Bel Canto Bookshelf: The Blog

Divas and Scholars Part III: Performance Practice for Singers

One of the big questions around ornamentation is “when did it stop?” At what point did composer begin expecting that singers would perform exactly what was in the score and no more or no less? While it is not considered correct practice now to add ornaments or appoggiaturas to the operas of Verdi (after all, he writes in so many himself) that does not explain the curious notation found in the score for Rigoletto. In the scene preceding the trio in the last act, Verdi wrote in the autograph score (which is reprinted in the Ricordi piano-vocal score on page 218) “This recitative must be declaimed with the usual appoggiaturas.” (Quest Recitativo dovrà essere detto senza le solite appoggiature.) This is a curious (and unnecessary) notation if the practice of adding additional appoggiaturas had fallen out of practice by that time. 

As for variations and cadenzas, when performing Rossini the singer should be aware that in many cases we have ornamentation and cadenzas in Rossini’s own hand written out for specific singers. It is not necessary for 21st singers to sing those exactly, but they do give us examples of the appropriate performance practice for singing the music of Rossini and his contemporaries. Those practices can and should be adapted to suit the particular singer’s unique voice and skills. Other useful examples come from the notebooks of singers of Italian and French opera and printed editions claiming to represent the ornamentation performed by a famous singer in that aria. We also have the examples printed in various treatises (like Garcia’s) which provide examples of the kinds of variations, ornamentation and cadenzas considered exemplary in that era. (Note: some of the published versions can be found at Many of those are quite interesting, for example, a published version of Nicola Vaccai’s “Api Erranti” with ornaments by the great castrato Velluti. Since there is text for two additional verses, one could sing a mostly plain first verse, a second verse with the singer’s own ornaments, and then the third verse with Velluti’s rather elaborate variations.)

By no means should a singer assume that one has to sing someone else’s ornaments, variations or cadenzas. As Laure Cinti-Damoreau wrote:

I do not offer them [examples of ornaments and cadenzas] to you to be performed at any cost, despite your physical capabilities and your character. I propose these models of variations, rather, so that later your taste will lead you, within your individual means, to invent others that suit you properly.” (p. 301)

(Note: Her treatise can be found here. It has ample examples of exercises and includes multiple variations on the same melody illustrating the practice of ornamentation.)

Gossett provides the following caveat regarding ornamentation and appoggiaturas in recitative in modern performances.

Nineteenth-century practice is a guide for modern performers, not a recipe, and performers must never lose sight of who they are as musicians or of the audiences for whom they are performing. Appoggiaturas, after all, lend weight to phrase endings, applying stress and adding rhetorical emphasis to the poetic structure. Modern performance style, on the other hand, in the spoken drama as on the operatic stage, tends to flow more quickly, avoiding what today are perceived as excessive rhetorical devices. The emphatic Shakespearean declamation of John Gielgud, for example, reflected a powerfully different style from that of the more conversational Derek Jacobi.” (p. 304)

As for the matter of cadenzas, they were longer than those to which we are now most accustomed. A good many of the ones that were considered “standard” in the 20th century are anachronistic. (The “standard” cadenza in “Una furtive lagrima” is a good example of that. It was borrowed from Verdi and inserted into a Donizetti opera. Gossett details some examples of cadenzas written by the composer or from the same time period which are more appropriate. Of course, this all depends on the individual singer. The process of constructing an original one is not that hard but it would require immersion into the style of the opera.)

Rossini provides us with the most examples of ornamentation. First, the ornaments written out in the score, and also those written for performances of his and other composers’ work, including a couple of examples of Rossini writing out ornamentation for music by Bellini. In Bellini’s autograph scores we can see alternations in which Bellini reworked music for particular performers and also alternations made for subsequent singers in the same roles. There are also changes in orchestra doublings or instrumentation (sometimes indicating that in the theater he found the orchestration too heavy. We have fewer examples of this in Verdi’s scores but there are similar alterations for Jenny Lind who created the role of Amalia in I masdanieri. 

As for alterations to the score like transpositions, punctature, and variations, they can mostly be explained by looking at how opera was performed in the 19th century. Opera in 19th century Italy was cast from an ensemble of singers hired for a season. Because of this sometimes a singer would find a role that they were required to sing sat uncomfortably high or how for them or at least had numbers out of their comfortable range. There were two ways to address this issue: transposition and punctature. Transposition is obvious enough and was quite common, and often happened with arias. But in ensembles where one singer’s music is too high or low but the music for the other singers is singable, they would alter the vocal line by changing the notes to other pitches that fit the underlying harmony. That is punctatura. We know that composers were aware of this and in many cases approved. We know, for example that when Verdi’s Ernani was set to premiere in Vienna in 1844 that Verdi made it clear that he would allow no cuts to the score, but in a separate letter to Donizetti, who was supervising musical matters at the theater at the time, that he trusted the older composer to make any necessary punctature. Gossett makes a strong case (with examples) that it is preferable to make a few minor alterations to the score rather than have a singer struggle with a note beyond their range (either too high or too low), an ornament they can’t manage or sputter to the end of a long phrase they cannot sing in one breath. There are limits. (And here as in many cases Gossett is not shy about calling out famous names!) He especially calls out Estelle Liebling’s edition of Una voce poco fa which one hopes has finally fallen out of fashion. 

Transpositions are a controversial matter, even though a good number of them are standard practice. (As in the previous mentioned version of “Una voce poco fa” in which the aria is transposed up to make it more suitable for a soprano voice.) This issue includes the problem of the constant rising of orchestral pitch. In Verdi’s day a=432 (even though in 19th century Rome it was a=450). That is high enough to make the highest pitches in a singer’s range unstable if a particular theater is using higher than usual pitch. 

One argument in favor of allowing transpositions is that in Verdi’s sketches we sometimes see that he had begun work on an aria in one key and later changed it to suit the singer in the first performance. That means that the keys were chosen for the singers, not singers chosen to be able to sing the notes of the arias in the “original” key. One of the arguments against transpositions is the belief on the part of some musicologists that specific keys were chosen for deeper meanings. (For some composers that is true. Richard Strauss, for example. But other than the association of C major with Sir John Falstaff in Verdi’s final opera, I know of know evidence of such associations for Italian composers.) Gossett points out that there are no key associations for which someone could not find some relationship, whether it was intended by the composer or not. Again, Gossett is able to provide multiple examples of arias and even ensembles that were originally composed in one key and then modulated by the composer before it reached its first performance. Does that mean that all transpositions are okay? As an example of an ill-advised transposition, Gossett again brings up the issue of “Di quella pira” in which it is not uncommon for the cabaletta to be transposed down a half or even a whole step just so the tenor can interpolate a high C (now B or B-flat) that the composer never wrote. We also see examples of Bellini transposing his own music not only for the first performances but for later performances with different casts. If Bellini did not mind transposing his own music, should we not be allowed to change the key to make a role more singable? “Casta diva” is given as an example as it was originally composed in a different key than we now usually here it in performance. Donizetti did the same for “Regnava nel silenzio” from Lucia di Lammermoor for the soprano Fanny Tacchinardi Persiani. (She abandoned the aria altogether the next time she sang the role substituting “Perché non ho del vento” which Donizetti had written for her in Rosmonda d’Inghilterra.) Lucia’s mad scene was also transposed. Donizetti composed it in F major but it was transposed to E-flat major where it remains in most printed editions. 

A number of transpositions in bel canto opera occur in those roles written for Rubini (or those Rubini later sang but had been written for other tenors). Even in his own time Rubini’s voice was exceptionally high which means that a great number of transpositions exist in operas like Il pirata when tenors found the music composed for Rubini much too high for their voices. (And this was before the switch to tenors singing everything in chest voice!) Some of these changes were made by Bellini himself, some by others, and some by publishers (who must have realized that few tenors could manage the F above high C!). 

What then does that mean for modern theaters wishing to put on operas like I puritani, Il pirata or La straniera? Do we transpose the music to suit the tenor? What does that do to the tessitura of other parts in the ensembles? What changes might that require in the orchestration? There are no easy solutions. Part of what makes this book so interesting is that Gossett’s view is that we need to be informed about all aspects of how these operas were composed and performed not only at the premiere but in the many productions that followed during the composer’s own lifetime. Only then can we make practical choices that will ensure successful revivals of this music. Strict adherence to a critical edition is no more a solution than blindly disregarding the original score and making hack-work of the music.