Divas and Scholars

by Philip Gossett

  As always in the world of opera, Muti’s decision reflected diverse factors, artistic, practical, commercial. Artistically, he was concerned that his largely American and Italian cast would prove incapable of pronouncing French idiomatically. That problem had emerged—too prominently and too recently to be ignored—with respect to a recording in French of Verdi’s Don Carlos, with the orchestra of La Scala under the direction of Claudio Abbado.1 Abbado’s cast, in fact, had featured a largely Italian group of singers, including Katia Riciarelli, Lucia Valentini-Terrani, Ruggero Raimondi, and Leo Nucci; the non-Italians—Placido Domingo as Don Carlos and Nicolai Ghiaurov as the Grand Inquisitor—did little to change the picture. Criticism of the French pronunciation was often severe.2 Historically, after all, the language of opera performances had much to do with who was singing, and there was a strong sense that operas were more effective when performed by singers employing their own language.

  Furthermore, from a practical standpoint, the entire La Scala troupe was to tour Japan during the autumn of 1988, so that the chorus needed to begin studying much earlier than normal and before materials for the new edition were available. It was therefore prudent to have the chorus relearn the Italian many were said already to know, at least in part, introducing corrections in the weeks preceding the performance. And finally, in commercial terms, opera in French by Italian composers has always perplexed the Italian public: while they are prepared to engage Bizet’s Carmen in the original language, they remain resistant to French-language performances of Rossini’s Le Siège de Corinthe, Donizetti’s La Favorite, or Verdi’s Les Vêpres siciliennes. When Muti led a fine performance of Rossini’s Moïse in the original French to open the 2003–4 season at La Scala, he still faced criticism from the critic of the Corriere della Sera, Paolo Isotta, for not performing the opera in Italian.

  And yet there was something inherently paradoxical in this situation: here was a conductor with a legendary reputation for his attention to dynamic markings in an edition, rhythmic details, pitch, expression, orchestral color, who was nonetheless prepared to have his singers employ a deeply problematic Italian translation. What conception of the function of text in opera would lead a Riccardo Muti to accept this compromise when he continued to find intolerable the much more limited compromise of a single high c at the end of “Di quella pira” in Il trovatore, even at the risk of infuriating the melomanes in the gallery and turning his opera house into a football stadium? After all, the high c is one note; in Guglielmo Tell tens of thousands of notes were changed from the opera Rossini actually wrote in order to underlay a metrically diverse text.

  That words somehow play a subordinate role in opera is inherent, too, in a 1995 article by Will Crutchfield in Opera News, where he lamented the loss of “one of my favorite examples of the power of melody in Italian opera.”3 He was referring to a moment in the Adagio of this same Manrico aria at the end of act 3 of Il trovatore, “Ah! sì, ben mio, coll’essere.” In his autograph manuscript, Verdi made a mistake in laying out the text in the varied repetition of the melody, anticipating the final line of a quatrain in place of what should have been the second line. In the original libretto, Manrico’s strophe is written as follows:

  Fra quegli estremi aneliti

  A te il pensier verrà!

  E solo in ciel precederti

  La morte a me parrà!

  With my last breaths

  My thoughts will come to you!

  And only to precede you into heaven

  Death will seem to me!

  But in Verdi’s autograph manuscript the text is underlaid as in example 11.1. Although every copy of the score and every edition (until the appearance in 1993 of the critical edition of Il trovatore) followed the autograph, the text of the second phrase, “la morte a me parrà, parrà,” should unquestionably be replaced (as in David Lawton’s edition) by the correct words “a te il pensier verrà, verrà!”4

  Crutchfield is too perceptive an observer of the operatic scene to suggest that singers should continue to perform the wrong words. His point is subtler: the force of Verdi’s art is independent of the niceties of the text/music relationship. It is “the melody, not the word, that speaks.” Indeed, he continues, after citing another gross mistake in the poetry of Ernani, finally corrected in Claudio Gallico’s critical edition (see below), “many generations of Italian tenors, baritones, conductors and teachers have handled these arias without apparently noticing that something unintelligible was being said.” Since the arias continue to speak to us despite this scrambling of their verbal messages, Crutchfield asks us to give “fresh regard for the absolute properties of voice, rhyme, melody, ‘idiot sound.’”5

  That Verdi’s art transcends such errors is certain. At what point, however, do they become utterly disorienting? If the order of verses 1– 4 –3–4 is acceptable because blessed by a hundred and fifty years of performance, why not 1–2–3–2 or 3–2–1–4 (“E solo in ciel precederti, a te il pensier verrà, verrà, fra quegli estremi aneliti, la morte a me parrà”)? Or why not a series of nonsense syllables? After all, is Verdi’s art better because the words are wrong? Would the composer’s melodic art not transcend the correct words as effectively as it does the wrong ones? In 2002 in Parma I sat through the presentation of a CD by a local organization interested in the history of singing, containing some twenty performances from early 78s of the Manrico aria. Whatever the positive qualities of these interpretations (and there were many), not one of the singers got the words right. That generations of tenors (tenors, so often tenors...) accepted this absurd text in Manrico’s cantabile and generations of opera fans listened raptly to their interpretations is utterly irrelevant. The words are simply wrong, punto e basta.

  EXAMPLE 11.1. GIUSEPPE VERDI, IL TROVATORE, ARIA MANRICO (N. 11), MM. 69–76, TEXTUAL ERROR IN THE CANTABILE.

  WORDS AND MUSIC

  Do words matter in nineteenth-century operas by Italian composers? As a theoretical question, the relative significance of text and music in opera or song has been debated by many generations of composers and critics. Prima la musica, e poi le parole (First the music, then the words), an eighteenth-century satire on operatic convention by Giambattista Casti, with music by Antonio Salieri, later became the inspiration for Richard Strauss’s last opera, Capriccio. Among Casti’s pointed barbs is a dialogue in which the composer misreads the libretto’s words, so that the phrase “and this costato [literally ‘rib cage’; figuratively ‘windbag’] will run me through” becomes “and this castrato will run me through.”6 When the poet objects, the dialogue continues:

  Composer: Castrato goes extremely well and I won’t change it.

  Poet: Are you making fun of me?

  Composer: What I wrote, I wrote.

  Poet: Have you gone crazy?

  Composer: I wrote castrato, it will remain castrato!

  Poet: And afterwards they will say, who was this poet who wrote such idiotic tripe.

  Composer: You won’t be the first or the last.

  The dialogue is so close to my dialogue with Piero Cappuccilli about Ernani (“your head, o traitor, there is no other choice” versus “your head or the traitor, there is no other choice”) as to be downright frightening: “traitor goes extremely well and I won’t change it.”7 More generally, composers have regularly been said to be indifferent to words (“Give me the telephone book, and I’ll set it to music” is a modern way of putting this), and Darius Milhaud proved the case in his witty Machines agricoles (1919), in which he set to music descriptions of farm machinery from agricultural catalogues.

  Without in any way undervaluing the terms of these debates or their significance, my own concern is a more practical one. Let us therefore rephrase our initial question. In the web of conflicting priorities that accompany every present-day operatic production, how much weight should we assign to the words? If we think about the question in those terms, it becomes apparent how many practical issues depend on our response. Should an opera be performed i
n its original language or in a language the local public understands? Should an opera house use supertitles? How should a singer balance the desire to produce beautiful sounds with comprehensible declamation of the text? What sorts of public spaces are most appropriate for the performance of opera? What balances should be maintained between the orchestra and the voices? Should singers’ voices be amplified artificially in order to allow poorly trained voices to be heard and to ensure that the words will be better understood, as in Broadway musicals since the mid-1960s?

  There are no unequivocal answers to any of these questions, for in each case the terms of the debate mean that choices need to be made. In each positive choice is embedded a negative one. If an opera house uses supertitles, members of the audience are able to follow the meaning of each phrase, but they lose visual and mental contact with the performers while their eyes and minds are focused above the stage (or, at the Metropolitan Opera, on the back of the seats in front of them). Performing in a nineteenth-century theater to 1,200 persons familiar with the language of the opera creates a different set of physical and vocal demands from performing in a twenty-first-century theater to an audience of 4,000 that for the most part does not understand the words. Yet when modern performers have to make choices—choices that may change from one theater to another, one work to another, or one performance to another—a knowledge of relevant history provides an intellectual and artistic context. Whether they embrace or reject this context, it is as much a part of the works themselves as the printed score.

  A large proportion of the theater-going public continues to believe that opera composers in Italy first prepared the music (taking Casti’s title literally, as a statement about compositional process), to which librettists subsequently added words. Wagner’s polemical reversal of the terms, with his insistence on the priority of the words in opera, from which the composer’s art would develop, became a club with which to pummel music from south of the Alps. But composers in Italy did not normally work this way. There is practically no evidence that Rossini ever composed music for his operas in the absence of a libretto, nor is there any reason to believe that Donizetti worked differently. When he wrote Il barbiere di Siviglia for the Teatro Argentina of Rome in January and February of 1816, with no time to spare, Rossini nonetheless waited until Cesare Sterbini handed him poetry, one piece at a time, before composing the music.8

  We have, however, a fine example of the limits of any such generalization: Rossini’s Neapolitan comic opera, La gazzetta. During the eight years in which he was involved with the Neapolitan theaters (from 1815 through 1822), Rossini wrote only one comic opera for Naples, La gazzetta, performed at the small, popular Teatro dei Fiorentini on 26 September 1816. As Rossini often did when facing a new audience, he recycled much earlier music in this second of his operas for Naples (as he had done in his first Neapolitan work, the serious Elisabetta, regina d’Inghilterra).9 After these two works, there is very little self-borrowing in Rossini’s eight subsequent Neapolitan operas.

  The borrowings in La gazzetta include a famous quintet at a masked ball from Il Turco in Italia. In Turco, Don Geronio, in Turkish garb, is searching for his wife, Fiorilla. But she and the Turkish woman (Zaida), together with the Turk himself (Selim) and Fiorilla’s admirer (Don Narciso), are also disguised in Turkish garb. The baffled Don Geronio doesn’t know which way to turn. In the final ensemble they all laugh at him, while he desperately calls out “Vo’ mia moglie... non capite?” (I want my wife... don’t you understand?). In La gazzetta, the husband Don Geronio becomes the father Don Pomponio, a character who sings throughout in Neapolitan dialect. The role was created by Carlo Casaccia, a legendary figure of the Neapolitan popular stage. Pomponio, seeking an appropriately rich and upper-class husband for his daughter, Lisetta, has announced a competition for her hand in the local newspaper (hence, La gazzetta), but Lisetta falls in love with Filippo, an innkeeper.10 And another couple forms during the course of the opera, Alberto and Doralice, in opposition to the wishes of Doralice’s father. Toward the end of the second act they all participate in the masked ball, dressed as Turks. Don Pomponio searches in vain for his daughter, while the young couples rush off to get married.

  For these two couples, the text of the quintet is essentially identical in the two operas. But a funny thing happened to Don Pomponio: in the quintet the librettist forgot to prepare a stanza for him to sing, even though there is a stanza for Don Geronio in Il Turco in Italia. In the libretto printed for the first performances of La gazzetta (and the opera was hardly ever revived during the nineteenth century), there is no text here for Don Geronio. In his autograph manuscript, Rossini includes words for the other four characters, but none at all for Don Pomponio. We have no idea what Carlo Casaccia sang at the first performances, although his powers as an improviser were legendary. Nineteenth-century manuscript copies of La gazzetta simply incorporate the textless part, paying no heed to the absurdity, while the only important nineteenth-century edition (prepared by Ricordi during the 1850s) imposes words from an earlier section of the quintet that have little to do with this situation.11

  Preparing the critical edition of La gazzetta for performance at the Rossini Opera Festival of Pesaro during the summer of 2001, Fabrizio Scipioni and I went back to Il Turco in Italia, borrowed Don Geronio’s stanza, but (with the help of Sergio Ragni, a Neapolitan born and bred) rendered it into Neapolitan for Don Pomponio. “Vo’ mia moglie” (I want my wife) became “Vuo’ mia figlia” (I want my daughter), while “Sarà questa, sarà quella” (she’s this one, she’s that one) became, with no change in meaning, “Sarà chessa, sarà chella.” For Rossini the problem of the placement of new words under previously written music is important, as we shall see, and colors our understanding of the relationship between words and music in his operas, but there is practically no evidence that the composer exercised his art for the theater in the absence of words.

  For Bellini the situation is different. The composer from Catania did indeed have a tendency to draft pages of melodic ideas, his “morning exercises,” as he was wont to call them, often in the context of thinking about particular operatic plots. Some fascinating pages exist from 1834 and 1835, as he began to consider the opera he was to compose for the Théâtre Italien during the carnival of 1834 –35 (I puritani), and then as he thought about a later project. The Puritani pages are filled with fragments of ideas, written down one immediately after another, sometimes a measure or two, sometimes an entire melodic period. Very little effort has been made to analyze what it is about these ideas that might have commended themselves to Bellini in the context of considering the plot of the new opera, but subconsciously, at least, there must have been something that brought them together in his mind.12 The technique may have worked for Bellini because he tended to distribute syllables in an extremely consistent way within a phrase and was less apt than Verdi, for example, to seek a particularized relationship between words and music.

  Subsequently, Bellini would develop a few of these ideas into full-fledged compositions, including the melody of Arturo’s cavatina (“A te, o cara”), Giorgio’s romanza (“Cinta di fiori”), and the theme of the duet for Elvira and Arturo in the last act (“Nel mirarti un solo istante”). In each case the sketched melodic fragment is significantly different from the operatic melody: the melodic cell for “A te, o cara” is notated in 4/4, rather than the characteristic 12/8 of I puritani; “Cinta di fiori” is represented by a two-measure fragment in B major (the operatic melody is in A); and the definitive version of “Nel mirarti un solo istante” is represented by just four measures in B major (the piece is in C major in I puritani), and is similar to the operatic melody only at its very beginning. Bellini knew that these melodies were destined for life in his operas, for when he drafted actual scenes incorporating these melodies, he crossed out the fragments, leaving untouched the themes that might still serve his needs. Turning these fragments into operatic melodies, however, was always tied to the verbal text of an opera
. Even when Bellini adopted or adapted one of his exercises, it would be transformed in the process of making it into an operatic melody.

  For Verdi we have vast amounts of evidence about the importance of words to his compositional process, evidence found in letters exchanged between the composer and his librettists and, increasingly, the elaborate sketches (particularly the continuity drafts) he prepared for his operas. But we also have at least one example of Verdi’s having prepared preliminary melodic ideas for an opera, a page of verbal indications and musical sketches for the first act of La traviata, which Verdi unquestionably drafted before having received any text whatsoever from his librettist, Francesco Maria Piave.13 On this page, the composer outlined the dramatic shape of the first act, still using the name of “Margherita” from Alexandre Dumas’ La Dame aux Camélias and identifying her beloved simply as “Tenore.” After specifying “Cena in casa di Margherita / Recc: Motivi d’orchestra. Brindisi” (Dinner in Margherita’s home / Recitative: motives in the orchestra. Brindisi), he wrote a long melody that he labeled “Brindisi del tenore”: it is a preliminary melodic draft of the brindisi “Libiamo ne’ lieti calici” without any words. The sketch is extraordinary: not only does Verdi invent the basic theme (written out in full, very similar to the final version), but he also specifies Margherita’s reprise, choral participation, the place where the soprano and the tenor alternate phrases, and even the conclusion with choral declamation beneath the melody.

  The page continues with a description of Margherita’s scene with the tenor, the reappearance of the guests, their departure, and Margherita alone on the stage. Then, specifying “Andante come segue,” Verdi drafted the entire melody of “Ahfors’è lui,” some forty measures, without any text at all. The melody differs in many respects from the final version, particularly when it turns from its minor mode opening to its major mode conclusion. What will become “A quell’amor” is here drafted in the relative major of the initial key, rather than the parallel major as in the definitive version. But the differences are matters of detail: the basic melodic material and shape are fully present, even though Verdi had received no words. Then, after some more sentences of dramaturgical explanation, he indicated “cabaletta brillante” (lively cabaletta), at which point he wrote down the first phrase of “Sempre libera,” noting that the tenor will be heard from offstage singing a phrase from the duettino.14