First, however, Debussy turned to another 1839 story, “The Devil in the Belfry,” producing a scenario and some sketches in August 1903. “The Fall of the House of Usher,” to which he turned in 1908, was also supposed to be entirely different from Maeterlinck’s Pelléas et Mélisande, despite being set in a destiny-laden ancestral pile with subterranean vaults and a pale-suffering maiden, and still having swift recitative-like vocal lines with narrow intervals and a sizeable orchestra as the main vehicle for continuity.
Debussy’s other problem lay in fashioning a suitable libretto from a story that had hardly any dialogue, took place over several weeks, and whose potential force for evil (the family doctor) was only mentioned in passing. Understandably, Debussy began with a long, lamenting monologue for the main protagonist, Roderick Usher, and it was not until he had finished a second libretto in June 1910 that he saw a way of converting his three scenes into two and beginning with a more dramatic interchange between Roderick’s saner childhood friend (L’Ami), recently summoned to the crumbling house, and his sinister adversary (Le Médecin).
Meanwhile, Debussy had produced a version of the final prelude (as a late birthday gift for his second wife, Emma, in 1909) and a few other, fairly rudimentary sketches based on Libretto B. As a demonstration of the problems involved, the first bar of this prelude contains both the only dynamic Debussy ever applied to Usher (pp) and its only tempo indication (Lent et douloureux). Even in the later sketches from Libretto C (1915–16), which embraced the whole of scene one and the start of scene two (Roderick’s monologue), together with much of the final melodrama (Poe’s medieval Mad Trist of Sir Launcelot Canning), there are invariably no key or time signatures, phrase marks, or scoring indications. Indeed, Debussy only put in accidentals that he thought he might forget later and, probably for the same reason, odd references from his text. So, apart from reassembling all the pages Emma gave away to musical friends after his death, there was some hazardous editing to do for the Durand Oeuvres complètes volume that appeared in 2006.
However, the in-depth detective work preparatory to this gave me confidence that I could complete the opera by reusing Debussy’s material in similar dramatic situations (like the meticulously spaced and scored “Usher chord”—C major with an F-sharp as a solo violin harmonic eerily suspended above), or the high sixteenth-note figure representing the “black wings”’ of fate (in the form of Poe’s favorite raven) which also dates from 1909–10, and is the only fast(ish) music Debussy produced for this doom-laden opera of psychological deterioration. To restore the balance, I needed to compose a nightmare scherzo for the passage where Roderick describes yet another nuit blanche (“sleepless night”)in the middle of the second scene. So, although my “original” contribution appears to be some 52% of the opera, in reality it consists of this scherzo, a few linking passages, some uncharacteristically happy and lyrical moments, and the final climax leading to the blood-spattered appearance of Lady Madeline as she comes to claim her brother from her premature burial (by the doctor) in the vaults directly below Roderick’s study.
So, while Debussy told his friend Paul Dukas in August 1916 that “destiny should allow me to finish it [Usher], for I shall not wish to rely entirely on Pelléas for the harsh judgment of future generations,” why did this not happen? The more likely reasons include his spending too long on the libretto and then feeling compelled to write more saleable music in 1915–17 to help reduce his by then substantial debt to his publisher, Jacques Durand. And by 1917 the outer movements of the Violin Sonata show that he was too debilitated by his rectal cancer to function consistently at his best, something that Dukas had spotted earlier, even if he blamed it on Debussy’s wartime “editorial work,” which “has impaired that marvelous musical intuition which directed his energies in the past.” But principally, I believe that Debussy saw Usher as a means of escape from his growing financial and domestic problems. We find him abandoning other work (Gigues) for Poe’s alluring world in 1909 and coming straight back to him after reworking the end of Jeux for Diaghilev in September 1912. And if he identified personally with the Hamlet-like Roderick Usher in his neurasthenic indecisiveness, then it was the ailing Lady Madeline who haunted him in his study, just as Mélisande had done. As he told Durand as early as 1908, “there are moments when I lose awareness of my surroundings; and if Roderick Usher’s sister entered my house I wouldn’t be at all surprised.” So a project quickly turned into an obsession that stretched romantically and enticingly into the foreseeable future, but was destined to remain incomplete for another century.
In reality, we know even less about Lady Madeline than we do about Mélisande and she must have one of the briefest parts in operatic history. In 1909, she sang three verses of Poe’s “The Haunted Palace”(originally sung by Roderick, but transferred to his sister in Debussy’s libretto), but in 1916 she was given only the opening verse. As the only female, the conniving doctor lusts after her, while an incestuous relationship is suggested in Roderick’s “soeur trop aimée.” (“sister much too loved”). But it is not until he hears her scream from below that Roderick is certain she has clawed her way out of her coffin and is on her vengeful path towards Debussy’s own conception of their rapid and terrifying denouement.