Divas and Scholars

by Philip Gossett

  Without cuts, the music of Semiramide runs approximately three hours and forty-five minutes. Adding a half-hour intermission and anticipating audience enthusiasm for a work that highlights virtuoso singing (on opening night as much as four minutes of applause greeted certain scenes), we knew that between fifteen and thirty minutes of music needed to be eliminated. Our challenge was to make those cuts in an effective and responsible manner. We can identify four categories of cuts: (1) recitative; (2) choral movements; (3) complete numbers; (4) internal cuts, usually of repeated passages.

  1. RECITATIVE. Performing a little-known opera in an enormous theater that in 1990 still refused to employ supertitles (a situation reversed a decade later) guarantees that details of the dramatic action and subtleties of character motivation will remain mysterious to an overwhelming majority of the audience. Although all the recitative in Semiramide is accompanied by the orchestra, is written with care, and is by Rossini (unlike Il barbiere di Siviglia, where nary a note of the recitative is his),35 significant cuts can be made. There are always gains and losses. After the massive introduction, a short recitative helps clarify (to those who understand the words) some of the action already witnessed. Omitting it saves three minutes. What are the negative results? Some useful dramaturgical development of individual characters is sacrificed, including elements important for later relations between those characters. There is an awkward tonal shift from the key of the introduction, F major, to the key at the beginning of the next musical number, G major. Although Rossini’s recitative mediates between those keys, audience applause after the introduction tends to alleviate the tonal problem. More unfortunate, cutting the recitative means the loss of a dramaturgically effective conclusion to the entire opening of the work. The High Priest, Oroe, is left alone. He addresses another prayer to the gods, and reenters the temple to a reprise of the solemn music that had opened the entire introduction. Still, loss of this recitative and a number of similar passages throughout the opera cannot be said to have damaged the work significantly. Indeed, many of the recitative cuts taken at the Metropolitan Opera were already being made during the 1820s.

  2. CHORAL MOVEMENTS. The massive choral interventions that introduce five of the thirteen numbers of Semiramide, and play an important role in two others, lend an air of solemnity and monumentality to the opera. The indiscriminate removal of choruses changes the character of the work. In some cases, such as the chorus opening Arsace’s scene in the second act, the cuts are not only disfiguring but remove genuinely distinguished music. In other cases, discreet excisions are possible.

  Often Rossini constructs a choral movement by providing an orchestral introduction (A), repeating that introduction with choral parts added (A'), providing a contrasting section (B), and then repeating the opening music with chorus once again (A'); additional cadences bring the chorus to a conclusion. Stage directors such as the late Jean-Pierre Ponnelle, who keep their forces in constant movement, know how to make choruses scenically interesting; those who are more visually oriented, sculpting beautiful stage pictures, such as John Copley, the stage director for Semiramide at the Metropolitan, find them excessively long.

  In deciding what to cut at the Met, dramatic needs, scenic needs, and judgments about musical value were invoked. As a result, three identically constructed choruses (with the form AA'BA', followed by cadences) were handled in three different ways. In the introduction to the first act, the structure was reduced to AA' cadences, a form Rossini regularly uses in other operas. In this case the excision of the B section removed one of the two extended passages in the opera in which the banda sul palco plays independently of the orchestra. (The other such passage, within the duet at the start of the second act, was left untouched.) The musically splendid chorus that opens the first-act finale, on the other hand, was left intact.

  The most intrusive cut occurred in the chorus at the beginning of the finale of the second (and final) act, which takes place in the subterranean tomb of the murdered King Nino. Although the entire chorus was rehearsed, the choral movement itself was judged musically weak and dramatically problematic by both the conductor and the stage director. Yet cutting this passage entirely was impossible for two reasons: first, it quotes material from the overture, transforming it from duple to triple meter, an effect that no one wished to lose (example 6.2). Second, fragments from the music appear in the recitative for Arsace that follows the chorus in the finale, and their appearance there would make no sense if none of this music had appeared earlier. As a result, it was decided to reduce this choral movement to the introductory orchestral statement of the theme (the very first A), with all participation by the chorus eliminated. This completely transformed the significance of the music. Instead of being heard as part of a chorus, it essentially became an orchestral introduction to the accompanied recitative—a structural device not dissimilar to musical techniques Rossini employs elsewhere (Calbo’s great aria in the second act of Maometto II is one of many possible examples). A listener familiar with Rossini’s style, but not with this particular composition, would have had no compelling reason to think the score was anything but intact.

  EXAMPLE 6.2. GIOACHINO ROSSINI, SEMIRAMIDE, SINFONIA, MM. 112–115, AND FINALE SECONDO (N. 13), OPENING CHORUS, MM. 17–20.

  3. COMPLETE NUMBERS. The least painful way to make cuts in an opera is to remove entire numbers. It is also the way Rossini most often countenanced in productions with which he was involved. As we have seen, the tenor, Idreno, is given two arias in Semiramide, one in each act. The pieces are musically attractive, but they function more as concert arias than as integral elements in the drama. In fact, the first-act aria was almost never performed in the nineteenth century, and it was probably omitted already within the 1823 Venetian season. The second-act aria, on the other hand, could not be so lightly removed, since it provides a necessary moment of repose for the mezzosoprano: without it, she would be forced to sing a twenty-minute solo scene, followed immediately by one of the most difficult duets in Italian opera (“Ebben, a te, ferisci”).

  Conductor, scholar, administrators: all of us looked longingly at Idreno’s first-act aria, a full eight minutes toward our temporal goal. The role of Idreno, however, marked the debut at the Metropolitan Opera of Chris Merritt, who in the early part of his career was known principally as a fine Rossini tenor. But Merritt was hired on the explicit understanding that he would sing both arias, and he was going to sing both arias. Ultimately, we were forced to settle for a series of internal cuts in each piece, cuts that displeased me musically and that were stylistically awkward. But opera is about people as well as about art: the singer’s will could not be ignored.

  4. INTERNAL CUTS. Performers of Italian opera often make internal cuts in musical numbers. Singers who understand the technique of vocal ornamentation, on the other hand, are loath to countenance cuts in their solo arias. They understand that repeated passages offer them the opportunity to demonstrate their art. Except for Idreno’s arias all solo arias in Semiramide were performed intact. Some cuts of repeated passages were made in duets, largely in recognition of the sheer endurance required for the soprano, mezzo-soprano, and bass to perform this score. For the most part, these cuts followed practices for which Rossini himself offers ample precedent. In the duet for Arsace and Assur, for example, the concluding cabaletta, “Va, superbo,” consists of a theme sung by the bass, its identical repetition by the mezzo-soprano, a short transition, a repeat of the theme for bass alone, then a repeat for mezzosoprano, now with the bass providing a contrapuntal line (AA trans. AA'). At the Met, the music after the transition was reduced to just the final A, with Arsace singing the melody and Assur providing a counterpoint.

  But at least one cut made in some performances of Semiramide at the Metropolitan Opera belongs to the well-known category of “vanity cuts.” Modern basses tend to conclude Assur’s “mad scene” by leaping up to a high f, rather than descending to the tonic as Rossini wrote. Leaving aside the advisability of th
is practice (a matter that will be discussed in chapter 9), when basses produce a solid f they want it to ring out for several measures. In the orchestral conclusion to Assur’s aria, however, Rossini quotes in abbreviated form an orchestral pattern that had served him earlier in the piece, with an alternation of tonic and dominant harmonies in two-measure groupings. After two measures on F major, the harmony shifts to the dominant chord, and the bass’s sustained high f cannot be continued without creating a harmonic clash (example 6.3). Basses yearn to cut those measures on the dominant, leaving the piece to conclude with seven consecutive measures on the tonic. Indeed, during rehearsals Samuel Ramey was so discontent about performing the music Rossini had written that at one point he showed his displeasure by alternating f and e (example 6.4). Nonetheless, the decision to respect Rossini’s score prevailed, and on opening night Ramey held his high f for two bars, then exited to a resounding and deserved ovation. But singers’ egos need constant stroking. When I returned for a later performance, the orchestral measures on the dominant had disappeared, and the audience had to listen to all those bars of tonic harmony, an effect Rossini would certainly not have countenanced, not even for a bass with a voice as beautiful and powerful as Ramey’s. I like to think, though, that Gioachino was watching these shenanigans: after two bars of high f, Ramey’s voice cracked, leaving the orchestra to repeat the tonic chord in solitary splendor.

  EXAMPLE 6.3. GIOACHINO ROSSINI, SEMIRAMIDE, SCENA, CORO, E ARIA ASSUR (N. 12), MM. 475–479.

  EXAMPLE 6.4. SAMUEL RAMEY’S MODIFICATION OF THE VOCAL CONCLUSION IN THE SCENA, CORO, E ARIA ASSUR (N. 12), MM. 475–479.

  Vocal Variations and Ornamentation

  The introduction of a high note at the end of a musical number, although an effect particularly dear to twentieth-century singers and audiences, was practically unknown to singers active during the years when Rossini was composing his works for Italian theaters and supervising performances of them. But there were many other ways in which singers were expected to modify the notes actually written down by the composer, and a study of various forms of contemporary evidence allows us to understand such practices quite precisely. Our knowledge of these performance techniques has increased enormously over the past thirty years. Disagreements among professionals can (and do) arise over details, and many singers have developed their own personal way of ornamenting their music. Certain techniques are recognizably associated with Marilyn Horne, others with Rockwell Blake, still others with Cecilia Bartoli. Yet it is only when a performance diverges profoundly from the accepted standard that the public or the critics tend to comment on the matter, and that certainly did not happen in the Met’s Semiramide. In most respects the singers participating in these performances provided fine realizations of the score in the appropriate style.

  Almost all the singers, for example, introduced appoggiaturas in an appropriate manner. Some understood from long experience exactly how to handle these small modifications in the vocal line, whose function is to provide a musical accent to match the accent that falls on the next-to-last syllable of a verse with a feminine ending (in a word like “[a]-mo-[re]”). Composers of the first part of the nineteenth century often wrote two equal notes for the next-to-last and last syllable (“[a]-mo-re”) but expected that the first would be modified by the singer (usually sung as the note a tone above). Although the need to employ such appoggiaturas in early nineteenth-century Italian opera is a principle that has been stated often and authoritatively,36 many professional singers, older or younger, continue to arrive at rehearsals with the music learned literally, and a surprising number of conductors do not know the difference. I have sat in early rehearsals for operas at Pesaro, New York, St. Louis, Miami, Chicago, and Rome, marking necessary appoggiaturas in the scores of singers.

  Only at one point in Semiramide, however, did I feel that the performance of appoggiaturas was wrongheaded. It came at the beginning of the cantabile section in the great “mad scene” for Assur. Sam Ramey’s performance of the scene was for the most part so extraordinary that a group of New York wags rechristened the opera “Sammy Ramey Day.” Yet Ramey got it into his head that at the beginning of this cantabile, Rossini’s notation, which clearly implies appoggiaturas in both the first two measures, should be interpreted by singing only the second measure with an appoggiatura (example 6.5). Is this just a matter of taste? No, it is not: Ramey sang the wrong note. An equivalent error would be to perform the beginning of Rigoletto’s aria in the second-act of Verdi’s opera without the appoggiaturas Verdi specified (example 6.6). But Verdi was composing thirty years after Rossini, when singers were less conversant with the conventions, and he consequently wrote out every appoggiatura required. He also sought to differentiate passages where he did not want appoggiaturas, often passages involving declamation over an orchestral melody, as opposed to recitative-like passages.

  EXAMPLE 6.5. GIOACHINO ROSSINI, SEMIRAMIDE, SCENA, CORO, E ARIA ASSUR (N. 12), MM. 278–279.

  Apart from appoggiaturas, the style of nineteenth-century Italian opera before Verdi (with some important extensions even into Verdi) required singers to add appropriate cadenzas where the composer did not write them in full, or to modify those the composer did write to suit their individual vocal abilities. The practice also was based on the assumption that singers would vary certain repeated passages, even though these were notated identically by the composer. Such interventions were not considered “optional”: the structure and style of the music were conceived in order to favor them. Nineteenth-century sources, including manuscripts prepared by Rossini for singers of his time, provide sets of variations used by singers from the first half of the century, and theoretical treatises on singing offer an enormous amount of information about these practices. There is no one “correct” way to proceed. Rather, the art of ornamentation is a creative interaction between historical style and individual personality. There are as many different ways appropriately to add cadenzas and to vary repeated lines as there are singers, and most performers today operate within perfectly acceptable boundaries.

  EXAMPLE 6.6. GIUSEPPE VERDI, RIGOLETTO, SCENA ED ARIA RIGOLETTO (N. 9), MM. 78–79.

  Certainly all the singers in the Metropolitan Opera’s Semiramide were well versed in these techniques, even though they went about their tasks quite differently. Some prepared their own materials anew; others combined variations and cadenzas they had sung before with new ones. Some had everything written out well in advance; others developed material during rehearsals. In one duet for Semiramide and Arsace, Lella Cuberli and Marilyn Horne sang a splendid joint cadenza that Richard Bonynge had developed for performances with Joan Sutherland and Horne during the 1960s. For the Met, I prepared a complete set of suggestions for the entire opera, which were sent to the singers before rehearsals began in New York. They used what they liked, rejected what they didn’t, and made substitutions and emendations to suit their needs. During rehearsals we fixed places that didn’t work. All eyes, of course, were on the conductor, and when Maestro Conlon grimaced, it was back to the drawing board. The last thing anyone was concerned about was “scholarship.”

  The world of the theater is not a place where one pays obeisance to a written score, but rather a place where one finds real singers performing with real orchestral musicians; audiences with trains and buses to catch; administrators who must watch both the cash box and the artistic product; successive generations of critics, each of which invokes a past golden age but fails to appreciate its own; costumes that come apart just as the heroine launches into a cabaletta, forcing her to sing complicated runs and variations while worrying that she’ll soon be naked; choruses that must be shepherded onto and off the stage because the director wants to make a “pretty picture”; massively inappropriate sets that must be changed in the middle of a musical number; ghosts that emerge in clouds of smoke sending prima donnas to their bed with coughs; tenors that jet from one theater to another and arrive for rehearsals barely knowing the music. That’s what the
world of the theater is like. And musicologists who truly love opera would have it no other way.

  An edition is not a performance, a performance is not an edition, and scholarship and operatic effectiveness are not mutually exclusive. Once we know the score, once a critical edition enters the world of the theater, it is used in much the same way as any other edition is used. It is to the issues that arise when performing Italian opera from any edition, critical or otherwise, that the rest of this book is dedicated.

  PART II

  Performing the Opera

  7

  CHOOSING A VERSION

  MODERN PERFORMERS AND HISTORY

  The relationship between modern performers and the history that should sustain them has rarely been so tormented as in our current musical environment. Increasing numbers of critics, conductors, scholars, and instrumentalists are calling into question many of the basic tenets that guided musical practice through much of the twentieth century. We distrust tradition, we distrust instinct, we distrust scholarship, we distrust our ability to judge among competing aesthetic positions and interpretive strategies. These concerns have become ever more troublesome as they have touched the performance of music written in the late eighteenth and the nineteenth centuries, the music we grew up believing was “ours,” a repertory that seems to recede further and further into the past with the new millennium.