One hopes that Debussy’s parents had eventually managed to be proud of their son’s fame and distinction – the Légion d’honneur, the seat on the Conseil Supérieure, the prominence of his (and their) name on the Paris billboards – even if they found it hard to understand why these accolades never seemed to convert into the kind of affluence that most parents hope for – not entirely disinterestedly – for their offspring. Whatever they thought, at the time, of his divorce and eventual remarriage, they had drawn closer thereafter. And whatever the Avenue du Bois de Boulogne did to his finances, it did no harm to his outward respectability. Manuel and Victorine Debussy had moved into an apartment nearby, and a late photograph of them with their son suggests the warmth and mutuality of a suburban homecoming.

Debussy was sincerely distressed by his father’s death. ‘Although we practically never had an idea in common,’ he wrote to Caplet, ‘it’s a loss I feel more deeply each passing day. He had a sort of admiration for L’Enfant prodigue, and the Chicago success [of Pelléas a week after he died] would certainly have enhanced that beyond all measure; one can, without irony, regret that he never could have that simple pleasure.’1 The ‘without irony’ tells us, of course, that Manuel Debussy would never have been capable of understanding Pelléas itself or its continued success. But, like most of us, he understood fame and would have been happy to refer that understanding back to his genuine liking for his son’s, after all by no means wholly conventional, earlier work.

Manuel would also have been impressed with a pair of agreements that Claude struck that autumn, the first a month before his death but when he was already desperately ill, the second six weeks after it. At the end of September 1910 the composer signed a contract with the Canadian dancer Maud Allan to provide by the end of February 1911 the music for an Egyptian ballet of her devising called Isis, for which he would receive the impressive sum of 20,000 francs plus a fee of 50 francs per performance. Then on 9 December he signed another contract, this time with Gabriel Astruc, to supply by early April the incidental music for a monumental drama about St Sebastian by the Italian playwright and literary adventurer Gabriele D’Annunzio, in return for a further 20,000 francs. Manuel might not have noticed or been troubled by the absurdity of his painstaking, unpunctual composer son undertaking to produce twenty or thirty minutes of dance music and an hour or so of theatre music for chorus and large orchestra within the space of about six months. But Claude, as we have seen, could be as devious about deadlines as he was meticulous about his work, and there is good evidence that he never had any serious intention of fulfilling his contract with Maud Allan to the letter. D’Annunzio’s initial approach at the end of November found him in Vienna, where he was conducting the Konzert-Verein orchestra in a programme of his own music, and in reporting to Emma about the concert and about his reply to D’Annunzio he added casually, ‘I mustn’t forget Miss Maud Allan’, as if that is precisely what he would have liked to do.2 It was true that by that time the work had changed its title, and presumably its subject matter, though not its locale. It was now Khamma; but ‘as we’re nevertheless staying in Egypt’, he wrote to the impresario Joseph Schurmann, ‘it’s much the same to me’.3

Maud Allan was a new kind of fish in Debussy’s aquarium. He had had dealings with women of doubtful reputation, had certainly slept with them and had even, in the opinion of some of his former friends, married one. But he had never written music for one; nor had he come across anyone who had elevated the louche and the seductive into something at least approaching high art. Maud was a dancer in the spirit of Loïe Fuller and Isadora Duncan – not, that is, a trained ballerina, but a free movement artist who performed solo in a partly improvised fashion, using the whole body as a vehicle for emotional expression. But there was more to her than pure expression. Her work was strongly angled towards the sensual, and tended to favour subjects of a sexually suggestive kind, such as Botticelli’s Primavera and, most famously, what she called The Vision of Salome, in which she danced the seven veils in an exiguous, mostly transparent costume of her own design clutching a gruesomely realistic head of John the Baptist; but she was also a good musician, a trained pianist, and an intelligent woman who saw her work, not just as staged flirtation, but as the visual realisation of an idea. It was in the nature of the time, and as it happened of her own background, that she attracted scandalous gossip. Her brother, Theodore Durrant, had been hanged at San Quentin prison in 1898 for the murder and necrophilia of two young women, and no doubt in people’s minds her dancing was the sort of thing you might expect of someone with a brother like that. It was also rumoured that she was lesbian. It was a combination of horrors that sent a shiver down the collective spine of the middle-class audiences who nevertheless flocked to her shows.

Probably Khamma and Isis were the same plot with the title changed. It had been extracted by Maud Allan herself, in collaboration with an admirer, the English writer William Leonard Courtney, from an Egyptian story found on a stele in a Theban temple by the Rosetta Stone decipherer, Jean-François Champollion, and retold by Gaston Maspéro in his Contes populaires de l’Égypte ancienne. It concerns a young dancer who is instructed by the high priest to dance before the statue of the god Amun-Ra in order to secure deliverance from the hordes besieging the city. She performs three dances, and when, at the conclusion of the third dance, the statue moves its arms and turns the palms of its hands towards the dancer, she launches into a fourth, more ecstatic dance, at the end of which she falls dead. Needless to say, and despite the need for a corps de ballet of worshippers and a supporting male mime for the priest, all this was really no more than a pretext for Allan to display her talent for sinuous, improvised movement to suitably exotic music. But she had reckoned without Debussy’s dilatoriness, and his instinctive obstinacy when faced with anything in the nature of an ultimatum. She was also careless over contractual details. She had omitted to include in the agreement any mention of the work’s duration, so that when Debussy finally, in January 1912, got round to producing a piano score lasting twenty minutes, she was in no position to grumble that she had expected forty, nor that she had wanted six or seven dances when he had given her only four. Bully as she might, he flatly refused to extend his score by a single bar, and when she tried to lecture him on the effect of his music, he exploded to Durand about the ‘detestable Maud Allan’.

But although he was rude about the dancer-librettist and claimed to hate the subject matter of her ‘wretched little Anglo-Egyptian ballet’, he took composing the score perfectly seriously and never for a moment lapsed into the kind of routine snake-charmer’s music that the subject might have seemed to invite. In fact, Khamma is surprisingly unremitting in its harmonic strangeness, from the sinister chromatic rumblings of the opening with its distant trumpet fanfares ‘that savour of riot and conflagration and send a cold shiver down your back’,5 through the chilly whole-tone solemnities of the high priest’s prayer to Amun-Ra, and on into the three dances that Khamma reluctantly offers, like the sacrificial victim in Stravinsky’s exactly contemporary Rite of Spring, knowing perhaps that it will cost her her life. The first dance is another of Debussy’s slow sarabandes, much varied in rhythmic texture but constant in tempo, and punctuated at one point by the distant fanfares of the opening. The second dance is quicker, more volatile, ‘light and fearful’ the composer indicates. The third is again slow, and the most overtly sensuous of the three, as if the dancer were straining all her physical resources to draw a response from the stone god. Only in the spontaneous fourth dance, Khamma’s dance of triumph, does the music develop a strong forward impulse, and the frenzy hardly lasts long enough or achieves enough abandon to explain the death of a healthy young girl.

At this point one can almost sympathise with Maud Allan’s regret at the ballet’s brevity. One can picture her (if she had ever danced the work, which she never did), exotically costumed, moving and gesturing with great improvisatory freedom, then coming too suddenly and literally to a dead end, and having to lie still for several minutes while the corps de ballet pirouetted round her and the high priest pronounced a blessing over her. But this is the inevitable fate of the solo dancer who aspires to something loftier than the variety turn of Pavlova’s dying swan or Allan’s own Vision of Salome. Debussy’s score is not his greatest, neither as rich in ideas as Le Martyre de Saint Sébastien, nor as brilliantly achieved as Jeux, but it is a well-organised, intriguing miniature drama, intensely personal in its harmonic idiom, and lacking only the radiant beauty of his very finest work. Orledge detects the influence of Stravinsky in the unusually abrasive harmonies; by the time Debussy composed Khamma he had heard Petrushka, and it’s true that analytically the chord structures are now and then similar, though Debussy’s subtle sound perspective, his way of distancing elements of the harmony behind or beyond other elements, together with the extreme refinement of his detailing, produces a totally different effect from Stravinsky’s habitually bold, forward textures.6 Orledge, a great admirer of Khamma, is probably right that it would have had a better fate if Durand had ever published an orchestral score. The reason he didn’t, presumably, is that Debussy himself orchestrated only a few bars at the start, then handed the finished piano score to Charles Koechlin, a composition pupil of Fauré, who duly orchestrated the rest. The result is stylish and effective, but it naturally handles the music respectfully rather than creatively. There are no discoveries, as in Debussy’s own orchestral scores, only colourings in.

If Maud Allan’s business-like, slightly controlling tone was distasteful to Debussy from the outset, D’Annunzio knew how to chat up a touchy composer as if he were the one person in the world worth knowing and working with. His first, unsolicited letter of November 1910 is almost like a proposal of marriage.

One faraway day, on the hill of Settignano that is the birthplace of that most melodious of Tuscan sculptors, Gabriel Mourey spoke to me of you and of Tristan in accents profound. I knew you and loved you already. I frequented a little Florentine group in which a number of serious artists made a cult of your work and were excited by your ‘reform’. Then as now I suffered from the inability to write the music for my tragedies. And I dreamt of the possibility of meeting you. That summer, while I was planning a long-meditated Mystery, a lady friend was in the habit of singing me the most beautiful of your songs, with that interior voice so necessary for you. At times my dawning work trembled at it. But I dared not hope for you … Do you love my poetry?7

It would be pleasant to report that these heart-warming sentiments were wholly sincere, though in fact Debussy was D’Annunzio’s third choice, after failed approaches to Jean Roger-Ducasse and Henry Février. But in any case he was probably seduced at least as much by the financial aspect of the collaboration as by the poet’s honeyed phrases. From Budapest, where he was playing immediately after his Vienna concert, he responded to D’Annunzio in kind. ‘How could it be possible for me not to love your poetry, just as the thought of working with you puts me, in advance, into some kind of fever?’8 But he told Emma that ‘this business doesn’t tell me anything worthwhile’, and there is more than a trace of irony in his reply to another ‘terrible and charming’ missive from D’Annunzio in January, ‘in which one can hear the triumphant reverberations of a march towards glory. And now all music seems pointless to me beside the endlessly renewed splendour of your imagination. Which is why it’s not without a certain terror that I see the moment approaching when I shall decidedly have to start writing …’9

The triumphant reverberations of D’Annunzio’s Le Martyre de Saint Sébastien echoed to an enormous, wordy five-act drama that, when performed with Debussy’s music, lasted the best part of five hours. The music was an essential part of the concept, and the composer himself figured in a lengthy prologue spoken before the curtain by the Nuncius (Messenger, or Announcer), in language akin to that of the author’s correspondence. It speaks of Claude Debussy,

who sounds fresh as the new

leaves under the latest shower

in an orchard of the Île-de-France,

where the almondless almond trees

light up the surrounding grass

in a Saint Germain grove

which remembers Gabrielle

and the faun king, and their love …

and so on for many lines. The play itself is mainly in blank verse, but includes more formal, rhyming sections that look as if intended for musical setting, though not all were set by Debussy. What he did write amounts in time to about a fifth of the drama, nearly an hour of music, composed in less than three months, an unprecedented rate for him, considering the size of the forces involved and the complexity of the whole project, with text arriving erratically and out of order. It necessitated, he told one correspondent, ‘writing music without stopping and with the regularity of a machine for manufacturing hats’.10

D’Annunzio’s actual title for his play was Le Mystère de Saint Sébastien, and it added a number of uncanonical details to what was known about the saint’s life. The (more or less) historical Sebastian had joined the Roman army in about ad 283 as a cover for his Christianity, became a captain of archers, made many secret converts but was eventually exposed and ordered by the Emperor Diocletian to be tied to a post and shot by his own archers. D’Annunzio’s long first act (or ‘mansion’, as he rather obscurely termed his acts) is dominated by the torture of a pair of Christian twins, Marc and Marcellien, observed by Sebastian, who subsequently converts the twins’ parents and sisters and a procession of other pagans, and in the process inevitably reveals his own Christianity. In the second act, Sebastian breaks into the magic chamber of the seven planetary sorceresses. In the third he confronts Diocletian in the altar room of the many ‘false gods’; in the fourth he is tied to a pole, while his archers plead with him to renounce his faith so that they will not be forced to kill him, but to no avail, and in the short final act the Soul of Sebastian is received into Paradise, in the garden of light and beatitude.

No doubt as for many of the artists who have painted the beautiful Sebastian naked and pierced with arrows, the sexual and sado-masochistic aspects of the subject were a major attraction for D’Annunzio, a noted poetic decadent and poseur who liked to extend the extravagances of his writing into the way he lived his life. Deeply in debt, he had abandoned his Florentine villa early in 1910, arrived in Paris in time to attend the Ballets Russes productions of Scheherazade and Cléopatra, and been fascinated by the performances of Ida Rubinstein in these two works. Ida, like Maud, was not a fully trained dancer; she depended for her effectiveness on stage on her slender, willowy physique, beauty of movement, and an intense presence, helped, from time to time, by her willingness to appear with next to nothing on. It’s possible that the idea for a Sebastian spectacular was hers; it was the initial production of a company she herself formed after breaking with Diaghilev at the end of the 1910 season, and it offered her a part that both precisely suited her somewhat androgynous beauty and fed her ambition to take on speaking as well as dancing roles.

The pagan elements in D’Annunzio’s treatment may well also have appealed to her, with her memories of Cleopatra and the Sultana Zobeïda in Scheherazade. But the handling was his contribution alone: the blurring of the division between the pagan and the Christian, the image of Sebastian as an Adonis, a Greek god led astray, as Diocletian seems to argue, or as a Roman officer almost too much loved by his men. The author shares their regret that this beautiful young man should renounce the sensual pleasures of the world for a painful death and the questionable joys of an afterlife without sin, and for the agnostic, beauty-loving Debussy, these ambiguities were surely a stimulus. They colour, for instance, the exquisite song of the Virgin Erigone in the second act, with its magical, serpentine polyphony of flutes and high soprano, and the rich chordings of the Syrian women’s lament for (as they suppose) ‘the beautiful Adonis’ in the third act. They even colour the more churchy elements, such as the Prelude, with its Cathédrale engloutie parallel chords, rendered more mysterious by their curiously irrational sequence of minor and major, not governed, as in the piano piece, by the obvious shape of a melodic top line.

Debussy’s Saint Sébastien score is hardly a unified piece of work. How could it be, consisting as it does of eighteen separate movements, some purely orchestral, some vocal or choral, some in the form of melodrama (speech over music), often separated by long stretches of heavily metaphorical poetic declamation? Instead its varied character might suggest a new set of preludes, a succession of images controlled, however, by a text and a single narrative line. Even within sections there is something of the same kind of montage. The solemn prelude, an unbroken sequence of tonic triads for woodwind followed by a haunting oboe melody in Dorian E flat minor, runs directly into the strange unaccompanied duet of the twin brothers, tied – as Sebastian will be – to posts, facing one another, a sadistic tableau with obvious sexual – not to say homosexual and incestuous – overtones, which Debussy seems to abet by casting them both as contraltos singing in unison or close harmony. The oboe melody comes back in the next chorus (‘Sébastien, tu es le témoin’), but a good forty minutes later and probably long forgotten by the audience. Finally, in this more than hour-long first act, Sebastian dances on the burning lilies, to music of an urgent refinement somewhat in the manner of Le Vent dans la plaine, but with a barely definable elevation that seems to echo the saint’s ‘Je danse plus haut que la flamme, sept fois plus haut’ (‘I’m dancing higher than the flame, seven times higher’).

Often in this marvellous but uneven score one is conscious of Debussy’s familiar style taking on new colourings in response to the overheated strangeness of D’Annunzio’s conception, in which early Christianity is – perhaps rightly – depicted as no less remote and cabbalistic in its way than the seven planetary sorceresses in their magic chamber. The Virgin Erigone sings a beautiful floating arabesque; Sebastian dances, and is mourned, to dissonant but finely designed harmonies that take to a new pitch Debussy’s preoccupation with the voicing and balancing of individual chords. Simple parallel chords like the ones that open the work crop up a number of times, often in mysterious sequences that suggest a secret and alien church. Some of these new colourings come from his work on The Fall of the House of Usher, and perhaps, in his haste, he even used material intended for the opera, but there are also less interesting patches. The persistent slow music occasionally drags its heels, and the choral final act, in Paradise, has that slight flatness that often afflicts the music of Heaven by composers who have no particular desire to go there. It has been suggested that the weaker moments are the work of André Caplet, who certainly orchestrated substantial parts of the score under Debussy’s guidance, as is confirmed by a precise memoir of Jacques Durand. But Durand says nothing about Caplet writing any music, only about his having filled in spaces in Debussy’s marked-up open score – a more or less mechanical process for a trained musician, and one easily checked by the composer. As Lesure insists, the character of the actual orchestration – its radiance and transparency – is plainly Debussy’s own achievement.11

D’Annunzio’s play, with Debussy’s music, was performed for the first time on 22 May 1911 at the Théâtre du Châtelet, with Ida Rubinstein as Sebastian, choreography by Mikhail Fokine and designs by Lev Bakst. Caplet conducted. All kinds of problems attended Armand Bour’s staging, especially to do with the difficulty of coordinating the music with the action in the scenes with chorus, a situation aggravated by Bakst’s insistence on spreading the chorus round the stage as part of his colour scheme. Then, a week before the performance, the Archbishop of Paris gave the show unintentional publicity by pronouncing anathema on it and threatening Catholics who attended with excommunication, a fairly reliable way, one would think, of ensuring that they would turn up in droves. In the end the long evening, which continued into the small hours, was a tolerably ramshackle affair. Ida as ever moved beautifully, showed off her wonderful legs to great effect, but spoke an impenetrable French ‘through her nose’, and rolling her ‘r’s ‘as if she had a mouthful of pebbles’.12 Louis Laloy, who adored and admired D’Annunzio personally, considered that ‘he has never had any sense of the theatre’.

*

Three weeks after the D’Annunzio premiere, on 13 June, Debussy attended the first performance of the new Diaghilev ballet, Stravinsky’s Petrushka, also at the Châtelet, and like many in the audience that night was stunned by the work’s mastery and originality, and perhaps also, with D’Annunzio still big in his ears, by its economy of means. He had also been at the first night, the previous June, of Stravinsky’s Firebird. But this was something different. His supposed reply when the young Russian asked him what he thought of The Firebird – ‘what do you expect, one has to start with something’ – may or may not be apocryphal, but he certainly told Stravinsky, ‘You will go farther than Petrushka, that’s certain, but you can already be proud of what that work represents.’ He was particularly struck by the scoring. ‘I don’t know many things’, he said, ‘as good as what you call: “Le Tour de Passe-Passe” [“The Conjuring Trick”, by which the Magician brings his three puppets to life]; there is in it a sort of sound magic, the mysterious transformation of mechanical souls who become human by a magic spell, of which you so far seem to me to be the sole inventor.’14 He imagined Stravinsky living with his characters in the way he had been living with the Usher family; and he must have talked about the work a good deal at home, because a few months later the seven-year-old Chouchou had ‘composed a fantasy on Petrushka to make tigers roar’.15 But he had nothing to say about the dancing, not even about Nijinsky, whose Petrushka was to remain one of his greatest roles.

Debussy himself may have had enough of the theatre for a while. Though he was nominally at work on Usher, and still more nominally on The Devil in the Belfry, there is scant evidence of even fragmentary progress on either of them. By the end of the year he was claiming to be bored by them both. He may have tinkered with Khamma but probably composed little of substance before December. He was still trying to complete Gigues, and by October was in a position to send Durand the four-hand reduction but not yet the orchestral score, which he took another whole year to complete. But there were the usual interruptions. In June, a month after Saint Sébastien, he went to Turin to conduct his Après-midi d’un faune and Ibéria and endured the humiliation of completely failing to manage the Scala orchestra and having to hand the baton to the young Vittorio Gui, who rehearsed the programme efficiently, then handed the baton back to him for the concert. ‘At bottom,’ he grumbled to Durand, without mentioning the more painful details, ‘the music of Claude Debussy means nothing to them, and at the first opportunity they’ll go back to their Puccini, Verdi, and whatever else in the language of si.’16 Then suddenly he was ill: ‘temperature, overwork, nervous stress, and forbidden to do anything for at least a month’.17 So off they went to the seaside, this time to Houlgate, near Deauville, where he hoped – presumably against doctors’ orders – to work on the Poe operas, but in fact managed hardly anything apart from picking at the orchestration of his Clarinet Rhapsody. As before he liked the sea but not the holidaymakers, and when not holding Chouchou’s bucket and spade he spent much of his time, as he told Caplet, reading cheap novels and the Chronicles of the Crusades of Jean de Joinville.

Finally, for 1911, he accepted an invitation to go to Boston for what looked like being a well-designed and musically strong Pelléas, conducted by Caplet. But at this point Emma, herself in indifferent health, put her foot down. Like many self-centred people, Debussy disliked confrontations, preferring to go his own way quietly and avoid the accusing gaze; he took it for granted that ‘people [l’on] would help me with all possible tenderness’. But ‘to detail for you all my arguments, my daily battles, was going to be so painful for me that I’ve put off writing to you for as long as I could! The worst of it is that I’ve got lost in all this, and have forgotten that calm egoism that is such an admirable strength, and that one knows so well how to use. In the end I’m extremely demoralised.’19

It was at this low point, however, that he set to work on what would in due course become a second book of piano Préludes. They came less quickly than the first book, were themselves interrupted by other commissions, and are mostly undated. But at least two of them, Brouillards and Feuilles mortes, the first two in the eventual book of twelve, seem to have existed in draft by Christmas 1911, and there were probably more by the spring, since no other music is known to have been written by him in the early months of 1912 after Khamma in January. At that point a series of theatrical projects once again intervened, and the piano pieces were probably put to one side until the autumn, and the volume completed only in the early part of 1913.

The first partial interruption came in March from the poet Charles Morice, an old acquaintance from the Mallarmé Tuesdays, in the form of a project for a ballet to be danced to settings of poems by Verlaine under the sinister title Crimen amoris (‘Crime of love’). Exactly what this was about is uncertain, because Morice’s scenario has not survived; but Orledge speculates that it was at least partly based on the poem of that name in Verlaine’s Jadis et naguère, in which the most beautiful of the fallen angels, a boy of sixteen, rebels against the diabolical goings-on in the silk and gold palace of Ecbatane and sets fire to it, killing himself and all his fellow devils and returning the land to ‘the clement God who will guard us from evil’. Debussy was intrigued by this idea and contributed some detailed suggestions for the staging that seem to prove he had every intention of composing the music. There was even a contract, dated May, and André Messager expressed interest in programming the work at the Opéra. But for some reason Debussy turned against the Morice scenario, recruited Laloy to supply an alternative version under the new-old title, Fêtes galantes, and eventually (in 1914 or 1915) sketched a page or two of music before abandoning the whole thing, as he abandoned every other sung theatrical project with a libretto not devised by himself.20

In this same March of 1912, rehearsals began in Monte Carlo for the Ballets Russes staging of L’Après-midi d’un faune with choreography by Nijinsky. Debussy must have given his permission for this somewhat improbable adaptation, but he seems not to have been involved in it in any way, did not advise on (or object to) the treatment of his music, and in fact attended no rehearsal before the dress rehearsal on 28 May, having been constantly told by Nijinsky, ‘It’s too soon, come tomorrow.’

Nijinsky had in fact probably been inspired by Egyptian vases, but he had also, with Diaghilev, visited the eurhythmicist Émile Jaques-Dalcroze at his school in Hellerau (Dresden) and had tried to adapt Dalcrozian rhythmic exercises to the needs of a stage choreography, more in the character of gesture than that of actual rhythm – hence the at the time widely held view that the movement of the dance contradicted that of the music, even though the whole point of Dalcroze was to match each gesture to the music. But Debussy was slow to express himself publicly on the subject because by the time he saw the ballet, or very soon afterwards, he was engaged in discussions with Diaghilev and Nijinsky about a new ballet commission, which, no doubt mainly for financial reasons, he was reluctant to put at risk.

This new idea was a disguised version of a sexual fantasy that Nijinsky later attributed to Diaghilev. ‘Jeux’, he wrote in his diary,

is the life of which Diaghilev dreamed. He wanted to have two boys as lovers. He often told me so but I refused. Diaghilev wanted to make love to two boys at the same time, and wanted these boys to make love to him. In the ballet the two girls represent the two boys and the young man is Diaghilev. I changed the characters, as love between three men could not be represented on the stage. I wanted people to feel as disgusted with the idea of evil love as I did, but I could not finish the ballet. Debussy did not like the subject either, but he was paid 10,000 gold francs for this ballet and therefore had to finish it.22

In the ballet, the Jeux (‘Games’) were played out on or near a tennis court, but they were not exactly tennis. When the curtain went up, a tennis ball would bounce on to the stage and a young man in tennis kit would leap across brandishing a racket. But then two young girls would appear, the young man would emerge from the shrubbery, and there would follow a series of flirtations, between the young man and the girls in turn, then between all three together, culminating in a triple kiss of a disturbingly ecstatic intensity, at which point a tennis ball would again bounce on to the stage like a passing policeman, and the ‘players’ would flee. The original idea – Diaghilev’s or Nijinsky’s – was for the ballet to end with an aeroplane crash-landing on the stage, but it looks as if the composer objected to this modish concept and was allowed to suppress it.23

Debussy was presumably as yet unaware of these details when, on 18 June, he contracted with Diaghilev to compose the music, but he certainly knew that he would be on his mettle to compete with other Ballets Russes commissions of the hour. On the 9th they had all attended the premiere of Ravel’s brilliant Daphnis et Chloë, a substantial, hour-long ballet much concerned with flirtation and abduction. The next day Debussy had been at Laloy’s house to hear Stravinsky play the completed parts of his new ballet about a primitive fertility sacrifice, The Rite of Spring, and been bowled over by it.24 ‘I still preserve the memory’, he told Stravinsky five months later, ‘of the performance of your Rite of Spring at Laloy’s … It haunts me like a beautiful nightmare and I try in vain to retrieve the terrifying impression it made.’25 Clearly his own ballet would not have much in common with Stravinsky’s, perhaps slightly more with Ravel’s. Above all it had to be written almost as speedily as Le Martyre de Saint Sébastien had been. Diaghilev wanted a complete piano score by the end of August and the orchestral score by the end of March 1913.

Quite apart from the plot, such as it was, Nijinsky had new ideas about the actual dancing. Diaghilev conveyed to Debussy his view of the work as pure dance:

Debussy must have read this letter carefully, since his score reflects most of its indications closely. His Jeux is, precisely, a scherzo-waltz, and it cultivates the lightness and elegance that Diaghilev emphasised in his remarks about dancing on point. Having completed the piano score, and as he contemplated the orchestration, he told Caplet that ‘I’ll have to find an orchestra “without feet” for this music. But don’t suppose that I’m thinking of an orchestra made up exclusively of amputees! No! I’m thinking of that orchestral colour that seems lit from behind and of which there are such marvellous examples in Parsifal!’27 It would mean concentrated work to achieve the necessary refinement and precision, and there would be no seaside that year for Chouchou. Now nearly seven, she had written a song, Debussy told Durand, announcing that ‘the Sea is cross at not having had a visit from Mr and Mrs Debussy and their charming little daughter’.

I used up a great deal of eloquence persuading her that this year the Sea has gone out so far that they despair of ever finding it again! That didn’t go down at all well, and there are times when it’s very delicate being a father.28

Delicate, too, being a ballet composer. His orchestra without feet announces itself so softly in the opening bars of Jeux that one wonders how much of it was audible as the curtain rose on that first night of the first ballet season in the new Théâtre des Champs-Élysées, 15 May 1913.29 Not until bar 70, some time after the curtain has risen on ‘an empty park’ (in Bakst’s atmospheric setting a densely tree-lined London square), does the dynamic level rise momentarily above piano, in honour of the bouncing tennis ball, and for the rest of the twenty-minute score forte markings of any category remain the exception and usually of brief duration until the closing pages, where (as in Act 4 of Pelléas) the repressed passion explodes in a series of (comparatively) violent outbursts. Debussy’s prevailing technique, which would so impress avant-garde composers half a century later, is to distribute tiny gestures round the orchestra like fragments of overheard conversation or like simulacra of the unspoken, barely expressed, only half-recognised passions that fleetingly develop between the three dancers. Every detail is touched in with a fine brush. Snatches of decorative arabesque melody flit from instrument to instrument; the exquisitely complex dissonances of Le Martyre de Saint Sébastien acquire here a transparency and airiness like the brush-strokes in an Impressionist watercolour. This is the Parsifal effect in practice. Debussy was perhaps thinking of the way the solo oboe floats on a wave of string triplets in the Good Friday Music, or the soft orchestration for full woodwind in the Grail scene of the first act, or in general Wagner’s method of scoring for full orchestra at quiet dynamics, an effect he had first explored, less translucently, in the second act of Tristan. But Debussy’s textures are not in themselves Wagnerian, any more than his debt to Wagner in other respects, brilliantly analysed by Robin Holloway, results here in a Wagnerian kind of music, as it did to some extent in the hastily composed interludes of Pelléas et Mélisande.30

The music’s needlepoint sensuality comes, of course, from the subject matter, while being inherently, profoundly Debussyan in itself. Its delicious suggestiveness is that of the early songs and the Après-midi d’un faune, albeit expressed in a more sophisticated language. But the fact that this atmosphere impresses so strongly through the music is due to something else that Diaghilev referred to in his letter, Nijinsky’s attempt at unifying the work through a conformity of dancing style (as opposed, presumably, to the differences between male and female ballet conventions). The music, though scattered in its elements and texture, is bound together by an overriding rhythmic impulse. Even though the rhythmic components are themselves diverse, they swing along together on a shared conveyor belt, and while the melodies are in a constant state of flux, they evolve from one another as in a stream of consciousness, stray thoughts connected by personality more than logic. In this, Jeux might be seen as a more abstract continuation of La Mer, a score whose tone was set by the pictorial subject, but whose coherence depended on subtle internal linkages worked at a purely musical level. Jeux may, as Holloway insists, be about sexual pleasure, but since Diaghilev it has flourished, when it has, in the concert hall, where subject matter is more provisional, and the music is enjoyed for its innate beauty and coherence, like any symphonic work.

However much Debussy may have been prompted in the kind of music he wrote by Diaghilev’s account of Nijinsky’s ideas, Nijinsky himself seems not to have fulfilled them. By all accounts the staging and choreography were heavy-handed, worked out literally on the hoof, and largely unresponsive to the music’s delicacy and refinement. There were absurdities that provoked laughter. Nijinsky had never played tennis, though he had watched it, and his sole concessions to anything related to that particular game seem to have been the dancers’ modern, white costumes and the ‘tennis’ ball, which was in fact more the size of a football. The movement, for this flickering, feather-light score, was again Dalcrozian, not fluid but posed, and again inspired (according to one reviewer) by the figures on Greek vases.31 The real problem for Nijinsky, and in the end for Jeux itself, was that he was simultaneously preparing his production of The Rite of Spring, which would have its turbulent premiere on the corresponding subscription Thursday two weeks later. Because The Rite was a complex piece involving a large corps de ballet and a revolutionary score, while Jeux had only soloists, it was the Stravinsky that claimed most of Nijinsky’s attention. Tamara Karsavina, one of the two female dancers, described the difficulty of working with a choreographer who was also dancing the male role, and who could not or would not explain his intentions clearly:

How much worse it must have been for the other ballerina, Lyudmila Shollar, who took over from Nijinsky’s pregnant sister, Bronislava, at short notice.

As for the composer, he was tactful to those involved, but did not hold back either in print or in private. On the day of the premiere, Le Matin printed an open letter from him containing a double-edged account of the ballet and a sharp satire of Nijinsky himself as ‘a gentleman very good at arithmetic’. Later he unburdened himself in detail to Robert Godet:

This man adds up demisemiquavers with his feet, does the proof with his arms, then, suddenly struck with hemiplegia, gives the music a dirty look as it goes by. It seems this is called the ‘stylisation of gesture’ … It’s horrible! It’s even Dalcrozian, for I consider Monsieur Dalcroze one of music’s worst enemies! And can you imagine what ravages his method can cause in the soul of this young savage that is Nijinsky?33

*

Between completing the short score of Jeux in September 1912 and the first performance in May, Debussy not only composed the remaining preludes for his second volume but delivered them to Durand, who published them in April 1913. At the last minute there was some equivocation about what to include. There was a piece called ‘Tomai des éléphants’, based analphabetically on Kipling’s ‘Toomai of the Elephants’ (in the Jungle Book), which Debussy persisted with but then decided was ‘impossible as a prelude’, and had to compose a last-minute substitute – probably the penultimate prelude in the eventual set, Les Tierces alternées. What happened to ‘Tomai’, who couldn’t become a proper elephant handler until he had seen the elephants dance, is unknown, but he may possibly have found himself, in a different guise, in the ‘Pas de l’éléphant’ of Debussy’s next ballet, La Boîte à joujoux.34

In subject matter, and to some extent in musical character, the preludes of book two map quite neatly on to those of book one. Occasionally the similarities are disturbing: for instance, Bruyères (‘Heathlands’) in book two starts uncomfortably like a re-run of La Fille aux cheveux de lin, but soon diverges, both musically and technically; Général Lavine – eccentric (about the American clown, Edward Lavine) has moments of sounding like a second-hand Minstrels, but varies the music-hall imagery and exaggerates the montage so entertainingly that it would be rank pedantry to object. There is a certain recurrence of imagery. Brouillards (‘Mists’) might recall Voiles, though certainly not musically: its left-hand triads and right-hand bravura are worlds away from the steady whole-tonerie of the earlier piece. Feuilles mortes (‘Dead Leaves’) could be a sad Baudelairean epitaph to the Sons et parfums of book one: the same slow triple time, the same saturated chromatic harmony, the same transference of Nature into music. La Puerta del Vino (one of the gates of the Alhambra, which Debussy saw on a postcard) matches the Sérénade interrompue, with similar violent contrasts, there of montage, here of dynamics. Canope (a funerary urn from the ancient Egyptian city of that name) resonates as an image with Danseuses de Delphes, but could hardly be less like it musically, with its solemn tread of parallel triads, and stifled fragments of chromatic plainchant. A closer precedent for this profoundly mysterious piece is the chord parallels that open Le Martyre de Saint Sébastien: the same solemn, stately tread, and something of the same tendency to suggest strangeness by breaking the natural sequence of these essentially very plain, uncomplicated chords.35

Taken as a whole the second book breathes a different air from the first. La Cathédrale engloutie notwithstanding, consecutive triads play a bigger part in the later set, often in that disjunct character that seems to have been a discovery of Le Martyre. At times there is more than a hint of Petrushka about these parallels: compare Brouillards and the opening of the fourth tableau of Stravinsky’s ballet, or the start of Général Lavine with Petrushka’s music in the second tableau. Sometimes the similarities are more noticeable with the four-hand piano version of the ballet, which Debussy had in his possession by Easter 1912, a gift from the composer, to judge by Debussy’s letter of 13 April: ‘Thanks to you, I have spent an exquisite Easter holiday in the company of Petrushka, the terrible Moor and the delicious Ballerina.’36 In general the tendency of the second book to combine or juxtapose different keys (bitonality) might have been prompted by the Stravinsky not only of Petrushka but also of the first part of The Rite of Spring, which Debussy had heard its composer play in June 1912. But the bravura of Les Fées sont d’exquises danseuses (‘The Fairies are Exquisite Dancers’) or the last two preludes, Les Tierces alternées (‘Alternating thirds’) and Feux d’artifice (‘Fireworks’) has a purely Debussyan lightness of touch, however Stravinskyan some of their figuration and harmony might look on the page. The French composer could somehow capture the mercurial grace of the Arthur Rackham illustrations that inspired Les Fées sont d’exquises danseuses and the eighth prelude, Ondine, without lapsing into the somewhat arch fayness of the original drawings. On the other hand, he was equal to the heavier, more picaresque comedy of Dickens, an author he adored. Hommage à S. Pickwick Esq. P.P.M.P.C. is like a multi-frame picture postcard from London, starting with ‘God Save the King’ (Union Jack T-shirt) and proceeding through what might be a series of street scenes culminating in a whistled tune straight out of the mouth of Sam Weller. The British national anthem is balanced at the very end of the book by a distant fragment of the ‘Marseillaise’ breaking in on the fireworks of 14 July.

In one other respect, the appearance of the preludes (as opposed to their sound) might have been prompted by the young Russian, though we have already noted something like it in Debussy’s much earlier Nocturnes. Beginning with Petrushka, Stravinsky’s orchestral scores are often layered, in a way that forms visible strips on the page, each strip being a continuing figure or texture in that particular section of the orchestra. In a solo piano work, of course, such layering is less obvious. Nevertheless, the preludes in book two have something of the same appearance. As in the second book of Images but not at all in the first book of Preludes, Debussy sets the music out on three staves for at least part of every piece, whether or not the music’s complexity seems to require it. Often this layout has the effect of isolating a middle voice, with accompanying chords or figures handed from right hand to left around it. In particular it undermines the automatic association between the upper and lower staves and the right and left hands. In pieces like the seventh prelude, La Terrasse des audiences du clair de lune, the music spans an immense range, greater at first sight than two hands could manage. But the three-stave layout not only identifies the separate layers of the music, it also helps explain how they can be performed by a normal human being not blessed with three arms.

The mysterious title of this seventh prelude, one of Debussy’s most fascinating piano works, calls for explanation, not least because in every existing book on Debussy that I’m aware of, including the relevant volume of the Oeuvres complètes, it is explained wrongly. Debussy is said to have picked up the title from an article by the historian René Puaux in the Paris newspaper Le Temps, ‘describing the coronation festivities of George V as Emperor of India’ (the great Durbar of December 1911). The text is usually quoted. It refers to ‘la salle de la victoire, la salle du plaisir, les jardins des sultanes, la terrasse des audiences du clair de lune, le couloir des reines’. Curiously enough this list of features is quoted correctly, and it is indeed from an article by Puaux. But it has nothing to do with the Durbar or with George V. It figures in the issue of Le Temps for 10 August 1912, and is a beautiful account of a visit Puaux made to the abandoned city of Amber, near Jaipur, some hundred and fifty miles from Delhi.

Debussy clearly read this article and was struck by the particular phrase he chose for his title. But it is hard to believe that there was not, behind that choice, a response to the whole description. Puaux relates his approach to the deserted palaces, not like the prince in The Sleeping Beauty, cutting his way through hedges of wild brambles, but like a normal visitor arriving at a well-maintained but completely uninhabited city ‘by way of a broad, paved ramp, into the imposing interior courtyard, which must, it seems, have been crowded only yesterday with chariots, elephants, ostlers, mahouts’.

And yet silence has reigned here for two centuries. The young Hindu functionary they have given me as a guide must have understood my desire to explore like the prince in the fairy-tale. He follows me without a word, and once he has taken off his shoes at the foot of the stairway of the Diwan-i’Am [the Hall of Public Audience], I no longer even hear him walking behind me. When we cross some room richer in decor than the others and whose former purpose I should like to know, I turn to him and, in an undertone, very simply, without architectural nomenclature, without a profusion of dates or details, he tells me: the hall of victory, the hall of pleasure, the gardens of the sultans, the terrace of the moonlight audiences, the corridor of the queens.

After contemplating the interior courtyard of the women’s quarter ‘where twelve young maharanis lent themselves each day to the whim of their master’, Puaux reflects:

Might there have been some remote connection in Debussy’s mind between Puaux arriving at the deserted city and Poe’s narrator arriving at the House of Usher? The two destinations have one thing only in common, but it is a big thing: they are actually or spiritually abandoned places, shipwrecks of history. Both narrators find their imagination exercised by the scene before their eyes.

What was it [Poe asks] that so unnerved me in the contemplation of the House of Usher? It was a mystery all insoluble; nor could I grapple with the shadowy fancies that crowded upon me as I pondered. I was forced to fall back upon the unsatisfactory conclusion, that while, beyond doubt, there are combinations of very simple natural objects which have the power of thus affecting us, still the analysis of this power lies among considerations beyond our depth.

 

What use [asks Puaux] would more precise information have, since it is my imagination that has to bring this silence to life? What use the extracts from learned works by [James] Fergusson, since my memory will only retain, as with every impression of life, that which has left the profoundest impression on my eyes and my heart?

There is no obvious sense in which Debussy’s piece represents Puaux’s Amber, in the way La Cathédrale engloutie represents the legend of Ys. It might, though, be regarded as a response to the activity of the imagination in the face of the beautiful but inexplicable relics of history. At first the eye sees and the ear hears only debris, musical fragments, at least seven apparently unrelated ‘objects’ in the first fourteen bars. Then gradually some kind of picture materialises, with a distinct shape, yet no clear connection with the fragments that preceded it. At the end, the visitor turns away, picking up a fragment or two, souvenirs, as he leaves, moved by what he has seen, but still in truth mystified as to what precisely it was.

Of all Debussy’s piano pieces, this is one of the most original in form. Jann Pasler has shown how the seemingly unrelated elements are linked by carefully planned voice leadings, by the way in which the melodic line of one will spill into the line of the next, or the way in which apparently random repeated bars serve a pivotal function between superficially disparate types of music.38 The technique is a kind of montage, but with a carefully devised hidden flow. In an uncomprehending performance, it barely hangs together. But Pasler has demonstrated how a grasp of the linkages can make sense of the continuities. Every detail is arresting, and to claim that anything is irrelevant or unconnected is merely to admit a failure to penetrate the mystery. It is precisely the enigma of great art that we know it is great even when we cannot put our finger on what distinguishes it from the trivial. About La Terrasse des audiences du clair de lune there is no hint of the trivial, but what it adds up to is as difficult to know as the phenomenon it purports to describe.