3. Verdi wrote two strophes because he was obeying a tired convention, and we need to save him from himself. This argument returns us to La traviata as verismo, a false path. Verdi was indeed obeying a convention, but two-strophe arias, cantabiles, or romanze are relatively rare in his operas, hardly a dying breed like the full cabaletta. Who are some of the characters who sing them? Well, there is François I, the French king. While he may have become the Duke of Mantua in Rigoletto (when Verdi was forced to change the setting), he never lost his Gallic charm: “Questa o quella” and “La donna è mobile” are both strophic songs, popular and elegant in style. Another figure characterized by romanze is the page, Oscar, in the very French court of Gustavo III of Sweden (also known as Riccardo, Governor of Boston) in Un ballo in maschera: “Volta la terrea” and “Saper vorreste” both grow out of the tradition of couplets in opera (particularly opéra-comique) and romances in the drawing room that dominated musical life in midcentury Paris. Verdi’s choices in La traviata, in short, were not casual.
In nineteenth-century Italian opera, in fact, the romanza in particular and strophic forms in general often had the character of French imports.86 By employing this form for both of Violetta’s solo cantabile moments (not to mention Germont’s “Di Provenza il mar, il suol”), Verdi has added important elements to her characterization: a simple vocal style that escapes sentimentality through the composer’s art, and a Parisian identity integral to the opera. The story of Marguerite Gauthier was a French story. Even when he was compelled for the premiere to place the story back at the end of the previous century (rather than treat it as a contemporary tale of mid-nineteenth-century vintage), Verdi never considered moving it to another location.
If we believe in Verdi’s La traviata and not some image of it seen through a late nineteenth-century lens, Violetta needs to sing both strophes of both her romanze. Once they are regularly heard performed well, arguments for their omission will dissolve. That does not rule out other small cuts that can be made in La traviata, as well as in Rigoletto and Il trovatore. Many cadential phrases in lengthy duets, for example, are repeated in Verdi’s scores. I rather like hearing them twice, but little is lost by omitting some of these repetitions. Less acceptable is the habit of snipping away parts of the cadenzas Verdi wrote in those same duets. There is nothing superfluous in the efflorescence of the linked voices of Gilda and the Duke in the cadenza at the end of the cantabile of their duet: What authorizes Serafin to claim that Verdi surrendered to base instincts when he wrote that glorious cadenza a due?87 Unless there are strong, practical reasons for trimming his scores (such as the inability of singers to keep a cadenza in tune), my general instinct is to trust Verdi.
VERDI’S MACBETH: THE OLD AND THE NEW
But Verdi himself occasionally fell into the very trap I have been describing, and it is fitting to conclude these two chapters on versions and cuts by examining the two versions he himself prepared of an opera he particularly loved, Macbeth. The operatic world of nineteenth-century Italy did not welcome, as we do, the coexistence of historically disparate styles in a repertory. As some of his mature operas fell out of fashion, Verdi sought to rescue them by bringing them up to date and removing the most blatant traces of an earlier style: Stiffelio of 1850 became Aroldo in 1857; Macbeth of 1847 was revisited in 1865; Simon Boccanegra of 1859 was heavily recast in 1881, with Boito’s assistance; even the relatively late La forza del destino and Don Carlos, of 1862 and 1867, respectively, were the object of revisions in 1869 and 1884. The motivations behind each of these revisions were slightly diverse: the heavily censored Stiffelio could not be performed as Verdi originally planned it; Macbeth was revised for Parisian performance almost twenty years after its Italian premiere; Simon Boccanegra, despite its many strengths, had ceased to circulate, and Verdi used its revision to test a possible collaboration with Boito; the altogether bleak conclusion of the St. Petersburg La forza del destino never fully satisfied the composer; and the French Don Carlos required more than a translated text to circulate easily in Italian theaters. These revisions provide interesting examples of how the composer himself, in revising an opera, would attempt to nudge an earlier work along the historical and stylistic spectrum to rescue it from oblivion. But rarely are the revisions altogether successful, despite critical efforts—even by such superb Verdians as Julian Budden—to show that they inhabit the best of all possible worlds. However much Verdi worked over a score, the seams between the old and the new always showed.
The two versions of Verdi’s Macbeth have already been discussed briefly in the previous chapter. When the first version, performed at the Teatro della Pergola of Florence in March 1847, was revised for performance at the Théâtre Lyrique of Paris in April 1865, Verdi carried on an elaborate correspondence about the project with his Italian and French publishers, representatives of the theater, and various friends and colleagues. Recall how the composer had described the first version of Macbeth in 1847 and 1848: dedicating the vocal score to his father-in-law, Antonio Barezzi, he called Macbeth the opera “which I love in preference to my other operas”; warning Vincenzo Flauto, impresario at the Teatro San Carlo in Naples, of the difficulties he would face in staging the opera, he wrote that “I hold this opera in greater regard than my others.”88 In short, the 1847 Macbeth was an opera of which Verdi was intensely (and justifiably) proud.
When the idea of a performance in French at the Théâtre Lyrique was raised in 1864 by Verdi’s French publisher, Léon Escudier, first in a letter, then in a visit to the composer in Genoa during the month of June, the composer was asked simply to prepare “three airs de ballet,” to adjust Macbeth to French taste.89 Looking over his opera from the perspective of the mid-1860s, however, he “was struck by things that I would not have wished to find,” and he identified several numbers “that are either weak, or lacking in character, which is worse still”:
1) An aria for Lady Macbeth in act 2
2) Various passages to rewrite in the hallucination scene of act 3 [sic; act 2 is meant]
3) Rewrite completely Macbeth’s aria in act 3
4) Retouch the opening scenes of act 4
5) Prepare from scratch the last finale, removing the death of Macbeth onstage.
What has he identified (leaving aside the small modifications in the finale of the second act)? Three solo numbers that embody pre-1850s conventions and a chorus in the Risorgimental tradition of “Va, pensiero.”
While none of these pieces has ever been considered among the greatest achievements in Macbeth (such as the first-act duet for Macbeth and Lady, most of the second-act finale, and the sleepwalking scene), Verdi did not have a negative view of them as he prepared and rehearsed the opera. Let us consider them in the light of Verdi’s own comments in his letters from 1846 and 1847.
1. An aria for Lady Macbeth in act 2, “Trionfai!” Hoping to counteract the prophecy of the witches that Banco’s heirs will rule, Macbeth and Lady lay plans at the beginning of the second act for murdering Banco and his son. After providing Piave revised text for this recitative (he had not liked Piave’s verses), Verdi concluded: “When the Lady remains alone, two quatrains are needed; but the old ones won’t do, and the first one in particular must be changed; so instead of an adagio I’ll write an allegro, which will be even better.” He informed the original Lady, Marianna Barbieri-Nini, that she would receive “another aria consisting, however, of a recitative and a single brilliant cabaletta.” Later he decided to write this cabaletta “for you in Florence so that it will suit your voice perfectly and be sure to make an effect.”90 And so it did. According to one reviewer, “Lady Macbeth [...] sings [...] an allegro with all the vigor and energy of which Barbieri’s lungs are capable. Endless calls and applause.” Another described it, however, as “an aria in the worst possible style and taste, which, however, was well sung.”91
The florid and brilliant aria, neatly described by Julian Budden as “a brash cousin of Elvira’s ‘Tutto sprezzo che d’Ernani,’ wit
h more than a touch of Abigaille [in Nabucco],”92 is perfectly consistent with the cabaletta of Lady’s first-act cavatina. Sung with conviction, it makes a strong musical effect, while dramaturgically it keeps Lady focused and forceful, rather than anticipating her psychological breakdown (as in the 1865 “La luce langue”). Would Verdi have written “Trionfai!” in 1864–65? Of course not. But “La luce langue,” however accomplished, seems to have wandered into Macbeth from another world. Not even Verdi’s perfumed reharmonizations of other sections of the score succeed in washing Macbeth clean of its origins.
2. Aria Macbeth, act 3. The return of Macbeth to consult the witches anew is the subject of the entire third act, which in its original form is an expansive gran scena for the protagonist: an introductory chorus of witches (“Tre volte miagola”), an elaborate recitative for the appearance of the apparitions, a cantabile as the heirs of Banco appear in ghostly procession (“Fuggi, regal fantasima”), a tempo di mezzo consisting of the descent of the aerial spirits and chorus (“Ondine, e silfidi”), and a concluding cabaletta (“Vada in fiamme”). In describing the piece to his original Macbeth, Felice Varesi, Verdi wrote: “There is a cantabile (sui generis) with which you have to make a big effect.” As for the cabaletta, “it does not have the usual form, because, after all that has preceded it, a cabaletta in the usual mold and with the usual ritornellos would seem trivial. I’d made another one that I liked when I tried it out by itself, but when I joined it to all that went before, I found it intolerable. This one suits me fine, and I hope it will suit you too.” But the third act did not excite much popular enthusiasm, and one reviewer suggested that “it might not be a bad thing to suppress all of the final cabaletta, ending with the chorus and ballet of the sylphs.”93
The original piece had a strong internal logic, which the revision ignored. Verdi added a lengthy ballet after the initial chorus and made subtle changes in the scene of the apparitions and in the cantabile, though leaving intact the basic ideas. The cabaletta, on the other hand, he simply suppressed. Yet the appearance of Lady Macbeth in the witches’ den is dramaturgically suspect, and the duet in which she and her husband goad each other toward “vengeance” is anything but inspired. While its style is unquestionably more mature than the forceful A minor/major cabaletta of 1847, in which Macbeth in the best baritonal fashion vowed revenge on Macduff, the revision does not improve the opera as a whole.
3. Chorus, act 4, “Patria tradita.” Verdi knew exactly how he wanted to begin the last act in 1847, having given the following instructions to Piave: “I’d like the scene to open with a grandiose, moving chorus, which would describe Scotland’s wretched state under Macbeth’s rule.” And he added: “Let this be a grandiose chorus. Beautiful and moving poetry, in any meter you want except decasillabi.” Later he was even more explicit: “I’ve tried to set the first chorus but haven’t been able to make it grandiose because, among other things, the meter is too short. So do me the favor of making four strofette of ottonari.” Verdi then explained why he wanted this particular verse form: “I’d like to do a chorus as important as the one in Nabucco, but I wouldn’t want it to have the same rhythm, and that’s why I ask you for ottonari” (“Va, pensiero, sull’ali dorate” from Nabucco, “O Signore del tetto natio” from I Lombardi, and “Si ridesti il Leon di Castiglia” from Ernani all employ the poetic meter of decasillabi). And the composer continued, “Don’t let this moment slip by, the only one in the entire opera that’s affecting. So do it with passion.”94
Verdi liked Piave’s verses enough to use them without change both in 1847 and in 1865, when he completely rewrote the music of the chorus. The 1847 chorus is within a strong Risorgimento tradition.95 For the first of Piave’s four strofette Verdi wrote an affecting melody in the minor, sung in unison by the entire mixed chorus. In the second, in the relative major, the chorus breaks into harmony (as when the Hebrew slaves in Nabucco invoke their “harps of gold”). The third brings back the music of the first, still in unison, while the fourth (related to but not identical to the second) moves to the tonic major. A twinge of minor in the final cadences helps moderate what would otherwise be a questionable triumphal conclusion. The harmony is simple, the melody tuneful, though not so tuneful as “Va, pensiero”: Verdi may have sought to avoid comparisons with his previous hit, yet only five years separated Macbeth from Nabucco. The 1865 chorus is very different, with its harmonic complexity, its orchestral semitone figures (both up and down), its massed harmonies and gentle counterpoints in the chorus, its avoidance of simple repetition. Then, up steps Macduff to sing a standard 1847 aria, and once again the contrast between 1865 and 1847 comes as a shock. To ask which opening chorus is better misses the point: Better for what? Better in which opera? Better in which context?
4. Morte di Macbeth. When Verdi sent a sketch of the text for this final scene to Piave, he instructed him, “Make two strofette, try to give them some touching quality, but don’t forget Macbeth’s character.” To Varesi he described the concluding moments as “a very brief death scene—but it won’t be one of those usual death scenes, oversweet, etc.” When he sent the music, he added:
You’ll be able to make much of the death scene if, together with your singing, your acting is well thought out. You will understand very well that Macbeth mustn’t die like Edgardo, Gennaro, etc. [Verdi is referring to the final cabaletta in Lucia di Lammermoor, with its invocation of Lucia as the “bell’alma innamorata,” and to the revised finale of Donizetti’s Lucrezia Borgia, in which Gennaro dies in the arms of his mother], therefore it has to be treated in a new way. It should be affecting, yes; but more than affecting, it should be terrible. All of it sottovoce, except for the last two verses, which, rather, you’ll also accompany with acting, bursting out with full force on the words “Vile... crown... and only for you!...”
After the premiere, Varesi (hardly an impartial observer) referred to “my death scene” as among the most inspired parts of the opera.96
As in the case of the chorus opening act 4, the death scene from Macbeth looks back to Nabucco, the similar concluding passage for Abigaille, also a character whose evil deeds are finally punished. These brief, intense death scenes, with a wounded or poisoned character staggering around stage, singing a tortured solo, then collapsing, were not particularly favored by Verdi’s contemporaries. Operatic deaths tended to come in duets, trios, or ensemble scenes (think of Rigoletto, Il trovatore, or La traviata), and early in the history of Nabucco the Abigaille scene was already cut in many performances. Still, Verdi knew that he wanted this effect in Macbeth, and even instructed Varesi about how to stage the ending: “You’re on the ground, of course, but for this last line [‘Vile... crown...’] you’ll stand almost straight up and will make as great an impression as possible.” When Verdi turned his back on this scene in 1865, he was criticizing not so much his specific music as a whole convention. As he told Escudier in December 1864, “I too am of the view that Macbeth’s death should be changed, but the only thing that can be done is a Victory Hymn.”97 And so it was, with a marchlike phrase in a minor key, “Macbeth, Macbeth, ov’è” for the soldiers and bards, followed by the soaring major-mode phrase of thanksgiving with which the opera ends. It is a lovely chorus, and its willingness to pack all its emotion into a phrase of four measures is something that we find only in the mature Verdi. Is it a strong ending for the opera? Not so strong that many modern productions have not taken to inserting back into this final scene the “Morte di Macbeth,” thereby trying to have the best of both worlds, and perhaps only succeeding in having neither.
As his work on the opera proceeded, Verdi made other significant changes, but the examples already cited provide sufficient insight. When Verdi and Boito revised Simon Boccanegra in the early 1880s, they faced again and again the kind of problems we have traced in Macbeth, and the collaborators invoked the metaphor of a wobbly table: when you fix one leg, another gets out of balance, and soon you’re eating on the floor.98 A good performance, of course, s
mooths out the frictions, makes the parts seem to belong together, compensates for the historically generated dissociations between elements of the opera. Even before Verdi made his revision, furthermore, performers made their own decisions about these places. To take one example, Pauline Viardot, who was to perform Macbeth in Dublin in 1859, informed the conductor, Luigi Arditi, of her transpositions and modifications, adding, “The cabaletta ‘Trionfai’ is not sung.”99 Viardot used a pair of scissors; Verdi substituted another piece. Both faced the same problem.
It would be possible to assume a stance to the problem of cutting that is absolute: the composer wrote it, the performer should follow his instructions. That is what Stravinsky told Ernest Ansermet in two letters written from Paris in 1937. In the first (14 October), he proclaimed:
There is absolutely no reason to make cuts in Jeu de cartes in concert performances, any more than, for example, in Apollo. [...]If you propose this strange idea of asking me to make cuts, the reasoning must be that the succession of movements in Jeu de cartes seems a little boring to you personally I cannot do anything about that. [...] I cannot let you make cuts in Jeu de cartes! I think it is better not to play it at all than to do so reluctantly.
Ansermet on 15 October persisted, although modifying his request: “I ask only one thing: permit me to cut from the second measure of 45 to the second measure of 58.” The furious composer explained to him why this cut was unacceptable, and concluded, “I repeat: either you play Jeu de cartes as it is or you do not play it at all. You do not seem to have understood that my letter of October 14 was categorical on this point.” 100 One feels a certain sympathy for Stravinsky in all of this, but his position has never been possible in the world of Italian opera. Perhaps it has never been possible in any performance art.
Divas and Scholars Page 40