We have explicit evidence of Verdi’s concern for keys and vocal range as early as 1843, thanks to a letter dated 30 May 1843 pertaining to I Lombardi.21 It is from the composer to the tenor Antonio Poggi, husband of Erminia Frezzolini, the prima donna who created the role of Giselda when the opera had its premiere at La Scala on 1 February 1843. Poggi was to sing Oronte in a revival of I Lombardi in Senigallia during the summer of 1844 (indeed, Verdi ultimately wrote a new cabaletta for Poggi to perform in this revival). But Verdi had additional plans for it: “I intend to put in F [instead of G major] the closing stretta of the finale secondo, just at the point where it passes into the major mode. It seems to me that it will be less tiring and stronger, since all those high bs would become high as, and this note is extremely effective for Signora Erminia.”22 Verdi did indeed go ahead with this transposition, which is present in many early editions of the opera. His reasons for doing so are crucial to an understanding of his developing art. While he might not have been quite so accommodating later in his career, matters of this kind were never far from his mind.
Not that the transposition, whatever its merits, is without cost. This conclusion of act 2 is a powerful moment. The act had begun in Antioch, at the court of Sultan Accian, who describes the terrible acts of the Crusaders (“Forti, crudeli, esultano di stupri e di rapine” [Strong, cruel, they exult in rape and pillage]). The Muslims, however, have taken as hostage Giselda, daughter of the Milanese leader of the crusaders, Arvino, and niece to Arvino’s brother, Pagano, who is living nearby—incognito—as a hermit. Needless to say, Giselda and the Sultan’s son, Oronte, have fallen in love. In the second scene, Pagano (whose identity remains unknown to his brother, from whom he is estranged) agrees to help Arvino and the crusaders rescue Giselda. The last scene of the act returns to the palace. Ladies of the harem taunt “la bella straniera” (the beautiful foreigner). Left alone, Giselda, in the cantabile of her aria, describes her own, unholy love for Oronte, and prays for assistance to her dead mother. In the tempo di mezzo, the Sultan’s wife, Sofia (a secret convert to Christianity), describes the barbarous attack of the Crusaders, who are spreading death and destruction everywhere and have already killed her husband and her son (in fact, Oronte survives the massacre). When Arvino enters to greet his daughter, instead of welcoming him, she turns on him and (“almost as if struck by dementia”) starts her cabaletta.
This unusual cabaletta is in several parts. It begins in G minor with an Allegro moderato, in which Giselda insists that God cannot approve of acts that shed human blood, and she concludes the section, with a modulation to the relative major (B major), to the words “No, God does not wish it.” Then she adopts a prophetic stance, the tempo gets faster (“Più mosso”), and she envisions a future in which “I vinti sorgono, vendetta orrenda sta nelle tenebre d’età vicina!” (The defeated will rise again and horrible revenge awaits soon in the shadows!). Her music returns to the dominant of G minor, but after remaining fixed there for four measures, Verdi changes the mode to major and launches, “con slancio,” the vehement final section, where Giselda again insists that God rejects the “empio olocausto d’umana salma” (impious sacrifice of human corpses). In a short transition, she prophesies that Europe itself will be invaded by those who are now suffering. Again she returns to the dominant of G minor, in a passage essentially identical to the earlier four-measure phrase, to prepare the reprise of the G major theme. Its new text features her words “No, Dio nol vuole” (No, God does not wish it) and adds the image of Christ descending to spread the message of peace, not war.23
Verdi’s transposition of the final section to F major creates havoc in this careful sequence of keys: G minor, B major, dominant of G minor, G major. Had Verdi started the whole sequence down a tone, in F minor, there would have been some logic, but he was concerned only with getting maximal effect out of what he hoped would be a show-stopping moment to conclude the act.24 While he may have been concerned that a transposition at the beginning of the cabaletta would push the part too low for Frezzolini, the result is an unfortunate compromise. It also greatly weakens the parallelism between the two four-measure phrases that prepare the section in major, since Verdi significantly modified the first of these passages in order to move down a tone. Did he not care? Yes, enough to have introduced the effect while he composed the opera. But did he care enough to refuse to help Frezzolini make a stronger impression? No. In this case, however, Ricordi later came to Verdi’s rescue: all modern printed editions of I Lombardi print the music in its original key. Does that mean one should never make the transposition to assist a modern-day Frezzolini? Of course not. But it would certainly be better to sing the music as Verdi originally conceived it.
Nonetheless, for Verdi, the problems with this cabaletta from I Lombardi and with the end of the La traviata duet are rare. Most of the composer’s transpositions precede the complete realization of the opera’s score, and when he does introduce a transposition while revising a completed opera (as elsewhere in the 1854 Traviata) he is careful to make the shift seamless and to introduce necessary adjustments in the orchestral fabric. Additional transpositions should be a last resort, but—as we have seen—are not without precedent in Verdi’s practice.
There is a related example in Donizetti. The role ofNorina in Don Pasquale was originally written for Giulia Grisi, who sang the first performances at the Théâtre Italien of Paris beginning on 3 January 1843.25 Her cavatina begins with an Andante in G major, in which Norina reads aloud a story about a lovesick knight, “Quel guardo il cavaliere.”26 She then laughs and throws the book away as two orchestral chords shift us from the tonic of G major to the dominant of B major, the key of the cabaletta (“So anch’io la virtù magica”). Despite this unusual motion from G major to B major, there is no evidence in the autograph manuscript of the opera that Donizetti ever considered composing any part of the cavatina in a different key.27 It may be that he wanted to make a sharp division between the Norina who reads a sentimental love story and the sophisticated, even cynical Norina, who boasts of knowing all the fine arts of love.
When Eugenia Tadolini assumed the role in Vienna in the spring of 1843, however, and then again in the spring of 1844, Donizetti apparently transposed the cabaletta up to C major for her, and he seems to have liked it there. In a letter of 16 April 1844 to Ricordi’s employee Giacomo Pedroni in Milan, the composer cites a melody from this cavatina in the new key.28 But he adds, “The only thing to say is that if it seems too high, I’ll lower it, but in B major it loses a lot.” Apparently, then, Donizetti wrote the piece in the lower key, transposed it to the higher key for a different singer, Tadolini, and actually preferred the latter key. Ultimately, though, the piece circulated and has always been published in the original B major. While no contemporary source has been identified that shows exactly how Donizetti made the transposition, it is not difficult to make a hypothesis. If we change two notes in the vocal line and rewrite two orchestral chords, we can allow the music fluently to address C major for the cabaletta. On one level, of course, this resulting G major to C major progression is more “natural,” more common, than Donizetti’s original progression, but can we really argue that the composer didn’t know what he was doing when he prepared the score for Paris? Whether Donizetti was correct when he later asserted that the cabaletta sounds better in C major, of course, is something each Norina needs to decide for herself, paying attention not only to her own vocality but also to the orchestral sound and the harmonic context.
If done responsibly, then, transpositions can assist a singer who has difficulty at a particular moment within the score but is able to perform the rest of a role well. What is an irresponsible transposition? Well, one could certainly argue that it is irresponsible (but not uncommon) for a tenor to insist that “Di quella pira,” the cabaletta of Manrico’s aria in Il trovatore, be transposed from C major to B major because, unable to sing an interpolated high c at the end of the piece, he wishes to transform it into a high b.29 The or
iginal key of Manrico’s “Di quella pira,” C major, is a bright, forward-sounding tonality, in part because of its extensive use of unstopped notes in the strings, with their greater resonance. When lowered a tone, the brightness of the original is sacrificed, a sacrifice all the more to be avoided in this martial context. There can be compelling reasons, of course, for making such a transposition in terms of the way a passage lies across the registral breaks in a particular singer’s voice, and no abstract principle should be allowed to stand in the way of helping a singer give the best possible performance. But simply interpolating an unwritten high note at the very end of a piece may not be one of them.30
BELLINI (?) TRANSPOSES BELLINI: NORMA
While most of Verdi’s modifications in tonality took place as he composed, Bellini transposed his own music constantly, during rehearsals, after an opera’s premiere, or when adapting the music for different singers. He did not always ensure that the job was well done. When modifying the original Parisian score of I puritani for performance in Naples, with a mezzo-soprano Elvira instead of the original soprano, as well as a tenor Riccardo instead of the original baritone, Bellini often wrote only the vocal line, leaving it to a collaborator/copyist to transpose the orchestral parts. On many occasions his copyist did not produce a satisfactory result.31 In the revised version of the Scena d’Elvira, for example, where Bellini lowered the entire scene from the original E major/A major down a third to C major/F major, the composer himself essentially wrote only the vocal lines and the names of the instruments at the start (see the facsimile of the autograph manuscript).32
While Bellini also made a few emendations in the copyist’s transposition of the orchestral parts, his corrections were insufficient. Here are some musical faults in the violin parts from the orchestration in the transposition of the Scena d’Elvira:
1. When Giorgio and Riccardo enter after Elvira has first sung her melody, “O rendetemi la speme,” the first violins in Paris played a series of triplets (ff. 23v–24); in the Neapolitan transposition, the first note would everywhere have been too low to be played on violins. Instead of rearranging the figuration, the copyist substituted the first available higher pitch in the chord, producing a series of unattractive octaves or fifths with the bass on successive downbeats (pp. 52–53).
2. When Elvira finally sings her theme over a triplet accompaniment, to the text “Qui la voce sua soave,” Bellini changed the direction of the triplets and used a consistent pattern throughout this presentation of the theme. In the first measure, however, the second violins descended originally to g (f. 25); since the parallel note for Naples would have been e, too low for the instrument, the copyist changed the figuration altogether, compromising the lovely effect Bellini had planned (p. 55).
3. After Elvira recalls the joyous music from the opening of the opera, she turns melancholy (“Egli piange...”), and in the original version the strings played a poignant, dissonant phrase, Largo, a memorable moment in the scene (f. 31). As originally written, all the violins and the violas are in unison, the cellos and double basses an octave below (doubled by the bassoons). When transposed down, however, a prominent pitch in the middle of the phrase is too low for the violins. Instead of rethinking the orchestration, the copyist substituted a rest for what would have been an f (p. 67). Bellini really should not have allowed this abrupt hole in the sonority to remain.
4. In the tempo di mezzo, as Riccardo and Giorgio are commenting on Elvira’s madness, the orchestra plays a theme. The melody is played by the first violins (doubled an octave higher by the flute), with an accompanying repeated chord below the melody played in the other strings. When transposed down, the melody is fine, but the chordal notes in the second violins are too low. The copyist therefore rearranged the notes of the chord, but he did so in an awkward way, so that the notes of the accompaniment in the second violins now sit above the melody, where they muddy the entire sonority.
These are all serious infelicities. While those opera fans who think that their favorite art form begins and ends with the voice may not be disturbed, musicians observe Bellini’s nonchalance toward the orchestra here with sadness.
As is well known, the cantabile of “Casta Diva” was written by Bellini in G major, the key of his autograph manuscript, not the F major in which it is heard today and which is preserved in most printed editions. This includes the very first reduction for voice and piano published by Ricordi in Milan on the occasion of the first performances of Norma, at La Scala, beginning on 26 December 1831.33 While it is usually affirmed that the transposition was decided by Bellini himself, perhaps during rehearsals, perhaps immediately after the premiere, to suit the vocal needs of the first Priestess, Giuditta Pasta, no evidence has emerged to document Bellini’s participation.34 It is certainly relevant that in the autograph manuscript of Norma the first page of the orchestral introduction to the cavatina was at one point crossed out and folded in half. Still, nowhere on these pages did Bellini make any annotations pertaining to a possible transposition, nor has any other source in his hand been found with such a change.
More important, the fluent progressions Bellini used in order to arrive at this G major at the beginning of the cantabile and to continue, after its conclusion, with the tempo di mezzo leave no doubt about his original intention. The recitative preceding the cantabile concludes on an A-major chord, heard as a strong dominant, but instead of resolving to the expected D, the chord is repeated by the strings alone, pizzicato, and is reinterpreted as the lowered second degree (a so-called Neapolitan sixth) in a progression that moves through G minor to a luminous G major. It is not too much to imagine that at this very moment moonlight floods the stage. After a single measure arpeggiating the tonic harmony, a solo flute intones the melody (example 10.3).35 This is an elegant progression, unexpected, logical, and exhilarating. At the end of the cantabile, the situation is simpler: the concluding tonic note, g, is reinterpreted as part of an E-major harmony in the stage band (it is the third degree of the chord), and the tempo di mezzo begins with a reprise of the march from the choral movement preceding “Casta Diva.”
Even if it was Bellini himself who decided that the cantabile should be transposed down a tone to suit Pasta’s vocal requirements, we can hope that he was not responsible for the way in which the transposition is introduced, for the movement to the new key is both harmonically suspect and ugly. Not only do string arpeggios begin immediately (instead of awaiting the arrival on the tonic major, which justifies and exalts them), but they begin on the first beat of the first measure, so that the modulation starts there instead of on the third beat (as in Bellini’s original). This causes a useless repetition of the G-major chord, initially in first inversion (with the third degree of the chord, b, in the bass), then in second inversion (with the fifth degree, d, in the bass). And that in turn leads to unfortunate parallel chords in second inversion (G major, then F minor). The passage as printed in the first edition of the vocal score is shown in example 10.4.
EXAMPLE 10.3. VINCENZO BELLINI, NORMA, THE TRANSITION TO THE BEGINNING OF THE CANTABILE IN THE CORO E CAVATINA NORMA (N. 3) IN THE ORIGINAL KEY (G MAJOR).
EXAMPLE 10.4. VINCENZO BELLINI, NORMA, THE TRANSITION TO THE BEGINNING OF THE CANTABILE IN THE CORO E CAVATINA NORMA (N. 3) AS PRINTED IN THE FIRST EDITION OF THE VOCAL SCORE, WITH THE CANTABILE TRANSPOSED TO F MAJOR.
I am not arguing that a transposition of the cavatina to F major is in itself objectionable (indeed, while we’re at it, why would we want to stop a particular prima donna from singing it still lower, in E major,36 or even higher, in A major?), but only that the progression that introduces the F major in printed editions of Norma is awful. What makes the situation even more ridiculous is that a simple modification of the passage could lead fluently to the new key. One need only use Bellini’s original progression a tone lower, employing the same pizzicato strings that he originally employed. The bass line (a–b–c–c–f) is elegant, and the succession of chords becomes thoroughly convincing
(example 10.5). At the end of the cantabile, on the other hand, there is nothing to do. The less said about concluding a cantabile in F major and beginning a tempo di mezzo in E major, the better. Still, given the applause that invariably greets the conclusion of “Casta Diva,” it is not a juxtaposition anyone is likely to perceive.
EXAMPLE 10.5. VINCENZO BELLINI, NORMA, A POSSIBLE TRANSITION TO THE BEGINNING OF THE CANTABILE IN THE CORO E CAVATINA NORMA (N. 3), WITH THE CANTABILE TRANSPOSED TO F MAJOR.
This transposition creates few problems for the orchestral instruments, which can perform the cantabile a tone lower without difficulty. Only twice did Bellini write a low g for the violins (their lowest note), and these can be adjusted for the F-major transposition by simple changes in the arpeggiated chords. When a critical edition of Norma is finally prepared, I hope the version of “Casta Diva” in G major will assume pride of place. The fear some performers express (or experience internally) when faced with the higher key may have as much to do with their expectations as with the actual difficulty of singing the music in G major. Had the printed editions transmitted the piece in its original key, many present-day Normas would have few qualms about performing “Casta Diva” in that tonality. Even so, I am prepared to grant that from the point of view of singing technique and the placement of the voice between registral breaks, the F-major tonality may be more accessible to a broad range of singers. Thus, there continues to be an important role for the transposed version, especially if the orchestral introduction is modified to remain closer to Bellini’s original progression.
Divas and Scholars Page 47