, except perhaps the many dresses. The original title of the first movement (‘Beautiful sea at the Sanguinary Isles’) looks like an inspiration in itself, before the music appeared, to deny its sense of placid Mediterranean radiance. The assumption that Debussy pinched the title from a short story of 1893 by Camille Bellaigue has been disputed, but chiefly on the grounds that the story has hardly any bearing on the music (apart from a brief description of the sea, it mainly has to do with the travellers’ experiences on the three islands), and that the story’s magazine publication was too remote in time for Debussy to have remembered it. But beautiful names may stick, and then be abandoned for want of real connection, or simply to avoid what might begin to look like a facile association. Debussy’s eventual title for his first movement, De l’aube à midi sur la mer (‘From dawn to midday on the sea’) already expresses the unsuitability of the original, and the music does so decisively. The middle movement, by contrast, kept its first title, Jeux de vagues (‘Wave games’); but the third movement also changed, from a dance to a conversation: Dialogue du vent et de la mer (‘Dialogue of the wind and the sea’). In all three movements, the static vision of the ‘Mer belle’ is replaced by images of motion, turbulence and fluidity. At one point, while still at Pourville in September 1904, he told Durand that ‘I would have liked to complete La Mer here, but I shall still have to finish off the orchestration, which is tumultuous and varied like the … sea!’15 So tumultuous, in fact, that it took him another six months to finish.

Of all Debussy’s greatest works, La Mer is the one that shows most clearly the benefits of his inflexibly meticulous, hyper-perfectionist approach to composition. In February 1905, excusing himself this time to Lilly for not visiting her at an agreed time, he claimed that he had spent the whole day orchestrating one page of the score: a pretext, perhaps, but a believable one. Lilly had sent him a heartbreaking letter imploring him to visit her in order to discuss their divorce rather than involving lawyers, and insisting that ‘whatever they may say or do to you, you have no better defender than the one you once loved (Lily or Lilo), the two!’ She uses the polite ‘vous’ throughout, but only, she explains, because ‘I no longer dare tutoyer you, not out of irony, but solely because it’s the normal thing in our situation. Please believe for always in the sincerity of my affection.’16 As Debussy later admitted, such pleadings did not leave him entirely cold. But they could not (whatever he also claimed) penetrate his work.

The importance of the orchestration in La Mer becomes immediately apparent from close listening. In the first movement especially, the whole form – the shape of what one hears – is articulated by it. It’s as if Debussy set sail, as one might, without a clear itinerary in mind, and allowed the conditions, the wind and waves, the sun and rain, to dictate his course. On the sea there is a certain monotony of context, a sameness or relatedness, quite unlike the varied character of a journey by rail or road; and Debussy is able to use this idea as a unifying factor while presenting his material as a sequence of more or less consequential events. At first each event is defined by its instrumental colour, as was also the case with Nuages. In essence, motives are not shared: a seagull is not a fish or a wave, though they all depend on the water. This is not babyish pictorialism. The music works, precisely, as a chain, with occasional repetition but no formal, structural recapitulations; these aesthetically clinching elements will come later in the work. Meanwhile the first movement proceeds, through the musical spray, in the manner of a journey, until (at about halfway, though of course this isn’t apparent as one listens) there is a pause and a sudden, dramatic change of colour, in the form of a rhetorical theme for divided cellos (Debussy requires sixteen of them). This is the start of a slow build-up to the movement’s coda, where, for the literal minded, the midday sun suddenly emerges and irradiates the entire watery scene.17

No wonder Debussy changed his title. This is not at all music you could put in a frame; it is too active and protean. It has at least a dozen important themes, some fragmentary, some decorative, some lyrical, some ceremonious, all linked by a sort of musical DNA, but differentiated by the astonishing diversity and refinement of Debussy’s orchestration. On the DNA side, some features pop up like the family nose. Most of the themes have triplet figures prominent, or sometimes triplets squashed into a Scotch snap,18 as in the very first oboe and clarinet motive, which also introduces a semitone rise and fall – a sort of aquatic sigh – that will infect several of the later themes. Some motives remain fragmentary, single figures attached to variants of themselves; others are extended into arabesques that nevertheless depend on multiple repetition. By the end of the movement one is conscious of having travelled in a coherent but progressive way; but as for the destination, that remains to be seen.

The second movement, though scherzo-like in general character, is more static in its sense of place. We are, let’s say, in mid-ocean, and the play of waves is to some extent a play of ideas already glimpsed in passing. Debussy establishes the position harmonically. The first movement ends brilliantly in D flat major (having started far away in B minor); the second opens in C sharp (that is, D flat) minor, but with a disturbing element of F sharp minor in the first chords. The D flat–C sharp linkage is exactly as between D’un cahier d’esquisses and L’Isle joyeuse, but here more sombre and unstable. As before Debussy tends to define his motives instrumentally. For instance, the main cor anglais theme (apparently a descendant of that instrument’s theme in Nuages) is only ever played by woodwind, usually oboe or cor anglais. Later, however, there is some fusion and a growing sense of culmination in the sharing and extension of ideas. The second theme (violins), whose DNA goes all the way back to the flickering semitones of the initial oboe theme of the first movement, is later recapitulated by flutes and oboes, and throws off a long-drawn string counter-melody leading to a genuine climax, before the waves die down and the music fades out quietly in E major.

The ‘dialogue’ of the finale turns out to be to a large extent a conversation between themes already heard. Debussy refrained from calling La Mer a symphony, no doubt all too aware of the expectations that would arouse. Instead he compromised with ‘symphonic sketches’, though that in turn carries implications – of some kind of improvisatory freedom – that might be no less misleading. His structures and linkages may be of a new kind, but they are far from casual. All the same, his third movement reminds us that the French symphonic tradition (Berlioz, Franck, Saint-Saëns, etc.) was cyclic, with themes transmogrified from movement to movement. This leaps to the ear with the trumpet theme a minute or so into the finale, a brightened-up version of the horn theme in the first movement introduction, while the increasingly dominant second subject is the culmination of the sighing triplet and Scotch snap motifs from the previous movements. Eventually the sea falls calm, and this theme sounds out over the waters, like the cry of a sea bird, in the first movement’s D flat major, leading finally – after further wind-blown dialogue with the trumpet theme – to a coda based on that movement’s ‘midday’ coda, at which point we recognise that the aesthetics of musical form have triumphed over the logic of imagery and anecdote.

Leaving aside all these colourful details, the brilliance of La Mer lies especially in its mastery of continuity and balanced form without any recourse to textbook procedures. Impressionist sea paintings like Monet’s Storm, off the Coast of Belle-Île or Whistler’s Arrangement in Blue and Silver: The Great Sea are organised partly, of course, by a deliberate limitation of motive, but within that by a big colour variation within a narrow range of tones. Essentially, Debussy does the same in La Mer. The subject is self-limiting and, like the two paintings, devoid of human interest (apart from a barely perceptible ship on Whistler’s distant horizon). Like them, it depends on brush-strokes to suggest the texture of wind and water, brush-strokes that combine to create the whole image, while remaining distinct as contributory detail. The differences are obviously generic. Music, since it operates by the clock, can create motion of a kind, where a picture can suggest it only metaphorically. But in the end it cannot, unlike Whistler’s watercolour, prove to us that its subject matter is what the title says it is. One hostile critic, reviewing the first performance on 15 October 1905, satirised Debussy as ‘dedicated to the writing down of the impalpable. He makes of the impalpable a sonority.’19 Another, by no means hostile, critic found that

for the first time, listening to a picturesque work by Debussy, I have the impression of being, not at all in front of nature, but in front of a reproduction of nature: a marvellously refined, ingenious and industrious reproduction, perhaps too much so; but a reproduction all the same … I don’t hear, I don’t see, I don’t smell the sea.20

Since this particular critic, Pierre Lalo of Le Temps, had written very sympathetically about Pelléas et Mélisande, Debussy saw fit to respond.

I love the sea; I have listened to it with the passionate respect one owes it. If I’ve transcribed badly what it dictated to me, that’s no concern of yours or mine. And you’ll allow that all ears don’t hear in the same way. In the end you love and defend traditions that no longer exist for me, or at least that only exist as representatives of an epoch, in which they weren’t all as beautiful or as worthwhile as one might care to say, and the dust of the Past isn’t always respectable.21

He thus moved his ground from a defence of his imagery to a defence of his music, about which Lalo had remarked that ‘the lack of development and logic that existed in his previous works, and which went by unnoticed, here becomes visible’. In effect, Debussy is insisting that whether or not we can see, hear or smell the sea in La Mer is of no great importance; what matters is that we feel the coherence and abstract beauty of his music inspired by it. Perhaps, mutatis mutandis, Whistler and Monet might have said the same about their paintings.

He had at last completed La Mer, after so many distractions, in early March 1905. There had been the usual phantom completions, and the inevitable interruptions for more or less fruitless theatre projects, of which only two fragments of music for a production of King Lear by André Antoine ever saw the light. The disasters of his private life (at least as seen from the vantage point of the Paris press) had fanned the flames of artistic notoriety arising out of the success of Pelléas. All of a sudden he was a public figure. Even the behaviour of his admirers became the object of scrutiny. The Symbolist poet Jean Lorrain satirised what he called the ‘Pelléastres’ in Le Journal:

As for Emma, it was widely supposed that Debussy was pursuing her for her money, though, as we saw, her disposable finances were in fact her husband’s. There was question of an inheritance from a rich uncle, but Lesure has suggested that she had no particular reason in 1904 to expect anything from that direction, and in any case the uncle was still very much alive, and died only in 1907.23 They were, nevertheless, well enough off to be able, in October 1905, to take an apartment in the Avenue du Bois de Boulogne (now the Avenue Foch, one of the smartest addresses in Paris). Debussy had just signed an exclusive contract with Durand, and though his own divorce from Lilly, which finally came through in early August, was expensive for him as the guilty party, Emma had managed to divorce her husband (in May) on the basis of shared responsibility, because of his misdemeanours with assorted actresses and call-girls, which will have secured her some alimony. Through all these trials and tribulations she was expecting a baby, and at the end of October she gave birth to a little girl, Claude-Emma, quickly and thereafter known as Chouchou. Fatherhood intrigued Debussy, and his letters henceforth contain frequent colourful references to his clever and personable daughter. It did not, however, help his finances, any more than the expensive address guaranteed the funds to maintain it. He never again moved house, but his finances soon reverted to their original state.

After finishing La Mer in March he had had to turn his attention to preparing various of his earlier piano pieces, including Pour le piano and the old (but newly titled) Suite bergamasque for publication by Fromont. With most of the older pieces – the Rêverie, the Mazurka, etc. – this was pure commerce. But he thought well enough of the suite to revise it, especially Clair de lune, work that might have partly served as preparation for the new piano pieces that he had contracted to write for Durand in July 1903 before even starting La Mer. He now wrote to Durand once again listing the titles of the first three of these solo piano Images and the three for piano duo. Then, unable to face the concluding confrontations of his divorce proceedings, he fled in late July with Emma to Eastbourne, on the south coast of England, and put up grandly in the Grand Hotel. ‘The sea’, he informed Durand, ‘stretches out with a very British propriety. In the foreground a combed and pomaded lawn on which important little imperialist Englishmen disport themselves.’

‘I think’, he added, ‘that I’ll be able to send you some music next week.’ But as usual the week stretched out like the sea, and three weeks later he was still writing to his long-suffering publisher excusing his delay in sending the pieces. He had corrected the proofs of the piano duo arrangement of La Mer. But he had, he explained, finally decided that he didn’t like the first Image, Reflets dans l’eau, and had made up his mind to write a completely new piece with the same title, but ‘based on new ideas and according to the most recent discoveries of harmonic chemistry’.25 At last, after another ten days, he sent the three Images, blaming the further delay on an attack of facial neuralgia caused, apparently, by the cool sea breezes. But he was proud of the music. ‘Without false vanity,’ he told Durand, ‘I think these three pieces will hold their own and take their place in the piano literature … (as Chevillard would say) to the left of Schumann or to the right of Chopin … as you like it.’26

The original Reflets dans l’eau (‘Reflections in the water’) was presumably one of the two pieces that Debussy had already played to Viñes in December 1901 (the other was Mouvement). If so, it is hardly surprising that he came to feel, in 1905, that it would no longer do. In between had come Ravel’s Jeux d’eau; and although Debussy’s new Reflets have rather little in common with Ravel’s Jeux in specifically musical terms, they certainly bear the imprint of the possibilities not only of ‘harmonic chemistry’ (an old phrase of Debussy’s) but also of keyboard sonority and imagery that Ravel’s piece had unlocked.

Not the least striking thing about Reflets dans l’eau is that, composed only a few months after the completion of La Mer, it presents such a startlingly different musical image of water. There, all was movement and flecks of colour, flying spray, wind and a sense of travel. Here, the waters are at first still and silent (how else would they give back reflections?); Debussy called the opening ‘a little circle in water, with a little pebble falling into it’.27 The pebble distorts and fragments the reflections, and the music gradually explores the resulting colours and shapes, peering more and more deeply into the spreading ripples of the dark pool, like Narcissus in ‘Auprès de cette grotte sombre’. Debussy’s musical image for this mobile stillness is a softly resonant series of D flat major harmonies in chords ranging up and down the keyboard, some consonant, some dissonant, but at first without chromatic notes, then gradually complicating the sound with chromatics and increasingly brilliant arpeggio flourishes, as if the pool had suddenly developed a cascade or two, before slowly subsiding into a stillness even more profound than the one with which it began. What is truly astonishing about this wonderful music is that Debussy presumably composed it without a piano at his disposal; at least he told Durand there were none, and one can hardly imagine that, even with his newfound affluence, he had a room with a grand piano in it. Yet the voicing and delicacy of the chords and the treatment of resonance and pedallings (implied but, as usual with him, never indicated) are of a precision seldom to be found even in keyboard music where voicing and resonance obey textbook procedures that Debussy had by now abandoned.

Is Reflets dans l’eau really about reflections in a dark pool, or is the water a metaphor for something in the human psyche? Is it just a pool, like the pools in Monet’s garden at Giverny, or is it a symbol of emotional or sexual perturbation, or a mother fixation, or general goings-on in the unconscious? If it comes to that, are Monet’s water lilies just water lilies, or are they ‘a vision of the secret, inner communication of things, a sense for the way flower and cloud, earth and sky, are bound together by sublime affinities’?28 The answer is, presumably, that Debussy’s music and Monet’s paintings can be all or any of these things, according to the preferences of the listener or viewer. It’s certainly true that Reflets dans l’eau rises very briefly to a fortissimo climax that is hard to explain in terms of pools or cascades or anything less spectacular in watery terms than a flash flood, though it is easy enough to explain in terms of the music itself, whose harmonic and textural build-up positively cries out for some such release. One of the problems about the silly argument between Impressionism and Symbolism where Debussy is concerned (‘useful terms of abuse’, he called them) is precisely the neither-or-both syndrome, as we saw in the discussion of La Damoiselle élue in Chapter 5. All one can say with confidence is that, like the Impressionist painters and the Symbolist poets (among others), Debussy achieved his greatest work by discarding traditional grammars and trusting his own formal and aesthetic instincts.

At least the title of Reflets dans l’eau obviously goes with the music. The title of the second Image, Hommage à Rameau, might almost strike one as ironic. In his Gil Blas articles Debussy had made no bones about his admiration for the eighteenth-century French composer Jean-Philippe Rameau, and his corresponding dislike of the German-born Gluck. It was a preference based only partly on nationalist prejudice, but partly also on what he saw as shared musical values: French elegance and lightness and a feeling for the subtle rhythms of the French language, as against Gluck’s ‘affectation of German profundity’ and his ‘need to emphasise everything by thumping the table or to explain breathlessly, as if to say, “You are a collection of particular idiots who understand nothing.”’29 But one can play Debussy’s homage, or listen to it many times, without detecting any hint of the French eighteenth century in any shape or form. There is no trace of pastiche, in the sense of Ravel’s Tombeau de Couperin. The pianist is directed to play ‘in the style of a Sarabande, but without strictness [sans rigueur]’, and it’s true that the music is in a slow triple time with an occasional sarabande stress on the second beat. But this is a purely technical point. The music has no flavour of the dance, the barring is irregular, and the phrasing often contradicts any conceivable dance pulse.

None of this, needless to say, is intended as negative criticism of Debussy’s piece, which is beyond question one of his finest. Like Reflets dans l’eau it works through a process of gradual intensification and elaboration from a simple, in this case unison, beginning. But there is no image; the idea is entirely musical. After the chant-like opening melody, again dominated by rocking thirds like the ones that linked the three pieces of the L’Isle joyeuse set, Debussy progressively intensifies the procession of chords and harmonies to a level of subtlety and richness even he has not achieved before. The music is in effect a study in soft, transparent, plangent sonorities, sometimes sequences of triads using the full spread of the hands, sometimes complex dissonances that seem to emerge from these triads and, so to speak, interrogate them. There are two or three loud climaxes, bracing the structure; but for the most part the chords, voiced with incredible care and precision, need to be quiet and attentively pedalled for their resonances to shine through. The whole piece is a vindication of Debussy’s argument (with Guiraud) for a criterion of beauty in harmonic design independent of rules or theory. Yet behind these seemingly arbitrary chordings is a masterly feeling for the architecture of very slow music, achieved, amazingly, without the help of harmonic conventions that guided Beethoven in his great slow movements or Wagner in his mauvais (as Rossini thought, but Debussy did not) quarts d’heure.

Mouvement, by contrast, reverts to bravura pianism in the manner of Pour le piano, but with many refinements. The particular subtlety lies in the hidden melodies, which float across the right-hand arpeggios, triple piano, but with stress marks on the melody notes and the instruction: ‘All the notes marked with the __ sign [to be] sonorous, without hardness, the rest very light but without dryness.’ The difficulty of coordinating this network of requirements, at high speed and without breaking the rhythm, can probably be imagined, even by those for whom the piano and all its ways are an impenetrable mystery. Mouvement is not, on the face of it, one of Debussy’s most musically interesting pieces, but in performance, with the right delicacy, the right distinctions of touch, and, once again, discriminating control of the sustaining pedal, it can make a brilliant conclusion to this first book of so-called Images.

Do they, as a set, justify the generic title? Not if one insists that an image is something exclusively visual. But there is such a thing as a verbal image, which amounts to a metaphor, a representation of one thing by another; so why not a musical image, for instance, representing movement (in general) by music that reduces movement to a series of rapid elemental figures articulated by melody? Hommage à Rameau perhaps escapes even this piece of special pleading. One of Debussy’s most completely self-contained pieces, it superbly defies classification.