Can a masterpiece still emerge from near-total disagreement between its creators? There have been many, including serious Stravinskians, to question whether Persephone amounts to anything so coherent. It was certainly a daunting hybrid to undertake, even in the context of Stravinsky's questing theatricality. Ida Rubinstein was of the breed of larger-than-life performer-patrons which (sadly? mercifully?) no longer exists. Imperious, fantastical and fabulously rich, she had made an instant impact dancing in the first Paris seasons of Diaghilev's Ballets Russes in 1909/1910. Then she had broken away to commission and star in D'Annunzio's kinky miracle play Le martyre de Saint Sebastien in 1911 with a score from Debussy, no less, and designs by Bakst - so inaugurating a career of mime and melodrame that was to culminate in her assumption of Joan of Arc in 1938 to music of Honegger and words from Paul Claudel.
Stravinsky had at least been interested enough in her artistic aims - or her money - to risk the wrath of Diaghilev by accepting her 1928 commission for the Tchaikovsky-based ballet, Le Baiser de la Fee. But now the great impresario was dead and Madame Rubinstein was no Diaghilev in conjuring a unity of taste from her collaborators. The commission was for a one-act presentation of the Persephone myth featuring a danced and spoken title role for herself and based upon an early, somewhat Ninetiesish poem Gide had derived from the Homeric Hymns. In classical mythology, the rape of Persephone by Pluto, and her subsequent rescue from Hades, were held to account for the rotation of winter and spring. However, Gide had Persephone voluntarily sacrificing herself out of pity for the lost souls in the underworld, thus insinuating a Christian gloss which the Russian Orthodox inflections of Stravinsky's choral writing, even in his most supposedly neo- classical phase, were bound to enhance.
Unfortunately, in elaborating his text for the theatre, Gide not only imposed a pretty leisurely pace on the action, but inserted sub-Wagnerian suggestions of how the music should go which were bound to irritate so anti-romantic a composer. Stravinsky, in turn, had strong views on the unifying of words and music derived from Slavic folk-singing in which syllables were treated as rhythmic units independent of spoken quantity or stress, and which he had already imposed enthusiastically upon the Russian words of Les Noces and the Latin of Oedipus Rex. The fact that Persephone was the first large-scale French text he had ever tackled seems scarcely to have restrained him, and by the time of the private preview in the salon of the Princesse Edmond de Polignac, Gide could be observed bridling more and more at each outrage upon his Parnassian prosody.
Nor was this the only hitch in a spectacle that was ultimately to encompass a speaking and miming protagonist, a tenor narrator, a troupe of dancers, substantial choral forces including a children's choir, and a large symphony orchestra. Stravinsky's hope that his artist son Theodore might design the premiere was pre-empted when Jacques Copeau, the producer, commissioned a buff-coloured basilica-like set from the little-known painter, Andre Barsacq. In the end, Stravinsky seems to have approved of little about the production except the choreography of Kurt Jooss. But the whole staging seems to have been strangely little documented; nor did its apparent failure encourage further performances until Frederick Ashton brought the work back to theatrical life with his striking new choreography for Covent Garden in 1961. More recently it has been taken up by a number of concert conductors, notably David Atherton who repeats his well-considered interpretation at the Proms next Wednesday. And, at last, after years in which the only recording in the catalogue was Stravinsky's own, there are two new ones to choose from.
It is certainly high time that one of Stravinsky's most subtle, yet least valued and studied achievements came into its own. True, the stylistic surface of Persephone can sound as inconsistent as any work of his magpie middle period. While the more elevated passages - the tenor narrations of the high priest Eumolpus, the sublime concluding rite of Persephone's dedication to her annual round - may sound like a combination of Gluck in his Elysian Fields mode and the Russian Easter liturgy, the nymphs in the opening scene celebrate 'the first morning of the world' to what sounds like nothing so much as a highly stylised can-can.
Yet somehow Stravinsky was able to blend and unify his sources in a gentle radiance rare in his output. Arguing not only from the subject matter but from the score's continual emphasis upon high voices, Robert Craft has suggested it may have been consciously conceived as a feminine complement to the marmoreal masculinity of Oedipus Rex of seven years before. Again, despite Gide's antipathy, it is evident Stravinsky made an exceptional effort to embrace elements of French musical tradition: pointillist touches of piano and piccolo, flowings of flutes and harps, blandishments of harmony in which some have even detected echoes of Massenet. In fact, long stretches of Persephone are among the most diatonic, the most purely 'white note' in all Stravinsky, offering a thesaurus of ways in which the sound of common chords can be freshened through novel spacing and juxtaposition.
Admittedly sustaining the radiance and overcoming the patchwork effect of the many short sections forced upon Stravinsky by Gide apparently depends upon a preponderance of moderate to slow tempi - which is doubtless why some have found the work bland and lacking the usual electricity. But here Stravinsky's own late, dogged and by no means brilliantly engineered recording may be part of the problem, precisely because he sticks for the most part to the metronome marks in the score. Craft, for one, is convinced that whatever relevance these markings may have had to the now- unreconstructable first staging, they are often impossibly slow for a convincing articulation of the music. His own new disc - part of a magisterial project-in-progress to conduct Stravinsky's complete works for the American record label MusicMasters - is often harder-driven even than Kent Nagano's recent version on Virgin Classics. Yet so crisp is the singing and playing that, although the work's charm is undoubtedly squeezed, its disparities are often overriden in a longer-term dynamism, reminding the listener that Elliott Carter once aptly described Persephone as 'the humanistic rite of spring'.
How to explain, then, that Atherton has often adopted even slower tempi than Stravinsky? Here the thinking seems to be that if each section is monumentalised the effect will no longer seem patchwork, and the approach certainly serves to lay bare the austere ritual beneath the sensuous surface. No doubt there remains something ambiguous about a score that seems to go better either at faster or slower speeds than its composer seemingly intended. But doubtless that ambiguity is also part of the explanation why, quite exceptionally in Stravinsky, the music lingers in the memory most of all for its sonorous poetry.
Royal Albert Hall, 7.30pm, 11 August: box office 071-589 8212; also live on Radio 3. Persephone, conducted by Robert Craft: MusicMasters 01612-67032-2Reuse content