In the autumn of 1899, just as Debussy was finishing the Nocturnes, his domestic life took a new and superficially encouraging turn. He married Lilly Texier. He was so poor that they were unable to pay a priest and had to marry at the mairie; in order to afford a wedding breakfast, he had to give a piano lesson to his Fontaine pupil, Mlle Worms de Romilly, on the morning of the ceremony, and their honeymoon after it was a visit to the zoo at the Jardin des Plantes.1 His letters to Lilly in the months leading up to their nuptials are evidence of an almost overpowering commitment, both emotional and sexual, to this pretty but intellectually rather ordinary young woman. Or perhaps what they indicate is a desperate need for the emotional security that he optimistically imagined would come with the sexual attraction he patently felt for her. One wonders what she can have thought of the gushing intimacy of these letters, so extravagant that the editor of the English edition of his correspondence preferred to leave all but one of them out. She must certainly have doubted their complete sincerity after returning to Paris unannounced one day in May 1899 and finding her Claude in bed with another woman, identity unknown. Briefly she walked out on him, then responded to his remorseful pleadings and avowals, if anything even more extravagant than before. He seems to have calculated correctly that she would respond to effusions of the ‘I yearn for the blush of your red lips’ variety that he would never have stooped to set to music, always for him the mistress whose character and judgement he respected above those of any mere mortal woman.

It looks, all the same, as if Lilly may temporarily have shut the door on this more demanding lover. After finishing the Nocturnes in December 1899 Debussy composed practically nothing for the best part of a year, nothing except the fragmentary music for the Bilitis tableaux vivants, composed the following winter. He was probably toying with ideas for piano pieces that began to take shape early in 1901, and there was the small matter of preparing a proper piano-vocal score of Pelléas et Mélisande with a view to a possible production in the 1900–1901 season. Two years before, the then new director of the Opéra-Comique, Albert Carré, had issued a vague promise to programme the opera, after struggling up to the composer’s fifth-floor apartment in the rue Cardinet with the theatre’s conductor, André Messager, and hearing him play it through. But for that to become a practical reality, it was necessary to produce materials that would be of use to other performers. On one level this will have been a routine task, since the music was fully composed and Debussy had given private performances on the piano on a number of occasions. But the need to produce a piano part that other people could play may well have confronted him with precise issues of keyboard texture and colouring in contexts with a visual or theatrical dimension beyond that of a song accompaniment. More generally, it may have given him a taste for keyboard composition on a scale that, for one reason or another, he had seldom previously attempted. But, if so, that certainly was not the only prompting that, in the next two or three years, turned him decisively towards the piano as a vehicle for his most advanced musical ideas.

For one thing, he was certainly acutely conscious of an existing repertoire of French piano music that to some extent avoided the – in his opinion – tedious conventionality of so much of his countrymen’s symphonic and operatic work. He was a particular admirer of Chabrier, whose piano works had a vividness and individuality very different from the kind of well-organised écriture that was taught at the Conservatoire (where as a matter of fact he had never studied). Chabrier had written for the piano with instinctive bravura, incorporating popular and pictorial elements, and rule-breaking harmonic procedures; there was a colouristic and rhythmic brilliance to his pianism and an absence of head-in-hands self-communing that surely appealed to Debussy, though he never remarked on it. Chabrier had died in 1894. But in March 1898 Debussy had attended a concert given by Marthe Dron and the Catalan pianist Ricardo Viñes that included a pair of piano duos entitled Sites auriculaires by a Conservatoire pupil of Fauré’s by the name of Maurice Ravel. Debussy had been so impressed by one of the pieces, Habanera, that he had borrowed the manuscript from the young composer. No doubt he was aware of the music’s debt to a piece by Chabrier with the same title; it even has an ostinato pedal C sharp picked up from Chabrier’s Habanera (where it is written D flat). But harmonically Ravel’s piece is more daring than Chabrier’s, as if the image of exotic Spain had inspired a specifically colouristic approach to chord design.

How long Debussy kept Ravel’s manuscript is not recorded, but within three years he had composed a two-piano habanera of his own, which he called Lindajara, after the courtyard garden of that name in the Alhambra, Granada. It was his first, but by no means last, keyboard homage to Spain, a country he barely visited (he had seen a picture of the Court of the Lindajara in a magazine).2 Nor would it be his last piece of modest plagiarism of Ravel. Lindajara is almost twice as long as Ravel’s Habanera, but harmonically more discreet and pianistically less colourful. Debussy would soon do better than this in the Hispanic field, while still retaining memories of the borrowed manuscript.

Something else happened in Paris in 1900 that might have influenced him to write for the piano. The Exposition Universelle of that year was in many ways as grand as its 1889 predecessor. It gave birth to the Grand Palais and the Petit Palais; it was a major showcase for Art Nouveau in general, and for a whole gallery of improbable scientific and technological concepts, including talking films, diesel engines, various applications of electricity, and Russian dolls. As in 1889, it had a colonial section that enabled French visitors to congratulate themselves on their superior culture and their noble, disinterested efforts at civilising the rest of the world, especially the parts colonised by them. Debussy was not the kind of liberal thinker who might be shocked by such condescension, whether or not he would exactly have condoned it. As before, his interest was purely musical, and here he was by no means disposed to regard the ‘colonial’ performances de haut en bas. Though there is no certain evidence that he heard the Javanese gamelan that was again on display, it would be hard to imagine that he did not. As for circumstantial evidence, there is a good deal of this in the music he composed in the next two or three years.

Roy Howat, in his superb Art of French Piano Music, argues plausibly that Debussy’s return to piano music early in 1901 was precisely because the recent experience of the gamelan had reminded him that the percussive nature of his own instrument was as much a matter of refined, precisely articulated sonorities as it was of the rich, blended textures of the romantic keyboard.3 His first essay of this kind, after finishing the Bilitis music, was a pair of toccata-like pieces – a Prélude and a Toccata so called – which he combined with a slightly revised version of the Sarabande from his otherwise unpublished set of Images of 1894, and published under the baldly functional title Pour le piano, as if deliberately hinting that this might be music conceived for some other instrument or instruments, then recast ‘for the piano’.

Whether or not that instrument might have been a refined percussion orchestra like the gamelan is an open question. The clipped, non legato opening of the Prélude and the bravura ‘touch’ music of the Toccata might suggest some such equivalence. But, leaving aside the Debussyan modal and whole-tone harmony, the parallel chord sequences and long sustained bass pedal notes, the music both looks and sounds more like something out of a Couperin ordre or a Rameau harpsichord suite, or even a Bach partita. Admittedly, influence is a slippery beast. How many composers, after hearing a Wagner opera, go home and start writing like Wagner, or planning an opera about gods and dwarves? They are at least as likely to do the opposite and write a comedy of manners for chamber forces, or go off at some tangent or other, directed by some remote association entirely private to themselves. Just occasionally, when influence takes the form of mimicry, we can cheerfully spot the connection. But these are not the most interesting cases.

Debussy completed Pour le piano in April 1901, and that same month, a married man with responsibilities, he finally yielded to the calls of Mammon and took a job as music critic for the twice-monthly arts magazine La Revue blanche. The hoped-for production of Pelléas that season had not materialised, and he perhaps felt momentarily at a loose end in his creative work. But he had no intention, as he told readers of his first column at the start of April, of acting as a common or garden reviewer.

He was as good as his word, but only for a short time. Between April and December he wrote a total of nine columns, then resigned, claiming ‘overwork and nervous strain’.5 Pelléas was now firmly timetabled and in need of orchestrating. The gamekeeper had to return to poaching.

The Revue blanche at that time, in the closing years of its relatively short life, was or had recently been associated with a number of the most prominent figures on the anti-establishment wing of the arts, many of them, including André Gide and Henri de Régnier, personal friends of Debussy. Insofar as it had a specifically political tendency, it was anarchist. Its managing editor, Félix Fénéon, had been arrested and briefly imprisoned in 1894 on suspicion of involvement in the bombing of the Foyot restaurant, and when the Dreyfus affair erupted soon afterwards, the Revue sided firmly with the Dreyfusards. But to call Debussy an anarchist in this sense, as Richard Langham Smith does in his invaluable English edition of the composer’s articles, is a shade far-fetched. On Dreyfus, he seems to have been neutral or to have blown with the wind, and in general, politically, he could best be described as disengaged. Intellectually, though, he was antinomian to the point of perversity. For reasons of artistic self-identification, he fought ceaseless battles with the fashionable tendencies of the day, refusing on principle to accept even the most normal or uncontroversial idées reçues. To some extent this was no more than a willingness to be rude about composers and conductors whose genius or talent he didn’t at bottom deny. Sometimes when reviewing new works by minor locals, he seems to have been curiously incapable of distinguishing gold from pinchbeck, but what music critic, however brilliant, was ever good at this? Schumann, for instance, was not. Hanslick, Ernest Newman, Andrew Porter were not. But Schumann (like Newman and Porter) saw it as his duty as a critic to keep himself informed and to describe and assess what he heard as sensitively and dependably as he could. Debussy was interested only in defining his own position in relation to what he was reviewing, with the result that his criticisms have a curiously lopsided feeling, something idiosyncratic that in the end relates to nothing except his own creative consciousness.

Thus, a symphony by Georges Witkowski sparks off a diatribe on the uselessness of the symphony since Beethoven. A group of Schubert songs are ‘bits of faded ribbon, flowers forever dried, and photographs of the departed’. Beethoven’s sonatas are ‘very badly written for the piano’. The main aim of Massenet’s operas is ‘to find documents to serve as a history of the female soul … [His] harmonies are like arms, the melodies like the napes of necks. We bend over women’s foreheads in order to find out at all costs what is going on behind.’ ‘What I like best is a few notes on an Egyptian shepherd’s flute: he is part of the landscape, and hears harmonies unknown to your treatises.’ For some reason, but perhaps with the Impressionist painters in the back of his mind, he calls for a music of the open air, which ‘would play in the free air and soar joyfully over the tops of the trees … Perhaps we would find there a way of obliterating those petty obsessions with over-precise form and tonality that encumber music so maladroitly.’ On the other hand, his composer’s percipience gave him insight into the genius of Musorgsky’s Nursery songs, three of which he had heard sung by Marie Olénine at a concert of the SNM in March.6 Meanwhile, he completely ignored several concerts that professional critics would have regarded as unmissable.

In order to justify such omissions and purvey such thoughts with a certain degree of impunity, he invented the figure of Monsieur Croche, a latter-day version of Valéry’s Monsieur Teste, a crotchety (though a ‘croche’ is a quaver in French), embittered, wizened old cynic who has seen and heard everything, describes himself as an ‘antidilettante’, and makes purposely provocative remarks such as: ‘To see the day dawn is more valuable than to hear the ‘Pastoral’ Symphony. What’s the use of your almost incomprehensible art?’ He rings the composer’s doorbell one evening, for no apparent reason, and walks or is invited in. ‘At once he aroused my curiosity,’ Debussy tells us disingenuously, ‘with his particular vision of music. He would talk about an orchestral score as if it were a painting, hardly ever using technical terms.’ Debussy then makes the classic mistake of insulting his readership’s musical perceptions, but by putting the insult into M. Croche’s mouth he can claim innocence. He even has the old brute insult him.

M. Croche is of course simply a reductio ad absurdum of Debussy’s own attitudes, which still reflect his student rebellions against everything the Conservatoire held dear. Obviously he regards the Prix de Rome as ridiculous. Predictably he makes fun of the Opéra, and in particular lays into Saint-Saëns and his latest opera, Les Barbares. He attacks narrow musical specialists, those who by implication know only music. ‘For me, to specialise is by as much to limit one’s universe … I know all the music there is, and all I’ve retained is the special pride of being protected against any kind of surprise. Two bars give me the key to any symphony or any other musical anecdote.’8 This is essentially Chabrier’s complaint about the tedious repetitiousness of Paris musical life, coloured by Debussy’s own preference for the company of non-musicians. The trouble is that M. Croche offers no solutions to the problems he identifies. All he seems to be saying, in the end, is that musical life is a waste of time and the sooner we’re shot of it the better. And Debussy, as devil’s advocate, reduces himself to the weak protestations of the ordinary punter who buys his tickets at the box office when he can afford them and expects to be allowed to enjoy himself. Taken as a whole, these early reviews are entertainingly, enjoyably perverse, but they change nothing. As a later composer, Hugh Wood, has memorably written, ‘“The composer speaks” – as great a moment as “Groucho sings”. Similar, too, in other ways: you hope that this is not what he does best.’9 In Debussy’s case, we can safely say that it was not.

*

Debussy’s final article for the Revue blanche appeared at the beginning of December 1901, and within a few weeks he was in the thick of preparations for Pelléas et Mélisande at the Opéra-Comique. Rehearsals with the singers began in mid-January, but even before that a quarrel had blown up between him and Maeterlinck over the casting of the role of Mélisande. At the end of December the Scottish soprano Mary Garden, recently the star of Gustave Charpentier’s Louise, had been announced for the part, which Maeterlinck claimed had been promised to his mistress, Georgette Leblanc, a more dramatic soprano who had sung Carmen. By the end of January, Debussy was claiming to friends that Maeterlinck was ‘in the bag’, but if so he refused, like the proverbial cat, to stay there. In February the composer had to endure tribunals at the Société des Auteurs together with Maeterlinck. While it seems to have been officially accepted that Debussy was within his rights to cast the opera as he wished, Maeterlinck remained on a war footing almost until the premiere. On 14 April, with the performance imminent (though not yet finally scheduled), he published a letter in Le Figaro disclaiming any association with the opera and wishing it ‘a prompt and resounding failure’.10 Worse, he plotted physical violence. Leblanc herself describes an occasion when her lover, brandishing a walking stick, announced that he was off to give Debussy ‘a few whacks to teach him some manners’. She goes on:

I waited in a state of anguish, certain there was going to be trouble. I couldn’t visualise Debussy, with his tragic features, taking correction kindly! At every instant I looked down the deserted street to watch for Maeterlinck’s return. At last he appeared at the top of the rise, waving his stick at the sky with comical gestures. His report was pitiful. While still barely in the drawing room, he had threatened Debussy, who had calmly sat down in an armchair, while Mme Debussy rushed distraught to her husband with a bottle of salts. She had begged the poet to leave, and, to tell the truth, there was nothing else he could do. Maeterlinck liked musicians no more than he liked music, and he went on repeating: ‘They’re all mentally sick, these musicians!’11

For Debussy, the Maeterlinck affair was merely a sideshow amid the stresses and strains of rehearsing Pelléas, music of a completely unfamiliar character both for the orchestra and for the cast, who, a fortnight before the premiere, were still singing their music as if it were normal opera, that is ‘at the tops of their voices’.12 Then with only a fortnight to go before the pencilled-in premiere in mid-April, the conductor André Messager suddenly started asking for extra interlude music to cover the changes of scene: initially between the scenes of Act 2, then for the Act 1 changes, and finally for the change between the second and third scenes of Act 4. Altogether, Debussy had to write almost 150 bars of new orchestral music in time, as it turned out, for the delayed premiere on 30 April. As Orledge points out, Debussy could compose fast when he had to, but without the time to be as meticulous as usual, and above all without the opportunity to prevaricate about the likely completion date, the result inevitably lacked the stylistic precision of the rest of the opera. Writing at speed, he tends to lapse into a refined Wagnerism; he himself drew a young composer friend Gustave Samazeuilh’s attention to a certain flavour of Parsifal in the Act 4 interlude, which he had supposedly written in one night just before the dress rehearsal.13 But though designed to solve a problem ad hoc, the interludes improve the opera’s pacing and formal balance even when it is staged, as most often now, with a single set and no drop curtain. The occasional performance omitting them tends to emphasise the brevity and multiplicity of the scenes, as against the continuity of the whole.

For various reasons the open dress rehearsal on the 28th was occasionally disrupted by laughter and some cat-calling, provoked especially by Yniold’s scene with Golaud (his scene with the sheep was cut), and by Mary Garden’s Scottish accent, which drew forth a good deal of mimicry from those audience members who could be relied on to see nothing absurd in Siegfried’s ‘Das ist kein Mann’ or Werther’s detailed description of his own grave as he dies of a self-inflicted bullet wound. The premiere on the 30th, however, went off smoothly and without any disturbance, and was even a great success. Debussy had evidently managed to impose a certain restraint on the operatic singers, to judge by one reviewer’s comment on the ‘sober declamation where not a word was lost and the orchestra was used with the utmost discretion’.14 Garden’s accent seems to have faded into insignificance beside her musical artistry and strong dramatic presence. And though the press was mixed, as it nearly always is where radical art is concerned, there was enough support for Debussy to take a whole paragraph of his ‘Reply to the Critics’ in Le Figaro listing those he wanted to exclude from the strictures and satires that followed.15 ‘Accusations’, Roger Nichols reports of the reviews, ‘were largely of formlessness (no arias), melodic and rhythmic monotony (no dances), lack of noise (sparing use of trombones) and unintelligible harmonic progressions (few perfect cadences).’16

The truth is that Pelléas was so unlike anything ever previously staged at the Opéra-Comique or, for that matter, at the Opéra, that it would have been astonishing if press reaction in general had been anything but bemused or hostile. Audiences are different in this respect from critics, being used to enjoying things that strike them as beautiful, without worrying how they are made or how they might best be described. Pelléas had an important precedent in Musorgsky’s Boris Godunov, a profoundly radical masterpiece detested by the St Petersburg critics in 1874 but so successful with audiences that it remained in the Maryinsky repertoire until Musorgsky died in 1881. It is perhaps no surprise that Debussy loved Boris and considered it an important influence on his own opera, though at this time he knew it only from the vocal score. Pelléas, too, was successful enough to be frequently revived, both in France and abroad, and it established its composer, more or less overnight, as one of the most important and influential figures in French music.

Earlier that same April a concert had taken place in Paris that was to have a profound effect on Debussy, though he was probably not present, being intensely involved with the Pelléas rehearsals. Ricardo Viñes, the young Catalan pianist of the Sites auriculaires, gave another Ravel premiere, this time of a solo piece with the Lisztian title Jeux d’eau. Debussy was already a friend and admirer of Viñes, who had given the first performance of his Pour le piano three months earlier. So even if unable to attend the Ravel premiere, he might already have known the work, which Viñes had previously performed privately, or he might have first encountered it when it was published later in the year.

Jeux d’eau is not a big piece (it lasts five minutes or so), nor is it a theatrical bravura study in the manner of Liszt, although its technical difficulties are on the Lisztian scale, and its graphic portrayal of the play of water in an artificial fountain was surely inspired by the Master’s Jeux d’eaux à la villa d’Este. Ravel completely escapes from the idea of the romantic showpiece, and concentrates instead on the image of the fountain as a mechanical construct, controlled indeed by a highly complex algorithm, but in no sense expressive of human fantasy or aspiration. The effect is achieved mainly by exploiting the glittering, more resonant upper register of the piano with, at the same time, both the soft (una corda) pedal and the sustaining pedal in play. The upper registers of the piano are in any case undamped – that is, reverberate freely whether or not the sustaining pedal is depressed. But they ring still more when the lower registers are undamped, and it is Ravel’s exploration of this aspect of piano acoustics that gives Jeux d’eau its unusual fascination.

However, there is another quite different yet in a strange way related feature of the piece that will certainly have intrigued Debussy. The lettering on the title page suggests an Art Nouveau derivation from some kind of oriental script, and the music itself is very obviously a reminiscence of the gamelan, which Ravel will already have heard aged fourteen at the 1889 Exhibition, and no doubt again in 1900. At the start, the right-hand arpeggios hint at a pentatonic (black-note) scale, and although the music quickly diverges into all kinds of more or less chromatic variants, the pentatonic motive returns several times, both in the form of arpeggios and as melody, for instance at bar 19. The black-note figures ring out more than the chromatically altered ones, because the notes are all on the same harmonic spectrum, and this means that the degree of reverberation constantly fluctuates, like sunlight reflecting off water droplets. Of course, the resonances also vary in an ordinary piece of tonal music: in fact, it’s partly because the notes of the tonic triad are the lowest notes on the spectrum that they create a sense of closure. But Ravel exploits this property outside the region of strict tonality, allowing the colours and shades to flicker in and out, until the final page, where the music settles on a pentatonic version of an E major chord (with added C sharp and D sharp), pedalled ‘jusqu’à la fin’ (‘to the end’).

Debussy’s Pour le piano predates Jeux d’eau and reflects different keyboard interests, more technical and abstract, less graphic. But according to Lesure he had for some time had in mind a series of piano pieces that he intended to call Images, in which some graphic element would presumably come into play.17 There is no evidence that any of this got composed before the actual signing of a contract with a new publisher, Jacques Durand, for two sets of solo Images and a set for two pianos or orchestra in July 1903. Instead he completed that same month a quite different set, which he called Estampes (Prints) and which Durand published three months later.

Since the Pelléas premiere, and its revival six months later, which also took up a good deal of his time, Debussy had toyed with other theatrical projects. A plan to turn As You Like It into an opera with a libretto by Paul-Jean Toulet never got as far as music paper, though it generated some lively scenic ideas. Debussy had spent a few days in London in July 1902, and had come home in a Shakespearean mood after seeing (though certainly not understanding) Forbes Robertson as Hamlet. But Toulet had gone off to Indo-China at the end of the year, had come back an opium addict, and that was that. Debussy’s ideas for the scenario show that he had an idiosyncratic vision of Shakespearean comedy that sheds some light on his other choice of topic this same year, Edgar Allan Poe’s farcical tale The Devil in the Belfry. He wanted, for instance, to introduce an offstage chorus to provide a commentary to the wrestling match in Shakespeare’s first act; he wanted the character of the clown, Touchstone, made more of; and he wanted the final scene, the betrothal of Rosalind and Orlando, stylised into a masque, strictly choreographed and ‘mixed in with songs, but in the antique manner, that is, forming part of the action’.18

This project was soon abandoned, but the Poe idea was the prelude to what was to become a lifelong obsession with the American writer, whose work, then as now, was far more highly regarded in France than in either the United States or Britain, and had been lovingly translated by two of her greatest poets, Baudelaire and Mallarmé. French writers of the Mallarmé generation found in Poe an echo of their own preoccupation with the fantasies and psychological terrors left uncared for by the materialism of the modern world. For them, in a word, he was a father figure of Symbolism. Debussy seems already to have tangled with one of his most disturbing tales, The Fall of the House of Usher, and would do so again. But The Devil in the Belfry is more of a satire on fantasy, in which, in the Dutch village of Vondervotteimittis, the devil decides to upset the ordered lives of the good villagers by making the midday clock strike thirteen. Debussy produced a scenario that shows that he planned to turn this humble nonsense into a wild ensemble piece, with dancing and choruses and a devil who whistles throughout. A few pages of sketches survive, and Debussy was still thinking about the work in 1908, when he included it in a contract with the New York Metropolitan Opera, and possibly even still in 1911, when the Opéra-Comique announced it, along with Usher, for its coming season. But the actual music made no progress after 1903, unless one counts the one-minute Morceau de concours that Debussy knocked up out of the sketches as his contribution to an identification competition run by the journal Musica towards the end of 1904.

In fact he had made no progress with anything of any substance since the completion of the Nocturnes at the end of 1899. Pour le piano had been only partly new, and there had been a commission for a saxophone piece (‘that aquatic instrument’, as Debussy called it)19 from the president of the Boston Orchestral Club, Elise Hall. But in the main he was drifting, marking time, waiting for Pelléas and waiting, probably, to find out what could conceivably productively follow that epoch-making work.

Then at the start of 1903 he again found himself employed as a music critic, this time by the literary journal Gil Blas. This was a more intense commitment than his Revue blanche column of two years earlier. For the same reason, it lasted a mere six months, but involved far more and more frequent writing and even some travel. His first piece took him to Brussels to review (very favourably) d’Indy’s opera L’Étranger; then in late April he paid a second visit to London in order to write about Wagner’s Ring conducted (marvellously, he said) by Hans Richter. Mostly, though, he was caught up with the usual Paris music, its occasional triumphs and frequent irritations. There are the same old obsessions, the same prejudices as well as some new ones. In only his second article he launches into an extended and only half-ironic defence of the barrel organ, as the basis for another diatribe about open-air music, but declines to discuss Berlioz’s Damnation de Faust, a work – the reader is left to deduce – that he dislikes. Another bête noire is Gluck, whom he consistently compares unfavourably with Rameau, thereby reviving a famous eighteenth-century controversy in a new form. Gluck is partly to blame for Wagner and Berlioz (‘always the favourite musician of those who knew little about music’20). Debussy loves Weber, loathes Mendelssohn and Grieg (‘Scandinavian one might be, one is nonetheless human’),21 admires Richard Strauss in spite of himself, and is warm in his praise of Franck’s Symphony and Rimsky-Korsakov’s Antar.

Behind all this is a certain incipient nationalism, a feeling (not unnoticed even now by visitors to France) that French is best – French delicacy, French precision, French restraint – even if ‘one may regret that French music has, for too long, followed paths that treacherously distance it from that clarity of expression, that precision and compactness of form, particular and significant qualities peculiar to French genius’.22 And the villains? The usual suspects: the Opéra, the Conservatoire, the Prix de Rome, the concert repertoire in general (too many operatic excerpts). But as before Debussy is better at identifying shortcomings than at suggesting remedies, with one striking exception. Reviewing Chevillard’s concert performance of Das Rheingold on Good Friday, he expresses regret at the whole idea of performing Wagner in this way.

You will object that there’s magic in the orchestra. But if one is a Wagnerian, is that enough? I shall dare timidly to propose film as a means of animation; we have seen in a recent play at the Ambigu how powerfully it could increase the emotion. We should be of our time, and we have no right to deprive Wagner of the expressive ingenuity of this invention, of which the ‘Music Halls’ make such marvellous use, in an infinitely less lofty cause.23

Not until the 1940s was film used in this way for Wagner, either in concert or on stage. And although, as Debussy admits, the idea wasn’t strictly his own, it was typical of his eclectic approach to the arts that he picked it up on the streets and, in his mind, transferred it to the art he knew best.

There was one other eclectic element about Debussy’s music column, namely that Gil Blas had decided – perhaps fearing his expertise – to employ a second music critic, in the attractive and not over-specialised form of the popular novelist Colette. Week by week the two columns were printed side by side, and Colette, though by no means musically unversed (after all, her husband was the prominent music critic Henri Gauthier-Villars, alias Willy), treated hers partly as an outlet for social comment, gossip and an occasional flurry of bitching. She was good at playing the poor little audience member. For instance, in praising Gabriel Pierné’s Concertstück for harp and orchestra (which Debussy also liked), she added that ‘between ourselves, I never know whether it is just a nice tickle I get from harp concertos or whether they make me want to go pipi’.24 When Strauss conducted in Paris at the end of March, she went into paroxysms of distaste.

No! Even if Debussy attacks me – or, worse, replies with a spiteful silence – I do not like that music! … I cannot accustom myself to this man, who chooses an aggressive tumult barbed with piercing piccolos for a sticky romance – a hedgehog garlanded with forget-me-nots! … Richard Strauss [also] conducted a ‘love scene’ from his opera Feuersnot. Once again my ears went bzi-bzi! That a love scene? My God, if I went into such tumultuous ecstasies I’d be afraid of what my neighbours downstairs might say!25

History is silent on whether or not Debussy attacked her for such remarks, though his own reaction was somewhat different. Over Feuersnot, he agreed that ‘one episode that raised orchestral torrents seemed ferocious for a love scene’, but conceded that ‘they were probably justified by the drama’. He rather liked Aus Italien (which was Colette’s garlanded hedgehog), was carried along by Ein Heldenleben, and concluded that Strauss was ‘a great figure [with] the frank and decisive appeal of those great explorers who walk among savage tribes with a smile on their lips’.26 Between the two of them, it must be said, a by no means entirely misleading picture of Strauss and his music emerges.

Debussy’s final Gil Blas column appeared at the end of June in the form of what he somewhat prematurely called ‘The Balance Sheet of Music in 1903’. Much of it was about opera, or to be exact the Opéra and the Opéra-Comique. But it included two deathless remarks of a more general and revealing character. ‘To conclude,’ he urged, ‘let us stay in France. Let us wish people there to like music a little less, only so as to like it a little better.’ ‘Those devoted to Art’, he had written a few paragraphs before, ‘[are] unshakeably in love with her, and on the other hand, one will never know to what extent music is a woman, which perhaps explains the frequent chastity of men of genius.’27

*

In the case of this particular genius, Art had returned unexpectedly, like Lilly three years earlier, and caught him in flagrante delicto with a news sheet. Early in August he wrote to Louÿs from Lilly’s parental home in the Yonne that he had just composed a piano piece called La Soirée dans Grenade, and had also been working on the Saxophone Rhapsody that was his belated, and somewhat reluctant, answer to the Boston commission of 1901.28 La Soirée dans Grenade is the second of the three Estampes, so he had perhaps already written the first piece, Pagodes, and must have added the third, Jardins sous la pluie, well before the end of August, since by mid-September he was returning corrected proofs of the whole set to Durand and making suggestions for the cover design. On the same day that he sent back the proofs, he told Durand that he was working on his new orchestral piece, La Mer, which may or may not have delighted the publisher, who had a signed contract for three sets of Images still on his desk, its ink barely dry. Truly, the floodgates had burst open. ‘I’ve been doing a lot of work for you,’ Debussy had written coyly at the start of August, ‘and my music is perhaps worth more than my literature.’29

Debussy scholars have tended to be guarded about whether this sudden rush of piano music had or had not been inspired by Ravel’s Jeux d’eau. On the face of it, it seems improbable. Why should Debussy, a newly famous composer entering his forties, allow himself to be led by the nose by a young whipper-snapper barely out of his Conservatoire short trousers? Rivalry would be absurd, not least because, however brilliant the Ravel, Debussy, with half a dozen or more masterpieces behind him, had no need to feel upstaged by one five-minute piece. And yet Pagodes responds to Jeux d’eau in so many ways that the connection is hard to deny, especially if you then turn to Soirée dans Grenade and consider its debt to Ravel’s Habanera. The simple explanation is that Debussy was intrigued by Ravel’s music and wanted to adapt its underlying concepts to certain ideas of his own, in a spirit more of cooperation than of rivalry. It’s as if, in hearing it or, more likely, trying it out, he thought: yes, excellent, but what if, instead …? For Pagodes is not just a kind of oriental clone of Jeux d’eau. It does certain things in a specifically Debussyan way; and Soirée dans Grenade might use Ravel as a starting-point, but is soon exploring a somewhat different musical terrain. ‘If this music isn’t exactly what they hear in Granada,’ he remarked to Louÿs, ‘well, so much the worse for Granada.’30

The first thing one might notice about Pagodes is the marking below the first bar: ‘2 Ped’, exactly the same as at the start of Jeux d’eau. Both pieces start pianissimo, with the soft and sustaining pedals both depressed, suggesting a very refined touch with the upper registers of the instrument creating a delicate harmonic aura around the melody. But such markings are a rarity with Debussy, who, as we saw, generally preferred to leave pedalling to the performer, in view of the great differences between instruments, hall acoustics and atmospheric conditions. In Pour le piano, for instance, there is not a single pedal indication. Pagodes, much more than Jeux d’eau, is built round unchanging harmonies that can be pedalled without unwanted clashes. Ravel’s piece starts with a simple pentatonic melody, but the left hand clouds the issue and soon creates complexities of light and shade that require extremely subtle half-pedalling – light touches on the sustaining pedal – to avoid harmonic chaos by limiting bass reverberation. Pagodes is more about harmonic ‘fields’, of the kind already identified in earlier works, including Pelléas, but especially a feature of gamelan music. Jeux d’eau may be gamelan-like in its texture and sonority, but Pagodes comes as close to imitating gamelan music as would seem possible without completely abandoning the aesthetic and intellectual world of Western music.

This difference already comes out in the titles. Jeux d’eau is a description, an image translated into sound, with an epigraph, from Henri de Régnier, that backs this up: ‘Dieu fluvial riant de l’eau qui le chatouille’ (‘The river god laughing at the water that tickles him’). Pagodes is obviously not graphic in this sense, but is more like a keyword, directing our gaze towards the lands where pagodas raise their upturned eaves. The intention is suitably vague, because pagodas are not characteristic of Java, so have no topographical connection with the gamelan. But Debussy may also have had a structural principle in mind, because pagodas are layered towers, and Pagodes, too, is composed in layers, both vertical and horizontal.

The vertical layers are the various strands of melody starting with the pentatonic melody in the right hand at bar 3, accompanied first by offbeat chords, then by a left-hand counter-melody plus chords. These elements grow in complexity, but never really interact; like a peal of bells, they form strands that resonate each other but remain essentially separate. The horizontal layering is more hybrid, because for the piece to make sense within its five minutes or so duration it needs a narrative element. It has, therefore, a kind of rondo form, with brief episodes of a more chromatic, less stable harmony, that lead eventually to a rhetorical climax of a distinctly Western character. But within this conventional-looking design, the music is a series of distinct units, linked by family resemblance but not really developing or evolving except through more and more elaborate ornament. Compared to Jeux d’eau, Pagodes perhaps leaves a somewhat primitive impression. But it has consequences in Debussy’s music that are anything but primitive.

There are hints of this already in La Soirée dans Grenade (‘Evening-time in Granada’). The only, slightly embarrassing, snag with this beautiful piece is that it is so obviously a parody (in the original, neutral sense) of Ravel’s Habanera – a better piece than the Ravel, no doubt, but shamelessly derivative of it. The borrowings are flagrant: the repeated C sharps, less persistent in Debussy but essentially the same figure; the sultry triplet chords at bar 23 (from Ravel, bar 9, etc.); the dotted left-hand figures at bar 33 (from Ravel, bar 14, also on F sharp), and in general the habanera rhythm and detailing. It’s true that Ravel’s piece is little more than a sketch, a montage of a few style features bound together by a persistent ostinato pitch and rhythm. If it had been found among Debussy’s papers and in his hand, it would probably have been accepted without much question as a preliminary draft for La Soirée dans Grenade. Instead, Debussy must consciously or unconsciously have thought: excellent, but a lot more can be done with that material. This piece is the result.

It may be that the montage idea was what initially appealed to him. Ravel, a native of the Basque country (but on the French side), seems to have thought of Spain as a set of picture postcards, a slideshow of characteristic scenes or fragmentary impressions: a hot sultry day, a girl in a mantilla, a man driving a donkey up a hill, a boy tuning a guitar, a general flavour of mañana. Why it should have been Spain that inspired this particular approach (rather than, say, Turkey or Morocco) is something of a mystery; perhaps it was its proximity and relative familiarity, combined precisely with those other exotic aspects, the Moorish, the African. But the first composers to think of Spain in this travel-poster way were Russian: notably Glinka, and Rimsky-Korsakov, a composer admired, and occasionally imitated, by both Ravel and Debussy. Glinka’s Souvenir d’une nuit d’été à Madrid (‘Memory of a Summer Night in Madrid’) might have been a specific precedent for Debussy, though he probably didn’t know it. He almost certainly had heard Rimsky-Korsakov conduct his Capriccio espagnol at the 1889 Exhibition, and perhaps the fourteen-year-old Ravel had as well.

La Soirée dans Grenade is a street scene: atmosphere with passers-by. The C sharp ostinato at the start hangs expectantly in the air, and soon a distant voice intones one of those laments, all triplets and augmented seconds, that announce deep Spain as surely as the ‘Song of the Volga Boatmen’ announces antique Russia. This in turn is interrupted by a strummed guitar phrase, and so it goes on. At first the events alternate, then briefly they combine, at which point Debussy adds a third stave in order to clarify the various layers of the texture. Towards the end a fragment of flamenco is spliced in, different from its context in both rhythm and tempo, and then on the final page the various episodes reappear in a more rapid montage of fading incident, leaving first the lament, then the ostinato, more restful because the harmony is now resolved on to the F sharp major that Debussy decides, on the basis of earlier unresolved dissonances, to treat as the home key.

The ancestry of Jardins sous la pluie (‘Gardens in the Rain’), unlike that of the other two Estampes, lies in Debussy’s own music. The content, like the title, of the third of his 1894 set of Images, ‘Some aspects of “Nous n’irons plus au bois” because the weather is unbearable’, had been obviously provisional. Bad weather now suggested a recast and more tightly composed piece based partly on that song, but more especially on another children’s ditty, ‘Dodo, l’enfant do’ (a lullaby of the ‘Rock-a-bye baby’ variety), a phrase from which is played by the left hand at the outset. Various stories were told of Debussy refusing to come in out of the rain while sketching this piece, but nobody seems to have had any theory about the connection between the children’s songs and wet weather. The alleged reason in the poem for not going to the woods was that the laurels had been cut, though as usual with nursery rhymes the subtexts proliferate and logic goes out of the window (where it may or may not be raining).31

Debussy of course chose these tunes, not for their texts, but for their value as melody; their simplicity of line was useful to him as a framework for a variety of watery figurations. There is, if you like, rain a-plenty here, though it may be harder to detect the gardens under it. But at bottom, Jardins sous la pluie is an abstract keyboard study in the melodic articulation of right-hand arpeggios. Debussy, like Schumann, was fond of left-hand melodies in the middle of the texture, and this piece is a prime example. ‘Nous n’irons plus au bois’, however, comes only high in the right hand until near the end, when it twice rings out fortissimo in the left hand, but in treble register.

It was also presumably important for him that the tunes were well known; anybody French would recognise them at once (foreigners perhaps less so). They are, in effect, inset images, and in this respect they work like ‘Au clair de la lune’ in the song ‘Pierrot’, or the slices of musical life in La Soirée dans Grenade, breaking into the natural discourse of the piece, usually in a strange key, so that the inset quality is doubly underlined. But whereas the Soirée is made up almost entirely of such images, in the form of montage, Jardins sous la pluie is carried along for most of its length by sheer bravura, and the layering, such as it is, is little more than the outlining of familiar tunes by the figuration itself. To this extent, it is not essentially different from any worthwhile nineteenth-century piano study, where the tedium of exercise is lightened by an unexpected richness of hidden melody.

How much weight should we give at this point to Debussy’s choice of the title Estampes for these three pieces? To compare them to Japanese prints as such would clearly be hopelessly laboured, even if Pagodes and Jardins sous la pluie are the kind of subject matter we might not be surprised to find in such prints. After all, the music barely represents these subjects in any case; and, as Paul Roberts has pointed out, Spanish flamenco is not a subject much treated by Japanese artists. What I think Debussy was trying to do was associate himself with a tendency in the visual arts with which he sympathised and which he felt himself to be in some sense adapting to the medium of sound. The changing colours and perspectives in La Soirée dans Grenade, the concentrated tone-colour motives in Pagodes, the flickering waters of Jardins sous la pluie: above all, perhaps the sense that, in pursuing these subjects in these ways, he was creating a new kind of music that disrupted the old grammar and the old routines, just as the Impressionist painters were replacing the figurative grammar of the Académie des Beaux-Arts. The comparison is interesting and useful, as long as it is not pressed too far. Two of the main qualities of these pieces, their simultaneities and discontinuities, have no real equivalent in the visual arts, being aspects of our experience of time. Maybe they suggest an experience of space, just as painting can imply the passage of time, but only as metaphor.