Divas and Scholars

by Philip Gossett

  Indeed, I would press the implied question of aesthetic judgment further. Where a composer has cobbled together what we judge to be a patently inferior revised version of an opera because he lacked adequate singers to perform the original music or bowed to external pressure to permit artistically questionable changes, must a critical edition empower performers to present the inferior version (by preparing scores, parts, etc.), or does it suffice to describe the changes?

  Just as a powerful prima donna can impose her will in evaluating alternative versions of a work, so too an edition of an opera can stack the deck. Placing music in the main body of a score or in an appendix might seem implicitly to favor one choice rather than another among aesthetically defensible versions, but providing only descriptions or incomplete materials places tighter constraints on performers. Rossini, for example, as we have seen, prepared two different endings to his Tancredi in 1813: a happy ensemble and a tragic, restrained death scene. Neither suited the desires of Giuditta Pasta, who in 1826 insisted to the composer that the opera should conclude with a showy aria for the hero en travesti. When Rossini equivocated, the imperious singer inserted a piece by another composer, Giuseppe Nicolini, but asked Rossini to write ornaments for her, which he obligingly did. The new critical edition prints Rossini’s variants in the context of the piano reduction of the piece into which he wrote them, but does not present a full orchestral score. From information provided in the introduction, a determined prima donna could still have this version reconstructed, but as the editor I did not feel obliged to facilitate her task.109

  If Italian opera in the twenty-first century is to remain a vital tradition, we need to recognize that part of its historical vitality lay in the process by which composers revised their operas for particular productions. The versions for which he was directly responsible offer modern performers a set of alternatives from which to make their choices. Measuring these choices against our grid of aesthetic matters, historical circumstances, and modern conditions does not ensure “correct” solutions. Indeed, the elements within this grid are so varied that no choice can ever be universally appropriate. But at least this approach recognizes that Italian opera is to be performed neither according to the idol of “tradition” nor according to the idol of “historical reconstruction.” If we accept a multidimensional approach to performance decisions, we ground our theories in the actual world of opera, a world of compromise and uncertainty, a world of strong opinions and even stronger egos. We should aim to provide, not a museum for the petrification of performances of the past, but a re-creation (within the context of our own social structures) of the characteristics that made Italian opera a vital art form in the nineteenth century and can help it remain so today.

  Although Le nozze di Figaro does not strictly fit into the subject of this book, I cannot resist adding a word about the tempest that blew over New York’s Metropolitan Opera in October and November 1998, when the Susanna (Cecilia Bartoli), with the approval of the conductor (James Levine), decided to introduce at some performances two arias that were not in the original score but were inserted by Mozart himself into the opera after its first production. Figaro was first performed at the Burgtheater of Vienna on 10 May 1786. On the occasion of a revival at the same theater on 29 August 1789, the composer provided replacements for Susanna’s original arias in the second act (“Venite, inginocchiatevi,” replaced by “Un moto di gioia”) and the fourth act (“Deh vieni non tardar,” replaced by “Al desio di chi t’adora”). He did so, of course, because he had a new Susanna, Adriana Ferrarese, hardly an inferior artist to the original Nancy Storace: in 1790, after all, Mozart wrote for la Ferrarese the role of Fiordiligi in Così fan tutte.110 But she was a different singer and must have been a very different presence onstage, if we can judge from Fiordiligi’s great aria “Come scoglio!” Mozart was not altogether enamored of her, as he wrote to his wife, Constanze, on 19 August 1789: “The little aria, which I composed for Madame Ferrarese [“Un moto di gioia”], ought, I think, to be a success, provided she is able to sing it in an artless manner, which, however, I very much doubt. She herself liked it very much. I have just lunched at her house.”111 Still, he did write the two arias, and Ferrarese performed them with success. Whether or not Mozart wished them consigned to the dustbin of history we will never know, since there is no documentation whatsoever, but the arias have long been familiar to anyone with the slightest knowledge of the history of Le nozze di Figaro.

  At the Metropolitan Opera, on the other hand, these arias gave rise to a public scandal when the stage director, Jonathan Miller, objected to their introduction. As I have tried to argue in this chapter, there is no “right” and “wrong” about such substitutions: arguments can be brought to bear from many different directions. Normally such questions are discussed backstage—whether amicably or not—and the matters are resolved before opening night. Miller chose instead to go public. Le nozze di Figaro is an opera regularly seen in our opera houses, season after season. Experimenting with Mozart’s own added arias, even from the stage of the premiere opera house in the United States, is hardly an example of lèse-majesté. It was not Bartoli’s playful substitution but Miller’s almost religious fervor on the subject and his misrepresentation of the historical evidence that were unacceptable. Still smarting over the experience almost five years later, Miller gave an interview to the Paris Review in 2003, in which he described the arias as “concert pieces” which “should be omitted from the stage production, because the words don’t match what goes on: they have nothing to do with the scene.”112 His description of the arias as “concert pieces” is historically wrong, whereas his aesthetic judgment is hardly absolute: it depends on how you conceive the scene, on how the stage director develops the action, on how the singers portray the characters. Besides, if Bartoli was “beguiling” the audience, as Miller himself affirms, something worthwhile had clearly happened in the theater.113

  8

  SERAFIN’S SCISSORS

  TO CUT OR NOT TO CUT?

  Should Italian operas be performed complete, or is it desirable to omit certain passages? The existence of multiple versions of Italian operas written during the first half of the nineteenth century already complicates the integrity of these operas, their status as “works.” If Bellini eliminated completed passages from I puritani during the rehearsal period in Paris, as we saw in chapter 7, what does it mean to talk about performing the opera complete? If Rossini—before he left Paris following the first performances of Guillaume Tell—struck from his opera two dances, several recitatives, and an aria for Jemmy (as he joyfully waits for his father to shoot the apple from his head), what should we be performing complete?1 If Verdi, returning to earlier works (Macbeth, Stiffelio, Il trovatore, Simon Boccanegra) between the late 1850s and the early 1880s, omitted entire sections of these scores and rewrote others, what search for textual purity would drive critics, scholars, and musicians to insist on integral performances?

  Making cuts in nineteenth-century Italian operas has a long pedigree.2 To pretend otherwise would falsify the historical performance record and impose a foreign aesthetic. Rarely, though, has the activity been subject to a critique that seeks to understand how that practice has been and continues to be motivated or the ways in which cuts have been effected. When a conductor omits a passage or an entire number because he honestly doesn’t like it, I may disagree while respecting his motivation. When a conductor makes a cut because it is “traditional,” on the other hand, he is acting without artistic integrity.3

  Decisions about cutting, furthermore, are unrelated to what editions a performer may be using, except insofar as the critical edition may make available passages omitted from earlier published scores. Before there was any talk about critical editions of Italian opera, some performers included the “Wolf’s Crag” scene in Lucia di Lammermoor, some did not; some tenors sang the cabalettas of their arias in Rigoletto and La traviata, some did not. While critical editions—prepared with as much knowl
edge as can be acquired about an opera and its textual history—tend to instill greater confidence in the credibility of the printed score, no one should need to be reminded that a performance is not an edition. Yet most objections to using critical editions in the opera house are centered not on the editions themselves but on their presumed implications for performance. There are conductors and singers who continue to believe, against all evidence, that a critical edition restricts their options, compels them to perform an opera come scritto (as written), abjuring cuts and performance traditions and slavishly following the dictates of the printed text. Nothing could be more remote from the position of those who prepare the editions. These fears can be understood best as a misreading of the aesthetic views of two conductors often associated with the editions, Claudio Abbado and Riccardo Muti.

  Abbado was the first to use the new critical editions of Rossini’s comic operas, and in his early performances he expected singers to adhere to the letter of the written text, refusing to countenance any cuts, variations, or cadenzas other than minimal ones. Given the sorry state of most Rossini performances before Abbado’s work (with the exception of that most Rossinian of earlier twentieth-century conductors, Vittorio Gui), this proved a welcome sweeping away of the abuses regularly heaped upon these scores. Anyone who has heard the 1992 Abbado recordings of Il viaggio a Reims and Il barbiere di Siviglia, however, knows that he has long since abandoned such a position. Indeed, one might well argue that he has occasionally gone overboard. Placido Domingo as Figaro was hardly a choice suggested by the nature of the written score, whatever Domingo’s extraordinary skills. And with Kathleen Battle as Rosina, Abbado retreated from the mezzo-soprano/contralto of Rossini’s original to the vocal rescoring dear to coloratura sopranos in the first half of the twentieth century.4 For reasons that have nothing to do with scholarly purity, I find the aesthetic results unconvincing. Also, in inserting into Il viaggio a Reims a citation of La Marseillaise—played by a trumpet, no less—Abbado served Rossini badly, an enormous irony, given the musical intelligence and spirit he lavished on other parts of the score. There are now legions of opera fans who think that Rossini bravely (or ironically) quoted the French Revolutionary anthem of 1792 in an 1825 performance of an opera at which the new King Charles X was present.5

  Riccardo Muti, on the other hand, who has worked particularly closely with the Verdi critical edition, has remained ferocious (some think too ferocious) on the subject of singers’ interpolations, refusing unwritten high notes, traditional trills and turns, and the like. There is much to be said for his position: many opera houses, as we have seen in chapter 4, continue to lament that they cannot cast the role of Manrico in Il trovatore, when what they really mean is that they can find no one to sing what passes today for the role of Manrico. But Muti himself is anything but rigid in his interpretation of the Verdi editions. The famous chorus of Hebrew slaves in Nabucco, “Va pensiero,” concludes with a pianissimo chord in the chorus, held for one bar and cutting off at the downbeat of the next, with two further pizzicato chords in the strings to bring the piece to a close. Presenting the first performances of the new edition to open the season at La Scala in 1986, Muti had his chorus hold that sonority for so long, with such a spectacular diminuendo, that even the audience gasped for breath. And after an interminable ovation, he gave the public the unthinkable: an encore.6

  Once any edition enters the opera house, it is subject to interpretation and emendation at every point. As in the case of choosing a version, performance decisions are dependent on the three-dimensional grid of aesthetic criteria, historical circumstances, and practical conditions, as described in chapter 7. These dimensions are not independent: our aesthetic criteria are in part determined by our historical knowledge, practical conditions tend to be closely related to our aesthetic criteria, and our view of history cannot be separated from who we are and why we are seeking information about performance decisions made at the time a work was composed. In a living art, there are no correct or definitive answers about performance decisions. Every situation is different, artists change, the same artists mature (or at least get older), instrumentalists have different characteristics from one pit orchestra to another. When a last-minute replacement tenor in L’Italiana in Algeri proved inadequate at the Rossini Festival of 1994, no one (administrators, conductor David Robertson, stage director Dario Fo, or scholars) had the slightest doubt that his second-act solo—a replacement piece not widely known by the public—should be snipped out of the score.7 We were only sorry that we couldn’t do the same thing with Lindoro’s first-act cavatina, but “Languir per una bella” was too familiar, too integral to the score, to be omitted without incurring the wrath of a public all too willing to show its displeasure by rotund booing.

  Discussions of cuts in modern performances are usually couched in aesthetic or practical terms that dehistoricize the process, or rather historicize it no further than the immediate past. It is important, instead, to measure cuts both against the social circumstances in which operas were originally prepared and against the conditions in which they are performed today. What is desirable or permissible may not be the same in San Francisco and Milan, in an American theater with supertitles and one without them, in an opera house that performs a work in the language of the audience and one that does not, in a barn that seats four thousand spectators and an intimate space that seats four hundred, at a festival in which an audience assembles from all over the world and in a regular theatrical season in which everyone has to get to work the next morning.

  THE INTEGRITY OF AN OPERA AS A HISTORICAL PROBLEM

  Our attitude to cuts depends on our perception of the integrity of an opera. The existence of multiple versions already problematizes that perception, as does the nature of the theatrical system for which operas were written. Equally significant is the historical position of a work within a stylistically defined continuum of nineteenth-century Italian operas. Most discussion about cuts is really about the acceptable limits within which a particular opera can be stylistically displaced along that continuum.

  Every work of art belonging to a vital and living tradition is created at a specific moment in the history of that tradition. Particular solutions to political, social, compositional, and stylistic issues, often at cross purposes, inform the decisions of each creative artist. Some decisions are fully conscious, others are unconscious adaptations to the immediate environment. Nor is our understanding of a work within its historical tradition unalterably fixed. For musicians and operagoers living in the middle of the nineteenth century, each new opera by Verdi changed inexorably their vision of the immediate past as they sought to reinterpret their understanding of works by Bellini or Donizetti in the terms imposed by Verdi’s accomplishment. Historical models developed by twentieth-century scholars are no less subject to reinterpretation: the widespread revival of Rossini’s serious operas over the past twenty years, after a hiatus of more than a century, has had a profound effect on the way we hear the music of Bellini and Donizetti. It forces us to confront their works with operas they knew well but about which, until recently, we were ignorant. A more profound knowledge of Donizetti’s operas of the early 1840s (works such as Adelia, Maria Padilla, Caterina Cornaro, and Maria di Rohan) will inevitably lead to a similar reevaluation of the operas of Verdi’s first decade.

  Various stylistic elements in individual works are tied to their position in this continuum. It would be as unthinkable for the protagonist of Verdi’s Luisa Miller of 1849 to present herself in an aria with the fluid structure and dramaturgical immediacy of “Mi chiamano Mimi” from Puccini’s La Bohème of 1896, as it would be for Cio-cio san in Madama Butterfly of 1904 to arrive singing a piece patterned after a great cavatina of the 1830s, such as Norma’s “Casta Diva.” It would be as unthinkable for the two servants, Ruiz and Ines, in Verdi’s Il trovatore of 1852 to have a lengthy conversation about the events befalling their masters, to fall in love, and to sing arie di sorbetto about their feelings
(as similar characters do in Rossini’s Ciro in Babilonia of 1812), as it would be for a chorus of guards at the beginning of the last act of Tosca in 1900 to break into a ditty such as “Squilli, echeggi la tromba guerriera” from the third act of Il trovatore.

  On the other hand, not every musical or dramaturgical element in a work is crucial to our perception of that work’s integrity. When we make a decision about cuts in the modern performance of an opera, we are making a judgment about integrity. The more we believe in the work’s integrity, the more hesitant we will be about making cuts. Yet the historical record suggests that few works for the theater lend themselves to absolutist judgments. Rather, there exists considerable evidence as to which elements seem more fragile at a given moment. Contemporary sources offer valuable insight about how operas were treated in their own time, and in some cases we can even reconstruct cuts a composer made during rehearsals.

  Although many of his operas continued to pay obeisance to eighteenth-century dramaturgical models, Rossini had little interest in those conventions according to which secondary characters participate in the action; nor was he concerned with scenes of secco recitative that elaborately narrate past events or comment on recent ones. Donizetti’s autograph manuscripts and letters, on the other hand, demonstrate his continued efforts to avoid the regular repetition of cadential formulas so characteristic of Rossini. In manuscript after manuscript Donizetti is drawn instinctively to these procedures, then crosses out the offending measures so as to streamline the musical dramaturgy.8 For Verdi one of the most problematic elements was the cabaletta, whose formal design made sense for Rossini and Bellini, associated as it was with the practice of singer ornamentation. Increasingly it seemed useless baggage from the past to Verdi’s contemporaries and to the composer himself. Revising his Simon Boccanegra from 1857 for performance in 1881, he tried to free it—not always successfully—from this earlier practice. Even when he did walk the path of the cabaletta in Aida, he constantly prodded his librettist, Antonio Ghislanzoni, with advice like, “If you could find a form somewhat more novel for the cabaletta, this duet would be perfect.”9