Small picture of Donizetti







Donizetti's Élisabeth, ou la fille de l'exilé

by Will Crutchfield.  

These extensive introductory notes were written for the first performance of this reconstructed French version of Donizetti's opera at the Caramoor Festival, New York, July 17 & 19, 2003, conducted by Will Crutchfield.

Donizetti composed Élisabeth in Paris during 1839-40, basing the work in part on a Neapolitan opera from his apprentice years that he had long wanted to revisit (see synopsis and detailed notes below). When the Paris production fell through for reasons as yet unknown, he adapted the score to a makeshift Italian text and sent it off to London, where he had repeatedly been offered engagements. Soon afterwards, however, his final illness struck, and the London score - transferred after a fire to a basement at the Covent Garden theatre - remained forgotten until it was rediscovered in 1984. A reconstruction of the London/Italian version was performed by Covent Garden in 1997. Further archival research in Paris enabled the reconstruction of the fuller original version of the opera (in French), which now receives its world premiere 163 years after its composition.

The Christian Science Monitor (June 27) and New York Press (July 17) published preview articles on the discovery of the Élisabeth score.  The performance was reviewed by the New York Times (July 19), The Journal News (July 19),  and Classics Today.



The Reconstruction at a glance

Notes in Depth


Donizetti in the Thirties


What did Donizetti write and when did he write it?

What happened to the score?

Rediscovery and Reconstruction

The Opera

Filling the gaps and recitatives


                  Appendix 1: Otto Mesi and Élisabeth

                  Appendix 2: Élisabeth and Elisabetta



Act One takes place in and around the cottage of Count Stanislaus Potosky in Saimka, the remote Siberian village to which he has been exiled on account of corrupt accusations by the Grand Marshal of the Kremlin and a conspiratorial Colonel, Ivan.

The men of the village return empty-handed from a luckless hunt. But to the alarm of the Countess, her daughter Élisabeth, who went hunting with the men, has not returned with them. Presently Élisabeth arrives, having broken off from the party and brought back enough game for a feast. Amidst congratulations from friends and family, she sings of her secretly cherished plan: she dreams of crossing the wilderness to plead with the Czar in Moscow for the release of her unjustly accused father.

Meanwhile Marie, a servant who has followed the Potosky family into exile, discloses that she has observed in the distance an Imperial courier arriving at the residence of the provincial governor. She intends to go and question the messenger, hoping for news of her faraway son Michel. To her astonishment, the courier turns out to be Michel himself, who has taken the job so that he could at last visit his mother in Siberia. Élisabeth sees in his arrival her golden opportunity. Despite his warnings of the journey's terrible rigors, she persuades Michel to take her along on his departure. The provincial governor orders Michel to the Siberian capital Tobolsk that very day, and so in great haste and secrecy, Élisabeth sets out on her adventure.

Act Two is set by the banks of the river Kama, between Saimka and Moscow. Ivan, the colonel who conspired against Potosky, has been banished in his turn, and now earns his keep as a ferryman on the river. He is in mourning for the death of his own beloved daughter, and is consumed with remorse for his past crimes. Ivan is startled by the arrival at the riverbank of a solitary female figure: it is Élisabeth, who has continued her journey on foot after Michel's subsequent orders sent him on a different route to Odessa. Ivan offers her food and shelter, and in the course of their exchange he discovers her identity. Stricken with guilt, he confesses his role in her father's exile, begs her forgiveness by the tomb of his daughter, and offers a letter exonerating Potosky.

In the meantime, a band of Tartar thieves approaches and demands that Ivan ferry them quickly across the river, which is threatening to flood its banks. Ivan hides Élisabeth in his hut, but she has been observed, and the Tartars insist that she be handed over to them. When Ivan refuses, they threaten to kill him, whereupon Élisabeth rushes from the hut and boldly interposes herself between the attackers and their victim. They are so struck by her bravery that they offer their help to the intrepid girl, and depart in peace without molesting her.

Suddenly, however, the flood breaks forth and sweeps away Ivan's boat. Élisabeth sees no escape from certain death, but miraculously, the wooden cross marking the tomb is borne up on the waters; she seizes it and is carried safely on towards her goal.

Act Three is set in an inn on the outskirts of  Moscow, managed by Nizza, Michel's fiancée. The new Czar is expected to pass by soon, on the way to his wedding festivities. Michel has in due course been sent back to Moscow, and Count Potosky, having learned of his daughter's perilous journey, has risked the death penalty to break his exile in search of her. Through the aid of the provincial governor's son, he too has returned to Moscow, and with Michel's help he is seeking news of Élisabeth.

Élisabeth herself soon arrives at the inn and is joyously welcomed by Nizza. A mysterious officer appears - obviously a nobleman in disguise. Nizza has reason to believe that he is the Grand Marshal and - knowing nothing of the Marshal's enmity for Potosky - enlists his help to carry Élisabeth's petition to the Czar. So the joy of Potosky, Michel and Élisabeth when they are at last reunited quickly turns to terror when they learn that the secret is in the Marshal's hands and that troops are surrounding the hotel. Potosky is arrested, and expects execution for his defiance of exile. But when the Officer returns, he reveals himself as none other than the new Czar himself, having traveled incognito to observe the condition of his subjects. He praises Élisabeth's valor, banishes the traitorous Marshal and restores Potosky to his rank and estates, amid general celebration.



As explained in detail in the "Notes in Depth" below, all of the musical numbers in our edition of Elisabeth are pure Donizetti. With one exception, they were either composed for Elisabeth/Elisabetta or adapted by the composer from Otto mesi in due ore for inclusion in Elisabeth. That exception (the opening aria of Act Two) is an amalgamation of music that may have been intended for Elisabeth with music from Otto mesi that, as far as we know, Donizetti did not adapt. This speculative combination fills what would otherwise be a gap in the score. Meanwhile, one number (the final aria) was completed by the composer only in its Italian (Elisabetta) form, and has been translated into French for this production. As described in the notes, two sketches lend support to the idea that Donizetti intended this aria or something very similar to it for Élisabeth as well as Elisabetta. Thus it is possible to be reasonably confident that the shape and sequence of the reconstructed opera reflect Donizetti's own intentions.

The orchestration of these musical numbers is also pure Donizetti - the composer succeeded in completing orchestration for all the new music, and indicated clearly what orchestrations were to be carried over from Otto mesi (usually revising them as well). There is one small exception: in No. 14 (Michel's third-act aria), Donizetti composed a very beautiful passage (about a page long) and then decided to cut it before completing its orchestration. Because of the sheer quality of this short episode, we have decided to second-guess the composer's cut, and to fill in the missing bits of orchestration so that the passage can be included.

However, we have decided not to use the spoken French dialogue that would have connected the musical numbers if Donizetti's Opera-Comique plans had been realized. Instead, following a long-established practice for the international presentation of Operas-Comiques, we have inserted sung recitatives between the musical numbers. In Act Two, these recitatives are simply the ones that Donizetti composed for the Elisabetta version, translated into French (with some minimal alterations to accommodate our speculative version of the opening aria). In Acts One and Three, where different dramatic sequences make the Elisabetta recitatives unusable, we employ an abridged adaptation of the surviving Élisabeth spoken dialogue, set to music using bits and pieces of the Otto mesi score where possible, and with free imitation of Donizetti's Parisian recitative style in the instances where no Otto mesi model fits.




Sometime around 1804, one Praskovya Lupulova made an intrepid journey, mostly on foot, from Siberia to St. Petersburg to plead for the release of her father from exile. In 1806, Madame Sophie Cottin produced a fictional narrative based on this event that became a best-seller in many countries and was quickly adapted for numerous plays, ballets and operas. In 1827, Gaetano Donizetti and his librettist Domenico Gilardoni joined the list of adapters, producing a "melodramma romantico" at the Teatro Nuovo in Naples entitled Otto mesi in due ore (the title alludes to an eight-month journey presented on the stage in two hours). Shortly thereafter, Donizetti became convinced that the story had better operatic possibilities than he had realized in Naples, and over the next thirteen or so years he worked at various times to bring them forth.

Élisabeth, ou la fille de l'exilé - the opera presented at the Caramoor Festival for the first time anywhere - represents the fullest fruit of those labors, and the story of its development, curious disappearance, and rediscovery is as complicated as any opera plot. It is a detective story in which, by sheer good luck, I was able to play a principal role, and the following note is an attempt to condense the story into reasonably economical form. As will be seen, that is more than a trivial task: but as we hope the audience will agree, the task is worthwhile, because the opera itself is full of verve, passion, tender melodiousness and a spirit of adventure appropriate to the tale. The musical world is still in the process of rendering full account of Donizetti: the sheer mass of his output guarantees that it will require years for posterity to sort through it all and take notice of the items of most lasting worth. Élisabeth has a strong claim to be counted among those.

Apart from the idea of the daughter's brave journey, the story told by Gilardoni corresponds only very loosely to Cottin's. Various incidents in the book find echoes in the opera, but they are invariably much altered, and some of the principal turns of the plot have no equivalent whatever in Cottin; they were instead gathered and modified by Gilardoni from previous theatrical versions.

The basic plot of Otto mesi as presented in Naples in 1827 was as follows: Elisabetta, daughter of the exiled Count Stanislao Potoski and his wife Fedora, resolves to traverse the Siberian wilds to plead with the Czar for her father's pardon. She persuades the courier Michele (the son of her parents' servant Maria) to take her as far as the Siberian capital of Tobolsk, intending to continue thence on her own. In the wilderness, by the banks of the Kama, she encounters Ivano, one of her father's accusers, now banished in his own turn. After a perilous confrontation with a band of Tartar brigands, in the course of which Elisabetta manages to save Ivano's life, he offers her a letter exculpating her father. But before she can take it, the riverbank is flooded, providing a spectacular second-act finale and a miraculous escape for Elisabetta, who floats away on the wooden monument that had marked the tomb of Ivano's own daughter. Arriving finally in Moscow, Elisabetta discovers Michele, who has made his own way thence and bears with him the precious letter, which he had the luck to receive from the hands of the dying Ivano. But the Gran Maresciallo - another of Potoski's persecutors - intervenes and nearly manages to snatch the letter away. No matter; the Czar, appearing as a deus ex machina in the final scene, has already learned of the Maresciallo's treachery and decided on his own to pardon Potoski, adding to his speech a tribute to Elisabetta's bravery and presenting her to her parents, whom he has already recalled to the capital.

Apart from some names, almost none of this is in Cottin. Her Élisabeth travels with a Christian missionary returning from China, never meets her father's enemies, convinces the Czar simply through the impressiveness of her efforts, and has to travel back to Siberia to bring her father the good news. The novel concentrates on the physical, financial and moral stresses of the journey itself, giving the reader meanwhile extensive lessons in Russian/Siberian geography and ethnography. The confrontation with the Tartar thieves is present (briefly and differently), but there is no letter, no courier, and no flood. The father in the novel, moreover, was not Russian at all but a Polish patriot resistant to Russian domination. Although Potocki is a famous Polish name, no version of the opera makes allusion to this. Cottin does have a "love story" (very chaste but touchingly developed); quite unusually for such a romance-driven art-form, the opera leaves this out entirely. (1)

The plot of Otto mesi, meanwhile, is not quite that of Élisabeth, because of dissatisfactions that Donizetti seems to have felt quickly, even though he had to wait over a decade to act on them comprehensively.



Throughout the 1820's Donizetti was a busy, successful composer but not yet dominant on the crowded Italian scene. Several successes at the end of that decade consolidated his position, and he spent the whole of the 1830's in a whirlwind of incredibly productive activity, traveling constantly up and down the peninsula, producing new operas (28 in that decade alone) and reviving - often also revising - the best of his apprentice works.

Otto Mesi had been one of those successes - Donizetti's first big hit, in fact - but Gilardoni's dramaturgy was lame in several respects, and the composer was keen to have it improved as he set about reviving the opera in other cities. Part of Donizetti's problem with the piece stemmed from the genre in which he had originally composed it. Notwithstanding the unusual designation "melodramma romantico" - which may have reflected Donizetti's as-yet unrealized ambitions for the work - Otto mesi falls into the well-defined category of "opera semiseria," which was a highly popular form at the time and might roughly be defined today as a serious story with some funny elements and a happy ending. Innocence threatened and rescued - usually the heroine's - was the standard theme. The convention generally called for, among other things, at least one character from the stylistic world of opera buffa, a comic bass who chatted with the audience, made humorous remarks on the situations around him whether they were intrinsically funny or not, and sang texts of a more colloquial sort than the other characters. (In Naples and its vicinity, he used the local Neapolitan language while the others sang in Italian.) In Otto mesi this character is Michele, the courier. He sings patter songs, rambles on garrulously in his dialogue, mispronounces Russian words for comic effect, and works himself into a tizzy over the thought that Elisabetta might be enamored of him when she seeks to talk to him in private. Outside the sphere of Neapolitan influence, this role was translated into Italian, but it remained a chattering basso buffo and was sung by the interpreters of Doctor Bartolo and the like. Donizetti eventually realized that this choice had been unfortunate for the tone of the opera. In 1834, writing to librettist Jacopo Ferretti about proposed revisions, he said "...the buffo is a dog - so make him pathetic rather than comic" (adding however that he would not have time to write a new aria; that had to wait for the Paris project).

The composer was also unhappy with the last act. Its original centerpiece is a comic trio of "who's got the letter" between Elisabetta, the courier, and the Grand Marshal. And at the end, the Czar casually announces that he has already discovered the Marshal's guilt and has spirited Elisabetta's parents to Moscow before even learning of the daughter's mission, which is thus rendered superfluous. All that trouble for nothing! Donizetti wanted a different situation for this act: he asked to have the Count come on his own to Moscow in search of Elisabetta, and to be arrested there for breaking his exile. He planned to include this plot device in a prospective 1834 production in Torino that never took place, but kept it (and elaborated upon it) in Élisabeth.

He also needed recitative texts to replace the spoken dialogue used in Naples, but was dissatisfied with the first version of these, made for Palermo in 1828. Two other sets of recitatives appeared over the years, probably both supplied by Ferretti, in which an intent to simplify and elevate the tone of the conversations is detectible. And along the way Donizetti made miscellaneous musical changes, not all of which he retained from one production to the next. For Rome in 1832, the original, rather conventional Prelude was replaced by a Sinfonia that began with an imitation of Russian folksong and employed thematic material from the body of the opera. The Benediction in Act One - its music having been borrowed for incorporation into Anna Bolena - was rewritten entirely for Milan in 1831. An alternate entrance aria for the heroine and a radically different finale for Act One were composed and later abandoned. Cameo arias for the Grand Marshal and the Czar in the last act came and went. Two entirely different ensemble finales for the opera were composed, and there were also productions that substituted a concluding solo (not always the same one) for the victorious girl.

These alterations and still others are detailed in a recent book devoted to Donizetti's career by Annalisa Bini and Jeremy Commons, who conclude that Otto mesi was - already in Italy - the opera he revised more often than any other. None of these revisions was as radical as the thorough re-making the composer eventually undertook in Paris, but several of them left their mark on that effort.



The capital of France was the Mecca of 19th-century opera composers. There were various reasons for this, but the principal one was economic. Paris was wealthy enough to support several busy theaters at once; French law permitted authors to profit from their copyrights instead of depending solely on one-time fees; and the country's musical life was advantageously centralized. Every show started in the capital and, if it succeeded there, went to the provinces in orderly and remunerative fashion, in contrast to the lively and chaotic competition of Italian cities with one another. Rossini, Donizetti, Bellini and Verdi all concentrated on Paris in the later phases of their careers. So did the Germans Meyerbeer and Offenbach and anyone else fortunate enough to gain a foothold there (Wagner tried but famously failed).

The Parisian system, with its different traditions, also offered certain artistic advantages to an Italian: bigger, better-trained orchestras and more rehearsal time; more focus on the overall theatrical statement; relatively less power in the hands of individual star singers, and more flexible - or at least, differently flexible - dramatic conventions. (To give one example from which Élisabeth profits: it was possible for principal characters to enter and exit freely to facilitate the plot, in contrast to the unwritten expectation in Italy that they should sing something substantial each time they appeared on the stage.) So the idea of reworking their old operas for the Parisian stage was an attractive one in which all the composers just mentioned participated - whether simply touching up the score and adding pieces, as Donizetti did for L'Elisir d'amore and Rossini for La Donna del Lago, or rethinking it radically, as Verdi did for I Lombardi/Jérusalem or Rossini for Maometto Secondo/Le siège de Corinthe.

Donizetti entered the Parisian fray at the end of 1838, and within two years Berlioz wrote ruefully that "it is a veritable invasion. One can no longer speak of the opera houses of  Paris, but only of the opera houses of M. Donizetti." The houses conquered included the Opéra-Comique, the Théâtre des Italiens (traditional home of Italian-language performances in Paris), the Théâtre de la Renaissance, and finally the Opéra itself. The scores Donizetti introduced in this period are his most richly developed and musically subtle, including La Fille du Régiment, La favorite, Don Pasquale, Les Martyrs, Dom Sébastien, Caterina Cornaro, Linda di Chamounix, Maria di Rohan, alongside reprises of several older Italian works (usually with new numbers added). Such productivity is astonishing even by his own earlier standards - and the sterling quality of those operas forbids any suggestion that this was mere facility. Their excellence adds special interest to the discovery that yet another work from this fertile period awaits hearing.



One of Donizetti's earliest Paris projects was to re-make the Otto mesi story with a new libretto. This was provided by the established team of "De Leuven and Brunswick" - the noms de plume of Adolphe von Ribbing and Léon Lévy (or Lhérie). The project nearly reached fruition, but then was postponed and, through extraordinary circumstances, dropped into an obscurity so deep that catalogues, biographies and encyclopedic work-lists for over a century did not list it, not even as a lost work.

This much is clear: on 9 August 1839, Donizetti asked his friend Tommaso Persico in Naples to send "the book of 8 mesi" to him in Paris, and he set to work in that same year on his French version. On the following New Year's Day he wrote Persico that he would "soon" begin to rehearse the work at the Opéra-Comique with "Mal Garcia." The idiosyncratic honorific presumably means "Mlle." and refers to the celebrated Pauline García, fresh from her first London and Paris triumphs and shortly to marry Louis Viardot, whose name she adopted. A newspaper report shortly afterwards, meanwhile, suggests that rehearsals were underway with a different García: Eugénie, Pauline's sister-in-law. This is probably a misunderstanding: the music fits very well to what is known of Viardot-García's vocal range and distinctive style, and in any case Donizetti could hardly have referred to Eugénie, the wife of Pauline's brother Manuel, as "mademoiselle" García.

Why this production was cancelled we do not yet know for certain. When the younger Italian composer Uranio Fontana tried, after Donizetti's death, to wed the new French libretto to the Otto mesi music he had access to, several Parisian journalists offered conflicting accounts of the abortive effort (these are collected in the Bini and Commons book mentioned above; for an account of Fontana's production, see Note 2 below). Several of the composer's Parisian autograph scores show distinct "first" and "second" versions, and some accounts associated the project with other theatres, which leads Alexander Weatherson (chairman of the London-based Donizetti Society) to speculate that Donizetti attempted his project on two distinct occasions in Paris. The present state of our knowledge does not permit confirmation or rebuttal of this hypothesis.

The surviving sources for the opera, meanwhile, consist of the following:

1) French "short-score" adaptations in Donizetti's hand for sections of Otto mesi that he intended to retain.

2) Copyist's transcriptions of the vocal parts from these, pasted on (or formerly pasted on) to Italian copies of the orchestral scores of the relevant pieces.

3) Full orchestral scores of the new or substantially revised pieces, again in Donizetti's hand.

4) The De Leuven and Brunswick libretto prepared for Donizetti, which comes down to us only in the form used by Uranio Fontana after Donizetti's death and published in connection with Fontana's score.

The first of these four groups is to be found today in the Bibliothèque Nationale's collection of Donizetti manuscripts. The second and third groups are partly in that same archive and partly in London, since the French materials were carved up (sometimes literally) and reassembled in the process of preparing Elisabetta, a last-minute compilation of parts of Élisabeth and parts of the old Otto mesi for a proposed English engagement. The fourth item - the printed libretto - is an invaluable document, since it is the only surviving source for the linking dialogue, but also a problematic one, since it obviously underwent some further changes for the purpose of  Fontana's production. We cannot always say with certainty which parts of it correspond to what Donizetti had before him and which were altered later.



We do not yet have any direct evidence of why the Élisabeth score was dismantled, reassembled in Italian, and sent to London. At least two invitations for Donizetti to provide an opera for that city are known, as well as correspondence with a librettist about the composer's efforts to accept one of them. But thus far we do not know whether Elisabetta was assembled in response to either of these invitations or to yet another, nor to which theatre the score was originally sent, nor when that happened. The outside time limit would be late 1840 to early 1845. There is reason to believe that it occurred late in this period: A notation in the same assistant's hand that appears in the Elisabetta packet refers, on the autograph score of one of the pieces that remained behind in Paris, to the incorporation of that piece into Dom Sébastien, which had its first performance in November of 1843.

The London-related questions are thoroughly presented, and possible answers weighed, in Weatherson's recent article for the Donizetti Journal (Vol. 7). Further research should be able to resolve at least some of these mysteries, now that the clues are better understood. What we do know is that the eventual resting-place of the score was a storage basement in the Royal Opera House, Covent Garden. This room, apparently undisturbed through most of the 20th century, held a hodge-podge of materials obviously transferred from other theatres' archives, dating from as early as the beginning of the 19th century and from as late as the 1890's. Some of these, including the Elisabetta score, had obviously been through a fire. The theatre director whose invitations to Donizetti are known, Benjamin Lumley of Her Majesty's Theatre, did indeed suffer a fire in the 1860's - but a detailed listing by him of Donizetti's works, compiled after the composer's death, fails to make any mention of this one, thus throwing into question the supposition that it had been in his theatre's possession. Some copyist's materials discovered alongside Act One, and a note in English saying that the opera had been abandoned incomplete, demonstrate that it was being prepared for rehearsal by some theatre; it could even have been Covent Garden itself (which also had its brushes with fire). For now that is the most we can say.

In the ordinary course of events, Donizetti would surely have gone in person or sent representatives to London to see his piece through production. But in 1844-45 distressing signs emerged of the malady that would leave the composer an invalid, paralyzed and insane, until his death in 1848. The musical world followed his condition with horrified fascination and gradually dwindling hope; all consideration of his pending projects was put on hold.



The 20th century first learned of Élisabeth/Elisabetta in 1984, when I unwrapped some packets of music bound in dark, sooty cloth in that storage basement in the course of an unrelated research project. It was immediately clear that I had stumbled on a lot of Donizetti autographs, which was interesting enough in itself. But, not knowing all of the composer's nearly seventy operas, I could not immediately say whether the music was published or unpublished, known or unknown. Julian Budden, London's great scholar of Italian opera and the author of brilliant studies of Verdi and Puccini, kindly offered his help and his reference books, and within a few hours we were able to associate the work with Otto mesi. But examination of all available manuscripts of Otto mesi led to the realization that a great deal of the music, perhaps the majority of it, was indeed "new," unknown alike to the Donizetti scholars and to the reality of public performance. Moreover, most of it was composed in French, though later overlaid with an Italian translation, whereas all known versions of Otto mesi by Donizetti had been in Italian.

The next logical step, which I took in that same month, was to examine the Donizetti materials in Naples (where the autograph of Otto mesi, or rather what remains of it, is housed) and in Paris (where the composer had presumably been working at the time of the project in French). On first examination, this search yielded the interesting information that the Naples autograph had been torn up and that the missing bits of it were divided between Paris and London, and that the Paris archives did indeed contain a good deal of material relating to Élisabeth, including significant further amounts of "new" music that had not been sent to London. This music - resting in the Bibliothèque Nationale and thus accessible to scholars - was not exactly a new discovery; but without the London score, there had been no way to see it as anything more than disordered fragments.

What I found in 1984 at Covent Garden was only the first and third acts of the opera, but early in 1985, the conductor Richard Bonynge mentioned to me in Toronto that he was on the lookout for rare ballet scores for an upcoming recording. I told him I had noticed that a good deal of ballet music was also present in that Covent Garden basement, and suggested he might find something interesting there. When he next had the opportunity (not until 1988) he gave a thorough look through this ballet material, and in its midst he found - what else but the mislaid second act of Elisabetta! This was in exactly the same format as the first and last acts I had found, and had obviously become separated from them in the jumbled process of carting the material from wherever it had been singed by flames place to the room where it subsequently lay.

It is hard to overstate the confused appearance of these manuscripts. Bits and pieces in various hands and languages, from various original sources, are thrown together in what clearly was great haste, with scrawled instructions for copyists telling them to jump from this page or bar to that, to take the words or vocal parts from one score and the orchestration from another, to transpose this bit into such-and-such a key and that bit into another, to skip one passage and play another twice. That is true of the London version taken by itself, and all the more true of the Parisian materials, which lack even the hints as to how to put them together. What eventually became clear, though, is that the London score was assembled in Paris, by Donizetti and a French-speaking assistant or assistants, and that what remains behind in the Bibliothèque Nationale is essentially the bits "from the cutting-room floor."

Now that at least one fully complete version (Elisabetta) was at hand, Covent Garden was naturally interested in mounting a production, although managerial changeover at the house and its closure for renovation complicated and delayed the plans several times. Eventually, Elisabetta was presented in 1997, edited by me and Roger Parker (co-Editor in Chief of the Critical Edition of Donizetti's operas for the Italian publisher Ricordi). The performances had to be in concert form, at the Royal Festival Hall, as this was during Covent Garden's renovation period. They served, among other things, to introduce the unknown young tenor Juan Diego Florez to the London public as a last-minute substitute in the role of Count Potosky.

At that time, though some of us had an inkling that Élisabeth was a preferable version, we all assumed that its reconstruction would be a much more hypothetical affair. Only later, when I began the painstaking process of transcribing the fragmentary Parisian manuscripts note by note and page by page, and after making a second trip to Paris that turned up further material (misidentified by the Bibliothèque's 19th century archivists), did I realize that in fact the jigsaw puzzle could fit together neatly. This determination depended on ascertaining exactly what page in Paris should link up to what page in London, what copyist's score should be mated with what short-score in Donizetti's hand, where the cut-up strips of vocal lines made by the copyists were once glued to which orchestral scores, which apparent fragments were meant to be inserted into another manuscript and which to be discarded, how long the gap produced by tearing off half-a-sheet might logically be, and many other calculations of the sort. The invention of photocopying has greatly simplified the business of putting together an evening-length work for voices and orchestra! In Donizetti's time, when the only way to copy was to start from scratch with pen and ink, it was necessary under deadline pressure to create instead a kind of treasure-map guiding the copyists through the maze of manuscripts, fragments, inserts and corrections. Putting together the French score that he had purposely carved up meant searching for the treasure without the map.

In the end, it became apparent that Donizetti indeed had a French version almost fully ready, stopping just short of the heroine's final aria - and that aria, which survives only in its Italian Elisabetta version, was sketched by him on the back of the last page he reached in the French autograph score, which lends some confidence to the supposition that he intended it, or something like it, for Paris as well. A very few gaps remain (they are discussed below), but by the time I finished studying the Parisian manuscripts, it was clear that there was a real opera, and a beautiful one, to be reckoned with.



What, in 19th-century opera's world of recycling, revising, mixing and matching, is a "new" score and what is a "revision"? The line is far from clear. If "new" can only mean conceived without reference to a previous work and composed from scratch, then Élisabeth does not count as an unknown opera but rather as an unknown revision of a forgotten one. However, proportion tells its tale. Donizetti was obviously happiest with the short middle act of Otto mesi, depicting Élisabeth's adventures in the wilderness; most of this was retained through all versions, and made its way, albeit much revised, into Élisabeth. The longer outer acts are a quite different matter: In Élisabeth, only about a third of Act One, and not a note in Act Three, derives from the Neapolitan score.

This is not to say that the other two thirds of Act One and the whole of Act Three were composed entirely new for Paris, because Donizetti drew also on three pieces written for intervening revivals of Otto Mesi. The process of reshaping his treatment of this story was gradual, but it flowed in one consistent direction: elimination of the traces of opera buffa and development of the more Romantic, lyrical, and passionate aspects of the adventure story.

These lie at the heart of Élisabeth's musical and expressive beauties. Donizetti does not seem to have read Cottin himself; his requests for improvements in the drama make reference only to a ballet version by Salvatore Gioja that he had seen in Naples and/or Milan in the 1820's. Nevertheless, he appears to have reached instinctively for a quality that is quite marked in the novel, where the focus is firmly on the bravery of the central figure, the difficulties she surmounts, and her devoted love for her exiled father. As presented and surrounded in Otto mesi, she risks seeming like a re-make of Rossini's Italian Girl in Algiers, merrily wrapping the world and its bumbling inhabitants around her little finger. Donizetti shrewdly judged that the story would shine in a better light with a tilt towards Romantic seriousness.

That tilt also corresponded with his own artistic development. Donizetti's last years were spent in rapid exploration and expansion of his musical horizons, as though some instinct told him that time was short. He sought ever more creative alternatives to the formal structures inherited from Rossini, and his taste for piquant harmonies became at once bolder and more refined. His orchestral palette was enriched by exposure to Parisian resources: the woodwinds are used as a flexibly recombining source of color (as opposed to the Italian tradition of block harmonies alternating with melodic solos), and the expressiveness of the string group is expanded by an increasing willingness to confide important material to the violas and to treat the cellos and basses as two separate groups. The score of Élisabeth is a treasury of beauties newly discovered, renewing and complementing the composer's undiminished melodic inspiration.

It is not as though this shift to seriousness turns the opera melancholy; after all, the story is still one of daring, high spirits and triumph. But melancholy is now felt in the background. Where Otto mesi opened with a chorus celebrating Elisabetta's birthday, Élisabeth begins with a piercing expression of suffering and forlorn hope from the Siberian exiles. Where Otto mesi's Conte Potoski is all vigor and bluster, Élisabeth's Comte Potosky (a longer and far more fully developed role, with almost completely different music) sings in tender consolation of his wife's fears, heartrending lament for the fate of an exile, anguished agitation over his illegal return to Moscow, and elation at the rediscovery of his beloved child. His presence puts some real tension into the third act. In the new plot, in which Michel's sweetheart unknowingly puts the crucial letter into the hands of an officer who seems to be the Grand Marshal, Potosky's other great enemy, Élisabeth risks not only failing in her mission, but exposing her father to summary execution.

Equally significant is the transformation of Michele into Michel. He is still a high-spirited character, but in an entirely different way: part mama's boy, part lover-boy, part Boy Scout. He has become a lyric tenor instead of a basso buffo, and most of his music is new: a bewitching song in Act One as he tells his mother of his efforts to visit her in Siberia, with harmonies and orchestral colors that look forward to Bizet; a warm-hearted larghetto as he attempts to dissuade Élisabeth from her risky journey; and a bravura showpiece in Act Three as he contemplates marriage and the risks he would face as a traveling husband who must leave his wife prey to rival admirers.

The effect of this change can be glimpsed by comparing (in English translation) two fragments from the courier's duet with Élisabeth; Donizetti, with impressive resourcefulness and with only minimal changes in the accompaniment, manages to reflect these differences of tone in his new vocal line.


To Tobolsk? You, really? Is it true, mademoiselle?

But you're joking, it's crazy, crazy, crazy! You wish to leave? 



Quiet, I'm telling the truth. Quiet, and let your zeal

I have decided; no one knows about it. Fulfill my resolve this very day!



A caprice has got hold of you; What mission are you contemplating?

I swear, you're crazy. I cannot consent to it.



Then I shall make my way Please, yield to my entreaty;

All alone to the Capital, I want to leave with you.

And shall ask the good Sovereign Think, that it is a father

For my father's freedom. Whom I must save from exile!



You're out of your mind. What mission are you contemplating?

For pity's sake, girl! I cannot consent to it!

And later, after he has agreed to help her:



(In all the fair sex (God, be her guide,

I have never found another Be her shield!

With more of the devil in her Her faltering heart

Than this one here.) Puts its hopes in you!)


The character of Élisabeth herself does not change so much in basic musical treatment, being already the strongest aspect of the parent opera. But her vocal color is shifted radically, from typical soprano to rich mezzo-soprano range, while her role is enriched by subtle new musical details and by her part in the Act Three scenes made possible by the plot change Donizetti requested from his librettists. This is a particularly strong new sequence: the dramatic aria of the Count (written for the proposed Turin production and amplified for Paris); the transporting duet of reunion when his daughter appears (new), and then the biggest piece in the whole opera, the multi-movement trio (also new). In the latter, Élisabeth, Potosky and Michel confront the news of the Marshal's intervention, plan the Count's escape using the courier's safe-conduct, and - when the arrival of Imperial troops foils that plan - defy the assembled soldiers. All this gives weight to the brilliant Rondo (new) with which Élisabeth closes the opera when things are set right.

Donizetti's Parisian treatment of the antagonist figure, Ivan, is a slightly more complicated question, due to a gap in the score discussed below. But in the meantime it can be observed that his long duet with the heroine, taken largely from the Otto mesi original, fits beautifully into the more mature surroundings of Élisabeth. It is the earliest full manifestation of the expressive qualities that Donizetti sought in his long process of involvement with this story.

Élisabeth is the culmination of that process, even though it is not the very last version the composer prepared for performance. The projected London Elisabetta represents a step backward, because several factors (the voice-types of the anticipated cast, the return to Italian text, and the use of recitative instead of dialogue) made it more convenient for Donizetti to turn back to Otto mesi at several points, and omit some of the new music he had composed. He did incorporate six new numbers (or substantial portions thereof) from the Paris score, but left out at least three others and substantial sections of several more, and he restored several Otto mesi pieces that had not figured in Élisabeth. The only new items for London were a cabaletta for the soprano-baritone duet - originally composed for insertion into Rossini's L'Assedio di Corinto - and the rondo finale for the heroine. Our production includes the rondo (which, as already noted, was sketched on the last page Donizetti reached in the French autograph), but not the Assedio cabaletta, since his Paris version of the corresponding piece survives and seems more appropriate to the dramatic situation.

The most decisive factor against Elisabetta is probably the decision to restore the talkative basso buffo version of the courier. He regains his entrance aria from the 1827 Naples score, and most of his duet with Elisabetta (including the passage quoted above) goes back to its 1827 form as well. Whatever bits of his Parisian music are retained, meanwhile, are appreciably compromised by the necessity to rearrange them from high tenor to middling baritone range (several bits of intricate singing à deux with Élisabeth, for instance, have to be abandoned and left as solos for the heroine). But even outside Michel's role some good new Élisabeth passages are lost, and the recitatives hastily scribbled for London's non-Italian-speaking public do little to give coherence to the story.

When Roger Parker and I prepared an edition of the rediscovered work for Covent Garden, the decision was taken to use the Italian version, since after all that was what Donizetti had intended to present in London. Fairness suggested that Elisabetta should have a hearing after waiting in the basement for more than a century. But both of us realized at the time that we were working on a compromise score, and many listeners, I think, perceived its patchwork nature even without knowing the details. Even so, the opera's beauties and spirit made an impression. The exciting thing about Élisabeth is that they have a chance to do so much more fully and coherently. Donizetti's protracted involvement with this lively tale was obviously a labor of love, and the opera he produced - finally restored to the form he intended - inspires love in those of us who have labored on it in our turn.



As noted above, the musical score for Élisabeth is more complete than anyone had initially hoped, and reconstruction of it involves only the most minimal speculation. Part of the heroine's entrance aria was lost when the Paris manuscript was torn up so that some of it could be incorporated into the London version, but it is easy enough to restore this following the 1854 libretto and the musical hints in the portion of the score that survives. Eight bars of the father-daughter duet in Act Three suffered the same fate, and in that case it is the London version that we hear at the Caramoor performance (luckily, the French words fit). These two gaps are relatively trivial; the only significant question-mark involves the opening of Act Two.

In all versions of Otto mesi, this act begins with a prelude, recitative, aria and cabaletta for the deposed colonel Ivan, and Donizetti put this scene unchanged into the London Elisabetta. But there is no trace of any French adaptation by the composer, which we have for every other Otto mesi piece that made its way into Élisabeth. Moreover, the published 1854 libretto contains an aria text for Ivan that does not fit the Italian piece in either rhythm or content. This at least suggests that Donizetti and his French librettists had another plan for the beginning of the act that either was abandoned (perhaps the composer decided not to give Ivan a solo), or else is lost to us thus far.

However, there are at least two tantalizing fragments that may relate to this gap. A contemporary French score of Roberto Devereux includes, as an appendix, an aria described as having been written specially by Donizetti for Paul Barroilhet (his favored Parisian baritone) for use in an 1841 Lyons production. This aria - of breathtaking melodic beauty and harmonic daring - bears traces of being a revision of some lost original: its text does not come out in any coherent poetic form, which is usually a sign of words written to fit pre-existing music composed for some other text. The date too is suggestive: this is exactly the moment at which Donizetti might have been expected to "farm out" pieces from his abandoned project, which indeed we know that he did in other cases (some material from Michel's Rondeau went into Rita in June 1841, and a movement from the 3rd-act trio went into Dom Sébastien in 1843). And the range and mood of the aria are strikingly consistent with the part of Ivan in Élisabeth. I am much indebted to Alexander Weatherson for bringing this aria to my attention.

In addition, the vast Parisian archive of Donizetti manuscripts contains a mysterious 40-bar orchestral prelude, in the hand of a copyist with annotations by the composer, and with the heading "new prelude to be inserted in an unpublished opera." At the time these notations were most likely made (apparently by an assistant helping Donizetti to organize his manuscripts), only one or two works besides Élisabeth could have been described as "an unpublished opera," and they are less likely candidates as hosts for this prelude. It may possibly have been intended to replace the Otto mesi overture: its key (E minor) and unresolved ending on the dominant would lead properly into Élisabeth's chorus of exiles, though it would tend to make the existing instrumental ritornello of that chorus redundant. It could also conceivably have worked as a prelude to Act Two, whether with a (lost) Ivan solo or without, since in the latter case the next musical piece would be the E major duet with Élisabeth.

Nothing at all concrete is yet known about this, but we have decided to use this prelude to lead into Ivan's solo scene at the beginning of Act Two, rather than employ the less interesting Otto mesi prelude that went into Elisabetta. Therefore our Act Two begins with:

    - the unidentified Prelude preserved in Paris

    - the Otto mesi recitative used in Elisabetta

    - the Robert Devereux aria

    - the Otto mesi cabaletta used in Elisabetta

This scene must therefore be clearly acknowledged as speculative. The reassuring thing is that it is the only substantial number in the opera for which this is so. If further research should bring to light the Élisabeth libretto in the exact form that Donizetti had before him, it might be possible to confirm or reject this speculation, but until then, it seems a satisfactory way to fill the only significant gap in the opera.


It remains finally to consider the linking material between the musical numbers. In Élisabeth, following the practice of the Opéra-Comique, simple spoken dialogue was to have filled this function. Luckily, the 1854 published libretto includes dialogue, so with minimal adjustments for the differences between Fontana's version and Donizetti's, this element can be reconstructed. For a performance in France, that could be the end of the matter. But the export of French dialogue-operas raises a problem all too familiar to operagoers around the world. Learning to sing good French is difficult enough, but speaking it convincingly is another matter altogether for polyglot casts. Moreover, audiences have never taken to spoken dialogue in foreign languages as easily as they have taken to the international convention of sung recitative. So it was long traditional for French opéras-comiques to be provided with recitatives when they moved either to the international scene or to the Paris Opéra (where recitative in French was the custom). Carmen, Faust, Mignon and Les contes d'Hoffmann are among the familiar operas for which this was done, either by their composers or by other hands. Donizetti himself supplied recitatives for La fille du régiment when he adapted it as La figlia del reggimento.

He did the same in arranging Élisabeth as Elisabetta, so the obvious solution for an international performance in French is to make a shotgun marriage between the words of the French dialogue and the music of the Italian recitatives. And in several scenes, that is exactly what we will hear tonight. But a problem arises: as noted above, the French librettists made several changes in the plot. In arranging Elisabetta for London, however, Donizetti turned back to many aspects of Otto mesi in due ore, thus creating recitatives incompatible with the French version. So in the end, the "shotgun marriage" had to become a ménage à trios. For Act Two, we use Donizetti's Elisabetta recitatives (translated into French, with minimal alterations). For Acts One and Three, where the plot changes occur, we use the Elisabetta recitatives when possible, but otherwise a loose abridgement of the Élisabeth dialogue, set to music in my own best effort to imitate Donizetti's Parisian recitative style.

So I hope the public will blame me and not Donizetti for any stray passages of recitative that may be found wanting, and meanwhile that it will join all of us in the excitement of restoring to the operatic stage a rich creation by one of its most inspired entertainers.



1. Élisabeth's destination is the source of some confusion. Cottin's heroine intends to go to St. Petersburg but finds to her surprise that her mission can be accomplished in Moscow, since the new Czar Alexander is in town for his coronation. However, the author specifies in her preface that the historical model, Lupolova, actually went the further distance to St. Petersburg itself. Some versions of the Otto mesi text follow this, with references to the westerly capital but with the dénouement nevertheless set in Moscow, for sketchily explained reasons. The Élisabeth libretto, in the only form of it that survives today, abandons all thought of St. Petersburg and makes Moscow the intended goal from the outset. This question matters more to the reader conversant with the habits of the Russian court than it does to the substance of the opera.

2. Uranio Fontana's effort to wed the De Leuven and Brunswick libretto to the Otto mesi score, premiered in 1853 at the Théâtre Lyrique, has also generated confusion, understandably enough in light of the complex situation. Several commentators, regrettably including the compiler of the New Grove Dictionary's latest work-list, have assumed that it was a completion of Donizetti's unfinished Parisian efforts. It was in fact exactly the contrary: an attempt to use the new libretto without benefit of those efforts, to which Fontana lacked access. He worked from an old Italian score; everything by Donizetti in his version comes from the parent opera, and for the texts that Donizetti had set in Paris - well over half the opera - Fontana was obliged to invent his own music from scratch. Fontana may have encouraged the impression that he had special information; he described himself as a "pupil" of Donizetti, though independent confirmation of this relationship is elusive. However, the only hints of Donizetti's Élisabeth in Fontana's consist of three brief and vague melodic resemblances. These suggest that he may have seen the score or heard rehearsals of it while Donizetti was at work (which he could have done, having been in Paris in 1840 for an opera of his own at the Théâtre de la Renaissance). But their pure-Fontana continuations, in contrast with Fontana's generally faithful transcription of the Otto mesi music, confirm that he did not have a score of Élisabeth in hand. (To date, no copy of the Élisabeth music other than Donizetti's autographs has been found, with the exception of one copyist's score of a passage from Act One. This was presumably prepared for the rehearsals that we know the composer began, and the fact of those rehearsals makes it certain that other such copyist's materials must once have existed. If these should eventually come to light, it is possible that the few gaps in the score could be authoritatively filled.)

Meanwhile, the libretto published in 1854 in association with the Fontana production corresponds for the most part quite faithfully to what is in Donizetti's autographs. We can confirm from those autographs that most of its plot changes were already in place when Donizetti set to work, and also that the librettists were working hand-in-hand with him as he composed: his mid-course revisions in Michel's first aria required some new lines of text, and those new lines are present in the 1854 publication. On the other hand, some further changes were clearly needed by Fontana, who was working in the dark, as Donizetti's new music had already gone off to its hiding-place in London. There are a few passages where this tinkering is obvious and can be corrected by reference to Donizetti's autographs, but there remain some uncertainties, especially in the dialogue sections.


APPENDIX 1               Otto mesi and Elisabeth

For those interested in the details, here is a piece-by-piece listing of Élisabeth as presented at the 2003 Caramoor Festival, with notes as to the origin of each section. The pieces described as being "from Otto mesi" (without further comment) come from the 1827 Naples score, and have some slight revisions attendant on their translation into French; only substantial alterations are noted separately.

Sinfonia: Added to Otto mesi in Rome (1832); revised for Paris


No.  1 Introduction:

          a) First chorus: new for Paris (based in part on a chorus from Maria Stuarda)

          b) Second chorus: from Otto mesi, slightly revised for Paris

          c) Cavatine Potoski: new for Paris

          d) Cavatine Élisabeth: based on original Otto mesi number but with heavy revisions

                            and new material for Paris

No.  2 Romance Potoski: new for Paris

No.  3 Cavatine Michel: first section based on Otto mesi, revised for Paris;

                            second and third sections new for Paris


No.  4 Duo Élisabeth-Michel: first and last movements from Otto mesi, revised for Paris;

                            middle section new for Paris

No.  5 Benediction: added to Otto mesi in Milan (1831); revised for Paris

No.  6 Final: from Otto mesi; revised for Paris


No.  7 Prelude et scene

          a) Prelude: from unpublished Paris ms., tentatively ascribed to Élisabeth

          b) Recitative: from Otto mesi

          c) Cavatine: from added aria in Robert[o] Devereux (18TK, Lyons), based on unknown original

          d) Recitative and cabaletta: from Otto mesi

No.  8 Duo Élisabeth-Ivan: based on original Otto mesi number but with revisions and

                            new material for Paris

No.  9 Choeur Tartare: from Otto mesi

No. 10 Trio: from Otto mesi, revised for Paris

No. 11 Inondation: from Otto mesi


No. 12 Choeur des Soldats: new for Paris

No. 13 Rondeau Michel: new for Paris

No. 14 Air Potoski: added to Otto mesi for unperformed Turin version, 1834;

                             revised and expanded for Paris

No. 15 Duo Élisabeth-Potoski: new for Paris

No. 16 Trio Élisabeth-Potoski-Michel: new for Paris

No. 17 Rondeau Final: new for London Elisabetta; sketch on final page of  Paris autograph of No. 16



APPENDIX 2: Élisabeth and Elisabetta

Changes, other than translation, made by Donizetti in adapting Élisabeth (intended for Paris and premiered at Caramoor) into Elisabetta (intended for London and premiered there in 1997)

Sinfonia: No change

No. 1: Élisabeth's cavatina reverts mostly to Otto mesi equivalent plus 23 bars inserted from Élisabeth

No. 2: Omitted in Elisabetta

No. 3: Reverts to Otto mesi equivalent

No. 4: Except for middle section, reverts to Otto mesi equivalent

Nos. 5-6: No change except reversion to the Otto mesi vocal parts

No. 7: Reverts to Otto mesi equivalent

No. 8: Slow movement reverts to Otto mesi equivalent; cabaletta replaced  with insert number

                                for Rossini's L'assedio di Corinto

Nos. 9-10-11: no change

No. 12: Omitted in Elisabetta (unrelated  chorus from Otto mesi substituted)

No. 13: Transposed and adapted

No. 14: Small cut for Elisabetta; no other substantial difference

No. 15: No substantial change

No. 16: Fourth movement ("De la prudence") omitted in Elisabetta

No. 17: Completed for Elisabetta; French translation made for the present   production

Note: Donizetti made a few cuts in the autograph scores, and it is not always clear whether these were done only for Elisabetta or were already intended as cuts for Élisabeth. The Caramoor production includes the passages in question. 




Page initially published in  2003