Writer’s notes

(The following notes appeared in the program for the 2009 Berkeley production.)

About the opera

When Jacques Offenbach died in 1880, he left his last opera, Les Contes d’Hoffman (The Tales of Hoffmann), unfinished. He had finished writing the music for the first two tales, about Olympia and Antonia, but the Giulietta act, as well as the prologue and epilogue, were incomplete.

Another composer, Ernest Guiraud, was engaged to finish the opera. For reasons lost to us, the decision was made to drop the Giulietta act entirely from the original production. Perhaps Offenbach had not completed enough of the act, perhaps Guiraud’s completion of it was clumsy, perhaps the opera overall was running much too long (which was the official explanation at the time, though for various reasons not likely to be the whole story), perhaps a combination of all these reasons.

In any case, the act was cut, and some of its numbers were moved to other acts. (The strong objections of librettist Jules Barbier were considered about as seriously as the strong objections of opera librettists usually are.) The barcarole was inserted into the Antonia act (which was then moved to Venice to justify its inclusion). Hoffmann sang his passionate love aria (“How long I’ve yearned” in my version) not to Giulietta but to his Muse in the epilogue. Dapertutto’s aria was given — with new words — to Coppelius in the first act, where it replaced a longer trio.

Some of these changes have become traditional and have persisted in most productions even to the present. The version of the opera most commonly performed today — the Choudens edition — was arranged around 1904, and is only one of many versions that were tried.

It is impossible to be completely certain about the opera’s shape at the time of Offenbach’s death, because many of his manuscripts for the opera were lost in a disastrous theater fire when the opera was performed in Vienna in December 1881. Since the Second World War, however, many of Offenbach’s sketches and preliminary manuscripts have been found, and we have a much better idea of his and Barbier’s intentions for the opera. Even with these new discoveries, though, we can’t be sure what shape the opera would have taken had Offenbach lived. It was his practice to resist making changes or cuts during rehearsal, preferring instead to wait until he had seen the audience’s reaction on opening night.

About this adaptation

Two editions have been based on these discovered manuscripts — one by Fritz Oeser in the 1970s and one by Michael Kaye in the 1990s. The opera as revealed in these editions is a brilliant, many-layered, and psychologically profound masterpiece, and makes the Choudens edition seem a clumsy patchwork by comparison. Yet both of these recent versions are, I think, too long, and contain dull stretches that it’s hard to imagine Offenbach would have been content with. My first goal, then, in writing this new adaptation was to restore the opera to the shape that Barbier and Offenbach planned, while at the same time tightening the story and getting everyone home to bed well before midnight.

I’ve used a mix of sung and spoken dialogue to link the musical numbers. Offenbach himself seems to have planned that there would be two versions of his opera, one with spoken dialogue for performance in small theaters like the Opéra-Comique (where the opera premiered), and one with recitative (sung dialogue between numbers) for houses like the Paris Opéra, which are too large for unenhanced speech to be effective. Most of the recitative in the opera was the work of Guiraud, and it often plods. Spoken dialogue works wonderfully in the intimacy of the Julia Morgan Theater, and as I wrote, I found there were many places where as much as five or six minutes of bland recitative could be replaced by a minute or two of speech. And as I wrote, I found that the unpredictable changes between speech and song created a fluid, dreamlike atmosphere that suited the story, a mood where reality and hallucination, order and chaos, the rational and the irrational, could blend into each other with disturbing ease. To the same end, I experimented quite a bit in this libretto with the juxtaposition of rhymed and unrhymed verse, a device I’ve made use of several times since.

If you’re only familiar with the usual grand opera version, you’ll hear some music in this production that is new to you, and you won’t hear some that you may be used to. In the Choudens edition, the Muse appears only as a sort of deus ex machina in the epilogue, but Barbier and Offenbach meant her to start the opera as well, and we’ve followed this. We need the Muse at the beginning if we’re to understand the opera as a single story and not an anthology of three independent one-act operas.

In the tale of Olympia, we’ve replaced Coppelius’s “eyes” aria with the trio Offenbach actually wrote for this scene (and the music for the “eyes” aria is returned to Dapertutto later in the opera, where Offenbach intended it). This aria has always seemed to me to rush by too quickly to make adequately the several points it needs to; the longer trio works better dramatically. And musically, it’s a delight; two cast members have even told me during rehearsals that the trio has become their favorite number in the opera. See if you agree.

In the tale of Antonia, we’ve given Nicklaus the “violin” aria that Offenbach wrote for him. The epilogue now ends with the soaring ensemble Offenbach intended. I’ll talk about the tale of Giulietta in just a moment.

The three tales are in the order Barbier and Offenbach intended. The Antonia act has often been moved to last place, because it has the strongest ending musically. This makes some sense if the opera is thought of as three separate stories. But there is a psychological progression in Hoffmann that runs through the tales, from self-absorbed naiveté to youthful zeal to disillusionment, that becomes significant when the opera is looked at as a single story.

As I researched the opera, a second goal presented itself to me. Barbier’s libretto, as brilliantly conceived as it is, was widely criticized at the time for distorting E.T.A. Hoffmann’s own stories, and when I read the stories on which the opera is based, I had to agree. It was Hoffmann’s method to make the fantastical and the horrific spring from situations that seemed on the surface to be ordinary and mundane. Barbier, though, often wrote in the style of the popular melodramas of the day, and there are places in his libretto where you nearly expect the villain to twirl his mustache and tie the heroine to the railroad tracks.

Here and there, then, I veered from Barbier and stuck closer to Hoffmann. The tale of Olympia, for example, is based on the story “The Sandman”. In that story, the sinister Coppelius is a vivid memory from the narrator’s childhood, but the horrific events involving Olympia are caused by an eccentric and seemingly innocuous, even comical, Italian immigrant named Coppola. It is Coppola, then, and not Coppelius, who appears in my version. In the tale of Antonia, I added a few details from Hoffmann’s story “Councilor Krespel” that Barbier didn’t use. In the prologue and epilogue, I’ve made Lindorf the sort of smug, comfortable bourgeois businessman that Hoffmann had a particular hatred for (more about that a little later), the better to contrast with the satanic mastermind he becomes in Hoffmann’s fantasies.

The tale of Giulietta was the most difficult to know what to do with. The tale as presented in the Choudens edition makes little sense; motivations are obscure and the plot is poorly constructed. Kaye’s version, which includes some previously unknown material discovered in 1985, is presumably the closest to what Barbier and Offenbach wrote. Yet once I’d read Hoffmann’s stunning “A New Year’s Eve Adventure”, on which the act is based (and you’ve just got to read it!), even Barbier’s own version seemed a muddled condensation. Hoffmann’s story, though complex, is clear every step of the way, and Hoffmann’s own ending — the main character, in effect, loses his soul forever — felt even darker than Barbier’s and perfectly right as the climax to the final tale. Here, then, I’ve rearranged Barbier’s order of events to clarify the story, and used more of Hoffmann’s ending.

The bittersweet music we’re using for Giulietta’s aria (“Quick the night flies”) is originally from Offenbach’s opera Die Rheinnixen. Oeser used it for this purpose in his edition of Hoffmann. The sprightly aria Offenbach actually wrote for Giulietta was among the 1985 discoveries incorporated into the Kaye edition, but the more I listened to both, the more I felt that Oeser’s choice for his version of the tale was also the better choice for mine.

We’ve chosen not to use either Dapertutto’s aria “Scintille, diamant” or the sextet with chorus from the Choudens version of the Giulietta act. Neither is by Offenbach; both were written for a 1904 production to fill out an act to which much of the music had been lost. Dapertutto’s “diamond” aria, beautiful as it is, is longer and weightier than Dapertutto’s simple statement warrants; the brisk, sinister aria that Offenbach wrote for this spot (usually given to Coppelius in the first act as I mentioned above) is more effective in the context of the tale. And the sextet, though a popular and impressive piece of music, is just too static — neither advancing the story nor illuminating the characters — to fit comfortably into this version.

Two notions in this adaption are (as far as I can remember) entirely my own inventions: first, that Hoffmann’s three tales began as nightmares he had after a bout of heavy drinking, and second, that Hoffmann’s companions take on roles and act out the tales as he tells them. The first is not part of Barbier’s libretto, but I think it’s strongly hinted at, and as I wrote I drew often on my own experiences with dream analysis. The second gave us a fresh take on the opera, as well as way to stage it without the need (or the budget) for four sets and the resulting long scene changes.

About E.T.A. Hoffmann

E.T.A. Hoffmann died penniless in 1822, but in Paris in the late nineteenth century his stories were tremendously popular — more popular perhaps than they had ever been in his native Germany. The veiled autobiographical nature of Hoffmann’s writing would have been well known to Offenbach’s audiences, as was Hoffmann’s unrequited obsession with a teenaged music student of his named Julia Marc. Julia was’t the only love in Hoffmann’s life, but she was the best known, and perhaps the one who caused him the greatest heartbreak.

In all, there were three great loves in Hoffmann’s life (none of them, alas, his own wife), and all three were singers. The first was Dora Hatt, the young wife of a middle-aged wine merchant. She and her husband were boarders with Hoffmann’s family, and she became one of Hoffmann’s first music students, when he was just eighteen. They fell in love, and their difficult affair lasted three years. It was she who first inspired Hoffmann to begin composing.

Hoffmann had been married nearly ten years, and had seen the birth and death of a daughter, when Julia Marc came to him for singing lessons. Hoffmann was soon obsessed with her, though the affair was never consummated — and Julia’s mother, suspicious of Hoffmann’s intentions, quickly arranged Julia’s engagement to a Hamburg banker. Julia appears in many fictional guises throughout Hoffmann’s work; sometimes he doesn’t even disguise her name. In his story “Don Juan”, Julia is the young soprano who sings Donna Anna as Hoffmann listens from a private box. In “A New Year’s Eve Adventure,” she and her husband (of whom Hoffmann was obsessively jealous) appear with their own names, and are also mirrored in the characters of Giulietta and Dapertutto.

Hoffmann was 40 when he met his third great love. Soprano Johanna Eunike played the title role in his opera Undine in Berlin in 1816. (The orchestral parts to Undine, curiously, were destroyed in a theater fire just a year after its premiere — a fate eerily similar to that of Les Contes d’Hoffmann. Fortunately, other copies of Undine survived the fire.)

It’s tempting to try to identify Olympia, Antonia, and Giulietta all as Julia Marc, or else as Dora Hatt, Julia Marc, and Johanna Eunike in some order. But reality and writers’ imaginations are not so neatly organized. Giulietta is certainly at least in part a grotesque portrait of Julia Marc, but there are probably aspects of all three women in Olympia and Antonia. In any case, for dramatic purposes Barbier condensed all three real-life women into one fictional, archetypal young singer named Stella, who appears in Hoffmann’s imagination in various guises.

In fashioning his libretto thus, Barbier falsified Hoffmann’s actual history but revealed a psychological truth. Hoffmann fell in love time and again not with real women but with an illusion of his own, an idealization of the female singer that existed in his own mind but that he thought he saw in others. The psychological notion of “projection” is a commonplace now, but it was a shrewd insight for Barbier’s and Offenbach’s own time. The tales of Olympia, Antonia, and Giulietta are ultimately the same story thrice told, the story of Hoffmann’s disappointed love for an ideal that existed only in his own fertile, troubled imagination.

— David Scott Marley