Most cuts introduced by thoughtful performers in the nineteenth century and in the modern world are not arbitrary manipulations of unstable texts, but rather attempts to eliminate elements considered to be least significant or characteristic for a work’s aesthetic integrity and historical position. In deciding which cuts to countenance in modern performance, we must determine how far to allow this elimination to proceed. At what point do such cuts distort the work? When does the process of cutting an opera result in such a distortion of its aesthetic premises that it would be better to abandon the effort? When should we quit trying to rescue Humpty Dumpty?
The significance of individual cuts, furthermore, will vary for different participants in an operatic performance. Too often the issues are presented as if everyone were working from a common viewpoint.10 Aesthetic issues of musical quality rarely play a role in decisions about cutting recitative, but for a stage director the effect of recitative on the coherence of the dramatic action can be crucial. The stage director, on the other hand, needn’t worry whether or not a secondary character sings an aria di sorbetto, but the practical capabilities of the performer (and clauses in his or her contract) may determine whether the piece is included.
In the following discussion I will consider three categories of cuts: recitative (either entire scenes or smaller sections); complete musical numbers; and portions of musical numbers.
MAKING CUTS: RECITATIVE
Early during his years in Naples, Rossini prepared an opera buffa, entitled La gazzetta (The Newspaper), for the Teatro dei Fiorentini, where it was first performed on 26 September 1816. Based on a Carlo Goldoni comedy, Il matrimonio per concorso (The Marriage Competition), it is a peculiar work. About 50 percent of the concerted music was derived from earlier operas still unknown in Naples, but Rossini nonetheless wrote out every note of this recycled music afresh, reorchestrating it, adapting it to new words, integrating it with new musical ideas.11 The opera also has vast stretches of secco recitative (including large sections in Neapolitan dialect), not a note of which is by the composer. Some of the text is hilarious, for most situations and details of dialogue reach back to Goldoni. A stage director with a particular affinity for verbal play might luxuriate in this recitative; a stage director like Dario Fo, on the other hand, with strong roots in the commedia dell’arte tradition and physical comedy, was less amused. For his production at the Rossini Opera Festival in August 2001 he insisted on massive cuts.
With the performance of the new critical edition of La gazzetta (edited by Fabrizio Scipioni and myself) less than two months away, I traveled to Milan to discuss these cuts with Fo. His apartment was in an uproar, for his entire entourage was about to transfer its operation to Pesaro, but Fo was totally absorbed in showing off his ideas for La gazzetta. He is a very visually oriented man of the theater, who prepares a project by drawing countless renderings of individual scenes, arranging the entrances and exits of characters and mimes, twisting and turning scenic elements in a profusion of inventive designs. He does not work from a musical score but from a libretto. While he conceded that concerted numbers had to be sacrosanct, he treated secco recitative as if it were expendable, like spoken dialogue. When I tried to understand exactly which text he wanted to cut, he found the question somewhat irritating. Couldn’t that be worked out in the theater? No, I tried to explain. Not only did the singers need to learn the music together with the words, but making cuts would require us to rewrite the music, so that harmonies made sense and the rhythm flowed easily. And, no, he couldn’t cut two words here and three words there without taking into account what those omissions would do to the poetic structure, rhyme scheme, and the musical setting. Most of all, if he waited until the first rehearsal to inform his singers of his intended cuts (by which point they would have memorized an enormous quantity of recitative), not much of Dario Fo would remain when they had finished with him.
Italian operas from the first half of the nineteenth century are what we call “number operas”: essentially closed musical forms separated by recitative, even if composers increasingly sought to override this separation. The division between recitative and numbers is not imposed by the composer. It is inherent in the structure of librettos, whose poets wrote their verses for recitative in the classical patterns of versi sciolti (free verses) of only occasionally rhymed settenari and endecasillabi, as opposed to versi lirici (lyric verses), rhymed poetry using regular poetic meters, employed for the musical numbers.
During the first two decades of the nineteenth century in Italy, recitative was almost always secco, accompanied by a keyboard instrument (usually the fortepiano, ancestor of the modern piano), with a violoncello and double bass doubling (and, in the case of the cello, harmonizing) the bass line.12 Influenced in particular by French operatic practice, strongly present in Naples during the reign of Napoleon’s brother-in-law Murat (1808–15), Neapolitan theaters in the 1810s began to require all recitative in serious operas to be accompanied by the orchestra. In 1813, for example, Giovanni Simone Mayr prepared secco recitative for his Medea in Corinto as he began composing the score from his home in Bergamo. When he arrived in Naples to begin rehearsals, he was compelled to rewrite all this recitative with orchestral accompaniment.13 As of the mid-1820s, when Rossini’s Neapolitan operas were being heard in theaters throughout Italy, the practice of writing only accompanied recitative in serious operas took root, although the use of secco recitative in comic opera persisted for several decades. Rossini’s last serious opera with secco recitative is Bianca e Falliero, which had its premiere at La Scala to open the carnival season of 1819–20. Significantly, the secco recitative in this opera was written entirely by a collaborator, not by Rossini himself.14 By the time he wrote Semiramide for the Teatro La Fenice of Venice during the carnival season of 1822–23, however, Rossini employed only accompanied recitative in this opera seria, and he personally prepared every note of it. While not every composer followed his lead during the 1820s, in the 1830s serious operas with secco recitative were practically nonexistent in Italy.
Recitative (whether accompanied or secco) had multiple functions in Italian opera, and recognizing them helps us develop principles for introducing cuts into modern performances. First, recitative has a narrative function: it presents developments in the plot, crucial to the dramatic structure of the opera, as they occur. Second, recitative has an expository purpose: it explains earlier events that prepared present dramatic configurations or comments on those events. Third, recitative has a formal function: it provides repose in the dramatic and emotional progress of an opera by separating closed musical numbers, which probe character, explore intense personal reactions to dramatic events, or carry emotionally charged confrontations. Fourth, recitative (and occasionally arias) can have a scenic aim, necessitated by early nineteenth-century stagecraft, in which set changes within an act were carried out in full view of the audience (or behind a drop curtain while the performance continued in front). To avoid lengthy pauses after a dramatic “long” scene (one that employs the entire depth of the stage), a bridging dramatic scene performed in front of a drop curtain (a “short” scene, often in recitative) allowed time for a new long or medium scene to be put into place.15
The relative weight of these functions changed during the course of the first half of the nineteenth century. Early nineteenth-century serious operas, such as Rossini’s Tancredi, employ a full complement of handmaidens, confidants, and lieutenants, following the practices of eighteenth-century French drama (the source of many librettos) and Italian opera to librettos by Pietro Metastasio or his contemporaries. They typically receive the laments, share the joys, and execute the projects of the principal characters: Amenaide’s Isaura and Tancredi’s Roggiero have ample opportunity to discuss the fate of their masters. In Rossini’s Ciro in Babilonia, as mentioned above, secondary characters actually develop sentimental attachments to one another, which must be resolved before the drama concludes.
Rossini considered the co
mposition of secco recitative to be a secondary activity, and for his mature operas he usually entrusted the job to another musician. Only in some early works, written before 1815, and in the unusual Viaggio a Reims of 1825, did Rossini prepare secco recitative himself. In his first surviving autograph, La scala di seta, an 1812 one-act farsa, all the recitative is his own. But in his very next opera, La pietra del paragone, a two-act comedy written for La Scala later that same year, the numbing quantity of secco recitative invented by the librettist discouraged the composer: he prepared music for its first act, but foisted the second onto a collaborator. In fact, Rossini prepared essentially none of the secco recitative in his most famous comic operas, L’Italiana in Algeri, Il Turco in Italia, Il barbiere di Siviglia, and La Cenerentola, nor in semiserious or serious operas such as La gazza ladra, Bianca e Falliero, and Matilde di Shabran.
That Rossini did not prepare secco recitative for Il barbiere di Siviglia or La Cenerentola does not mean the job was poorly done; the musical setting is often sparkling and witty, the style fully in line with Rossini’s procedures in those earlier operas for which he composed recitative (La scala di seta, the first act of La pietra del paragone, and Tancredi). Il Turco in Italia and La gazza ladra, both prepared for the Teatro alla Scala, offer a marked contrast. The classical rules for recitative called for notation in 4/4 meter with strong syllables falling on the first and third beats (always with the understanding that delivery would be rhythmically free and that there was no real “beat” involved). Rossini’s own recitatives generally accord with this model. But his collaborators in Il Turco in Italia and La gazza ladra followed a modified convention, observing the accents on the first and third beats only when the harmony changes, while allowing strong and weak syllables to fall wherever they might during continuous recitation over a single harmony. (A still looser modified convention, adopted often by Donizetti and by Verdi in Un giorno di regno, his only opera with secco recitative, was to abandon barlines altogether, or to place them only at harmony changes, creating long run-on “bars” and abandoning all pretense of meter.) Woe to a performer who tries to sing these recitatives in a strict 4/4 meter with accents on the first and third beats. Italian singers instinctively understand this, but American singers need to be taught. Samuel Ramey and Lella Cuberli, in the first Pesaro production of Il Turco in Italia during the summer of 1983, painstakingly had to unlearn the accent patterns they had assumed from the notated rhythm. Nor is it possible to fix these rhythms, as some later editions tried to do: such manipulations falsified the temporal relations between syllables, a far worse sin.
Despite the fluidity of the Barbiere recitatives, however, the pages devoted to this music by Serafin and Toni (to be discussed more fully below) are absurd.16 Since these authors rather like the recitatives, they want the settings to be by Rossini, and so they invent a scenario in which the manuscripts within Rossini’s autograph of the opera are the “fair copy” of a previous autograph manuscript (now lost). This fair copy “must have been” prepared by an associate from Rossini’s “vocal score,” because Rossini had neither the time nor the inclination to write them out a second time. How many distortions and errors can dance on the head of a pin? The recitatives in Rossini’s autograph score of Il barbiere are composing scores, not copyists’ renderings; Rossini never prepared a preliminary “vocal score” before writing out his full score; and there is no evidence that Rossini ever wrote any part of his operas, not even the most complex ensembles, “a second time.” But we will return to Toni and Serafin below. The crucial issue, after all, is not who wrote these recitatives. That Rossini prepared all the secco recitative for Tancredi does not mean that every note must be played in a modern production: some dialogue, reflecting practices already old-fashioned in 1813, was always eliminated in contemporary performances and can quite properly be eliminated today.
Secco recitative can be declaimed quickly; accompanied recitative, punctuated by orchestral chords or, increasingly, enriched with sustained singing or playing, moves more slowly. Thus, the length of the verbal text decreases when recitative is accompanied, while the musical importance of these passages increases. During his Neapolitan years, Rossini invented a style for using accompanied recitative throughout an opera (and not only for the elaborate scenas that precede major arias). Much of the recitative in Elisabetta, regina d’Inghilterra, of 1815 is little more than secco recitative with interpolated string chords. (In the first performances of the new critical edition at the Rossini Opera Festival during the summer of 2004,17 the conductor, Renato Palumbo, strove to make these recitatives highly expressive; that the music sometimes resisted his efforts reflects the limitations of Rossini’s writing.) By the 1816 Otello the recitative is more flexible, and in the third act it reaches a consistently high level. In later works (the 1818 Ermione, 1820 Maometto II, and 1822 Zelmira), Rossini’s accompanied recitative achieves a level of musical and dramatic intensity far different from that of his contemporaries. It comes as no surprise, then, that for operas employing accompanied recitative throughout, Rossini tended to engage collaborators only when under intense time pressure. We know, for example, that La donna del lago, which had its premiere at the Teatro San Carlo on 24 October 1819, was hurriedly planned by Rossini after Berlin authorities refused Gasparo Spontini permission to fulfill the contract requiring him to prepare a new opera for Naples. To help meet the unexpected deadline he was facing, Rossini assigned some accompanied recitative for La donna del lago (as well as the aria for Duglas, “Taci, lo voglio”) to a collaborator.18
This reduction in the amount of recitative continued throughout the first half of the century. When Verdi first collaborated with the librettist Francesco Maria Piave on Ernani, he complained to Guglielmo Brenna of the excessive number of verses of recitative that Piave had written.19 He persisted in recommending brevity to his librettist, year after year. Only a little more than a month before the premiere of La traviata in 1853, he commented to the president of the Teatro La Fenice, “Piave has not yet finished polishing La traviata; and even in the things he has finished there are some long-winded passages that will put the public to sleep, especially near the end, which must be very quick if it is to have an effect.”20 The function of recitative changed significantly as the century unfolded. With librettists and composers turning to new dramatic sources (particularly the Romantic theater), and with secco recitative giving way to accompanied recitative, dialogue involving secondary characters was reduced to a minimum. In works like Rigoletto and La traviata, a few brief phrases must suffice to characterize the Count of Ceprano, Merullo, the Baron Douphol, and Flora Bervoix. Only sensitive performances, both vocally and scenically, can bring such characters to life. By midcentury, the emblematic secondary role in Italian opera is the servant in Flora’s party in La traviata: his classic and single line is “La cena è pronta” (Dinner is served).
There were also important changes in the external shape of an opera. The two-act design that dominated Italian opera during the first three decades of the century (itself a reform of eighteenth-century practice, where most operas were in three acts) grew to three or even four acts by 1850. Of Rossini’s thirty-four Italian operas, only three Neapolitan works (Otello, Armida, and Mosè in Egitto) have more than two acts; of Verdi’s roughly twenty-five Italian operas, only the early Oberto and Un giorno di regno have but two. Although it may seem paradoxical, this shift did not increase the amount of music in an opera: if anything, there was sometimes less music, for the intermissions obviated the need for additional scenes to bridge set changes within an act.
When Rossini wrote Italian operas in three acts, there were always unusual circumstances. Time was needed to prepare the final scene of Mosè in Egitto, in which the Israelites cross the Red Sea, or the beginning of the final act of Armida, where the action shifts from the second act’s pleasure palace to an enchanted garden. Rossini and his librettists could have provided bridging scenes but preferred to introduce an intermission to enhance the
dramatic and emotional resonance of the work. On scenic grounds alone, the third act of Rossini’s Otello, which takes place in Desdemona’s bedchamber (a “short” scene), could easily have followed directly the aria she sings to conclude the second act (a “long” scene). Surely the composer wished to separate this final scene of the opera not only for dramaturgical reasons but also to allow Desdemona much-needed rest. A few years earlier, however, Rossini and his librettist might have bridged the gap by introducing a short scene (both temporally and spatially), containing recitative or an aria di sorbetto. Indeed, in the first London performances of Otello, at the King’s Theatre in 1822, without the participation of Rossini, a scena e duetto for Emilia (Desdemona’s “handmaiden”) and Emilio (Desdemona’s father) was introduced, making it possible for the opera to be performed in two acts. (At least they don’t fall in love.) When Rossini himself took responsibility for a London revival in 1824, he promptly restored Otello to three acts.21
The 1830s were a period of transition, during which the problem of bridging scenes remained. The final act of Lucia di Lammermoor provides a perfect example. Lucia’s mad scene—after she has murdered her husband, Arturo, on their wedding night—takes place in the great hall of the castle. In the final scene of the opera, Lucia’s beloved Edgardo awaits her brother Enrico, whom he has challenged to a duel, among the tombs of Ravenswood. When Edgardo learns of Lucia’s death, he kills himself. Donizetti and his librettist had two problems: (1) How, after the mad scene, do they get Lucia, the chorus, and everyone else off stage? (2) How do they most effectively allow time to prepare the new setting for the final scene? They solved these problems by introducing accompanied recitative to bridge the scenes. In it, Lucia is led off by her ladies, and the remainder of the chorus exits with them. Overwhelmed by the events, Enrico too departs, while Raimondo and Normanno (Lucia’s tutor and the captain of Enrico’s guards) engage in dialogue about the appalling events they have witnessed. This dialogue surely took place in front of a drop curtain, behind which the final scene could be prepared. The recitative, in short, had an exquisitely practical purpose: absent these practical requirements, it could not be more superfluous. On the other hand, cutting the scene without solving the practical issue is an absurdity: is it really better for an audience to sit, bored, during a scene change than to hear a well-performed recitative fill in details of the plot?
Divas and Scholars Page 35