Musicians and critics who posit the existence of a continuous musical tradition linking the present to nineteenth-century practices are faced with increasingly persistent evidence that this apparent continuity is a chimera.1 Instruments and the techniques for playing them, vocal styles and the use of ornamentation, the social role of music and its implications for musical structure and performance—all have undergone vast changes since the time of Beethoven, Chopin, or Rossini, not to mention Grieg, Elgar, and Bartok. Some conductors arrogantly dismiss this evidence as ignorant posturing by nonmusicians or the refuge of poor souls unable to play in tune. But even those most contemptuous of efforts to find for the so-called standard repertory a performance practice more responsive to the historical circumstances under which it was created and first performed cannot dismiss a movement whose popular reception has been enthusiastic.2
The principles embraced by musicians and critics committed to the development of more sensitively historicized performance styles, on the other hand, are equally under siege. Battles rage in the scholarly literature and popular press about numerous details: the best way to apply ornamentation in various repertories, the interpretation of tempo markings, and the choice of instruments to be employed (whether in the orchestra, for solo keyboard works, or in the accompaniment of secco recitative). But the very possibility of developing such alternative styles with a sensitivity to historical circumstances has been attacked by those who insist that conductors from Toscanini to Roger Norrington are merely demonstrating the triumph of “modernism” in performance, an anti-Romantic sensibility that has nothing whatsoever to do with the historical past and everything to do with a present-day revolt against more proximate performing styles.
We are surrounded by proponents and critics of either a loosely defined tradition, or a questionable search for authenticity, or a pessimistic, even nihilistic insistence that all our efforts lead to our merely embodying the modernist or postmodernist (or whatever tomorrow’s “-ist” may be) tendencies of our current society. We also belong to a generation of musicians and music lovers formed by a century of mechanical reproduction of music. Our memories merge evenings in the theater with innumerable recordings. We lose track of what real performances sound like, so confused and enchanted are we by the doctored sounds served up by record companies. Small wonder that a minor industry has arisen in comparative analysis of early recordings: its device, a stopwatch to compare performances of Beethoven symphonies frozen on disc with the composer’s own metronome markings, whatever those highly contested numbers might mean.
Let me state at the outset that I do not consider modern performers to be under any obligation (moral or otherwise) to respect either the particular qualities of a work or the general characteristics of its composer’s style. To take an example that some musicians found disturbing, I had nothing but profound admiration for Peter Brook and his La Tragédie de Carmen, theatrical art at its most memorable. This adaptation and transformation of a novella by Prosper Mérimée, using melodies by Georges Bizet reorchestrated intelligently for a small instrumental ensemble, had no pretense to be Bizet’s Carmen. It aspired to be a work in its own right, with its own aesthetic integrity.
Once they set aside the Peter Brook model and claim to be presenting a nineteenth-century opera, however, performers find themselves in a different situation. Even here, I have no interest in invoking the language of moral obligation. Yet conductors, singers, and stage directors (even the most iconoclastic) inevitably approach their tasks enriched or encumbered with their knowledge of previous performances of a work or similar works, whether through personal experience in the theater, secondary accounts (the opinion of teachers and coaches, for example), or recordings. Everything has been filtered further through sensibilities reflecting those experiences that make us all international citizens of the early twenty-first century. Performances, in short, are already “historical through and through.”
The call to historicize further our knowledge of Italian opera, then, is not a subterfuge for escaping our modern identity or the personal passions of an artist in favor of historical models. It stems rather from the belief that performers can find more satisfactory answers to some of their concerns if they make the effort to supplement their present knowledge with awareness of a work’s original historical, dramaturgical, musical, and social context: indeed, much in the shape of a typical nineteenth-century Italian opera is meaningful only in such a context. Neither invocations of the authority of tradition (which too often is a euphemism for the status quo) nor obeisance to the idol of historical reconstruction will offer as promising a path.
This chapter and the ensuing ones will provide access to this context for a variety of issues that performers confront each time they perform an Italian opera written during the first half of the nineteenth century. By combining historical evidence, theoretical models, and actual examples, these discussions may help all those concerned with Italian opera—whether as performers, critics, or listeners—to understand the ways in which the lessons of history and the realities of today’s musical culture can interact effectively in the opera house.
“IT IS ALL BELLINI’S WORK”
Before rehearsals for an opera begin, performers must decide what music to incorporate in the particular production being planned. A first set of decisions involves the choice among alternative versions, when these exist, as they almost always do: that is the subject of the present chapter. Once a basic version has been chosen, performers need to address the problem of whether to make any cuts: that is the subject of the next chapter. Too often these questions implicitly or explicitly are couched in terms, whether aesthetic or practical, that limit the history they invoke to the immediate past.
In an opera such as Lucia di Lammermoor, traditionalists insist on adhering to a rigidly defined, heavily cut version, whose historical roots they ignore. Whether consciously or not, they are essentially following the version of Lucia recorded under the baton of Tullio Serafin in the 1950s. This version, furthermore, was given a semblance of permanence in one of the most wrongheaded books ever written about Italian opera, entitled Style, Traditions, and Conventions of Italian Melodrama of the Eighteenth and Nineteenth Century, published (in Italian) by Ricordi in 1958, to which we will return in the next chapter.3 The failure to think through the performance history of individual works, though, extends well beyond operas like Lucia, which did, after all, have a fairly continuous performance history from the time of its composition to the present.
In 1977, when directing at New York City Opera the first performances of Rossini’s Il Turco in Italia in the new critical edition, Tito Capobianco, a well-known stage director, told me that certain decisions about versions and cuts in the opera were traditional, as if those decisions went back to the nineteenth century. But there was no continuous performing tradition for Il Turco in Italia, which had disappeared from the repertory for a hundred years before being revived under the direction of Gianandrea Gavazzeni at the Teatro Eliseo of Rome in 1950 (with the participation of Maria Callas). The opera was recorded at La Scala during the summer of 1954, then presented at the theater the following March, always with Callas. I know these performances, of course, only from the recording, but on that basis one observes the intelligence of most of Gavazzeni’s musical choices and the stylistic mastery of Callas’s interpretation.4 Writing in the program book, John Steane comments that “she excels in recitative.” And referring to the conclusion of the opera, after the great quartet at the masked ball (“Oh! guardate che accidente”), he adds, “Then in the final solo of the opera, Sì, mi è forza partir (Yes, I must leave), the sentiment is real; the tragic accent now marks a genuine development and the character gains a new warmth and depth in a way which perhaps only an artist with Callas’s skills and experience could bring out.” What Steane failed to understand is that this part of the opera has been decimated. In the version they chose to perform, surely unbeknownst to Callas, Gavazzeni omitted her principal aria
and one of the most stunning single compositions in the opera, “Squallida vesta” (Squalid clothes). This is the aria in which Fiorilla, having learned that her husband has barred the doors of his house to her, repents her flighty behavior. The trivial secco recitative that follows (“Sì, mi è forza partir”)—written by one of Rossini’s collaborators, not by the composer himself—is utterly void of any dramatic meaning without that previous scene. Who told her that she would have to leave Naples for her parents’ home in Sorrento? Of course, one wonders what Callas would have had to say about performing Il Turco in Italia in Gavazzeni’s version had she been aware of this cut!
On the other hand, in their monolithic insistence that every note of a score such as La traviata be played in the theater, reformers show equally little awareness of the textual and performance history of the works they champion and the theatrical conditions that prevailed, even in the best of circumstances, during the first half of the nineteenth century. As Charles Rosen has argued, it is highly suspect to historicize the performance of a musical artifact, concerning ourselves with its notes, rhythms, orchestral forces, and vocal style, while abstracting the work from the historical and social circumstances that gave it birth—to presume, in short, that one can create an authentic performance in an inauthentic environment. And yet, who would feel it obligatory, while improving the musical text of a nineteenth-century opera, also to reproduce the appalling circumstances under which many acknowledged masterpieces of our musical heritage were born?5
These rigidly drawn lines succeed only in obscuring the issue: how can we use our knowledge of the history of performance, our grasp of the social realities of nineteenth-century Italian opera, our analytic skills, and our aesthetic perceptions to develop a more responsible approach to choosing what music to include in modern performances? In order to address this question, we need to recall the social realities of Italian opera and their implications for the nature of operatic “texts.” Can we talk about an opera that exists in multiple versions as being, in any meaningful sense, a “work”? Or must we consider these operas to be loose agglomerations of music that can be manipulated in any way we find convenient?
However much we may avoid the language of moral obligation when discussing performance, it remains commonplace to speak of the obligation of a critical edition to print a work in a form as close as possible to an author’s intentions, even though the philosophical and practical difficulties in understanding “authorial intention” are legion. These difficulties are well known to literary editors: the multiple texts of King Lear, Dreiser’s Sister Carrie, and the poems of W. H. Auden are notorious examples. Different versions of a poem or novel can be contemplated and compared by a solitary reader, even if critics disagree about which text to favor through the physical organization of a book. Should they put alternative versions face to face, in footnotes, or in a principal text and a series of appendixes? Although plays in their theatrical realizations present problems similar to those of opera, there remains a need for reading texts prepared with criteria similar to those employed for other literary works.6
For composers of Italian opera, on the other hand, publication was important primarily as a way to facilitate performance. Indeed, through about 1820 even the most popular operas circulated only in manuscript or in incomplete reductions for piano and voice (lacking, for example, the recitative). Ricordi began publishing complete vocal scores in the mid 1820s, as we have seen, but it was not until near the end of Verdi’s life that some full orchestral scores of his operas appeared (those for Otello and Falstaff), and even then the composer was only tangentially involved with the process.7 It is not surprising, then, that authorial intentions in Italian opera are often bound up with the requirements of particular productions.
Although it was more lucrative for composers to prepare new works than to mount new productions of old ones, various circumstances could lead to their involvement in revivals: they might be preparing a new opera for a theater that was simultaneously mounting a production of an earlier one (as Rossini did when he presented a new version of Maometto II to open the Venetian season of carnival 1822–23, for which he wrote Semiramide);8 they might agree to participate in a revival for a special circumstance, such as the opening of a new theater (as Bellini did when he revised Bianca e Fernando for the opening of the Teatro Carlo Felice in Genoa);9 they might seek to accommodate the wishes of an admired singer (as Donizetti did when he transformed a secondary tenor role in Maria di Rohan, Armando di Gondì, to a principal role en travesti for the contralto Marietta Brambilla at the Théâtre Italien of Paris a year after its Viennese premiere);10 they might genuinely feel an opera needed revision on aesthetic grounds (as Verdi did when he prepared a major new version of La forza del destino for Milan in 1869, seven years after the St. Petersburg premiere).11 We can usually document, reconstruct, and publish these versions, however they may originally have been motivated (whether to meet changing needs of the performing forces or the diverse tastes of different publics), but we need to be cautious about assigning them relative worth. This is true even in the case of a composer such as Verdi, whose treatment of his own autograph manuscripts suggests he intended to leave his operas in what he considered to be definitive versions.12 It is not intuitively obvious, however, that the Verdi responsible for the 1865 Macbeth, revised for Paris, could look with sympathy or even with equanimity at the original 1847 version of that opera.13 The availability of multiple versions prepared by the composer, then, creates a first problem that must be addressed by modern performers of Italian opera.
Some commentators, such as Carl Dahlhaus, blinded by the influence of social and theatrical structures on this process of revision, withhold altogether from Italian bel canto operas the status of a “work,” preferring to see them as amorphous collections of individual compositions given temporary shape by particular productions.14 By emphasizing abuses of the Italian theatrical system, this position attempts to ignore the evidence of composers’ letters (particularly those of Bellini, Donizetti, and Verdi), who railed against theaters and performers tampering arbitrarily with their works. One could argue, to be sure, that some of that railing was itself theatrical, a product less of deep artistic conviction than of an effort to ensure control over theaters (and of the royalties that accrued from such control). Thus, during the late 1840s Verdi regularly asked Ricordi to insert in every rental contract with theaters a clause of this type:
With the aim of impeding the alterations often made in opera houses, it is prohibited to make any change in this score, any mutilation, to lower or raise the key, in short to make any alteration that requires even the smallest change in the orchestration, under penalty of a fine of 1000 francs, which I will demand from you whenever any theater whatsoever makes such an alteration in the score.15
But there is considerable contemporary testimony, even before Verdi’s diatribes to Ricordi, that composers, singers, and theaters were sensitive to the status of Italian operas as “works,” whose structure should be respected, where possible, for both practical and aesthetic reasons. Let me offer examples from 1828, 1834, and 1836, pertaining to Bellini and Rossini.
A Composer
From the viewpoint of a prima donna, Rossini’s Otello (first performed at the Teatro del Fondo of Naples on 4 December 1816) had one major flaw: Desdemona appears on stage first in a scena and duettino with her friend Emilia. Later in the opera, to be sure, she has a great deal of solo singing (the second-act finale is essentially an aria for Desdemona, and the first half of the third and final act is devoted to her willow song and prayer), but major singers longed then—and continue to long today—for the opportunity to display themselves in solo song when they first enter. Printed librettos of the epoch demonstrate that Desdemonas of the period often introduced a cavatina before the scena and duettino with Emilia, sometimes omitting the duettino altogether.16 The revival of the opera in Rome during the carnival of 1820 is infamous for Rossini’s having prepared a “happy ending” f
or the opera, at the request of the theater, inserting a duet from his Armida and a final ensemble from Ricciardo e Zoraide. In the same production, however, the prima donna, Girolama Dardanelli, appeared on stage in the first act with a chorus (“Esulta, patria omai”) and cavatina (“Quanto è grato all’alma mia”) borrowed with only a few adjustments in the text from Rossini’s first Neapolitan opera, Elisabetta, regina d’Inghilterra, of 1815. We don’t know whether Dardanelli’s insertion was made with or without Rossini’s approval (he was not present for the rehearsals or the performance).17
It is quite certain, on the other hand, that Rossini had nothing to do with the first production of Otello given in Paris at the Théâtre Italien in 1821. On that occasion, Giuditta Pasta made her debut in the role of Desdemona. She wanted a scena and cavatina, and in the absence of one in the original version of Otello she had no compunctions whatsoever about adding a piece that Rossini had prepared for Malcom in La donna del lago: the recitative “Mura felici” (Happy walls) became “Mura infelici” (Unhappy walls), while the cavatina “Elena! oh tu, che chiamo!” (Ellen! you whom I call) became “Palpita incerta l’alma” (My uncertain soul trembles). Since the original piece was written for a contralto, Pasta transposed the cavatina up a fourth (from E major to A major), and in that key it was printed in all French editions of the opera.18 When she brought the opera to London in 1822, she also brought with her the added scena and cavatina. And when Rossini arrived in London and Paris to supervise productions of Otello in 1824, he did not tamper with this fait accompli.
Yet Pasta’s added cavatina for Otello caused a chain reaction. When Rossini introduced La donna del lago to Paris on 7 September 1824, he replaced Malcom’s scena and cavatina (by now too well known to the Parisians from Otello) with the scena and cavatina for Arsace from Semiramide, “Ah! quel giorno ognor rammento,” creating yet further confusion when Semiramide was first sung in Paris in 1825, where the cavatina was simply omitted.19 By 1827–28, Rossini finally succeeded in setting matters right. Henriette Sontag assumed the role of Desdemona in a revival of Otello early in 1828, and Rossini took the opportunity to get his operas back into the form he had written them. Fearful, though, that Sontag might be criticized for not singing the scena and cavatina to which Pasta had accustomed the French, the composer sent a letter to the Journal de Commerce:
Divas and Scholars Page 29