By the mid-20th century, the Italian operas that had entered the standard repertoire had settled into standardized performing traditions. Many of the misprints in the scores were easily corrected by a teacher, coach, or conductor who was familiar with the errors. Books were published showing the cuts, alterations, and additions to the scores, and sometimes publishes included these in their editions. The fact that some of these were added to the scores decades after the death of the composer mattered little. Gossett documents a number of cases where attempts to return the score to something closer to the form in which it was first performed were often met with anger and derision. The “tradition” (such as it was) had now become canon. The prevailing attitude was “If it was good enough for Toscanini, it’s good enough for me.” In some cases, these additions alter the way a role can be cast. Two examples are found in Verdi’s most often performed operas. The interpolated high e-flat in “Sempre libera” in La traviata and the tenor high C in “Di quella pira” in Il trovatore. Verdi wrote neither note. Gossett comments”¦
Give me a tenor who can sing Manrico as Verdi conceived the part and chooses to add a singing high c, and I will join the loggione in applauding him. Failing that, let Manrico, in Rossini’s famous words to the same Tamberlick, leave the high c on the hat rack, to be picked up on his way out of the theater.”
Enter the critical edition. The first task of the editor is to locate reliable sources, in particular the autograph scores, wherever they may be: in libraries, public or private, collections of noble families, or even in bank vaults in Switzerland. Searches of libraries in the 1960s revealed scores thought to be long lost. In addition, part books belonging to famous singers (with alterations to the part written in by the composer), orchestral parts and other rich sources were hiding waiting for scholars to find them. Some of these finds included the source material for Rossini’s Il viaggio a Reims (much of it reused in Le comte Ory), the tragic ending of Tancredi, and original materials necessary to reconstruct Verdi’s Stiffelio (which the composer had reworked into Aroldo). Verdi was a particular problem as Verdi’s heirs still own many of the original sources of his work and are reluctant to share them with scholars.
In addition to restoring the opera to its original state, a critical edition more accurately represents the articulations, dynamics, and slurs of the original score. This allows Rossini to sound like Rossini, Bellini like Bellini, etc. Too often the cuts and alterations served to make Rossini sound more like Donizetti or Bellini, and Bellini or Donizetti to sound more like early Verdi, none of which is in any way an improvement. Fortunately, there is a growing acceptance of critical editions. Obviously, an edition is not a production and conductors, directors and singers must make choices among the available options and sometimes make cuts to conform to the realities of producing opera in the 21st century, not least of which the overtime that kicks in begins when a performance lasts more than three hours.
Many factors must be considered when deciding what of the many options to perform. One factor is the very different cultures in which these works were created. (19th century audiences clearly had more stamina than 21st century ones.) Next, we must choose between the various versions of the opera, including those created by the composer, those created with limited input from the composer and those created without any input at all. (Donizetti supervised most productions of his operas while he was alive so there are fewer contemporary nonauthorial versions. Those exist much later and only in his most popular works.) This creates its own problems as sometimes singers, conductors and audiences are well attached to “traditions” that have nothing to do with anything the composer wrote.
Gossett points out that we are under no obligation to recreate the opening night of any particular opera. Composers often made drastic changes to their own works to make them more appealing to audiences. This is especially true of works created for Italian theaters and later revised for Paris (in French translation). So which alterations to the score are acceptable and which are not? That is not a question easily answered.
Another aspect of 19th century theater must be considered at this point. For the most part, works were performed by a roster of singers engaged for the entire season. (Season for this purpose does not necessarily mean the entire year but might only mean a few months of continuous performances, for example the carnaval season between the day after Christmas and the start of Lent. One ensemble of singers sang all the important roles. This practice often required alterations if a role had been written for a singer whose voice sat a bit higher or lower or who had exceptional skills that the current occupant of the role lacked. Such alterations are less often required in modern theaters where each opera is cast separately. If the theater is producing a role requiring special skills, the singers with those skills are hired. Transpositions and punctature were quite common in the 18th and 19th centuries (and a few transpositions are still commonplace).
Another type of alteration made to accommodate the singers or the local taste was to cut portions of the score from performance. Cuts are made for a variety of reasons. The first of these is length. Modern audiences are not used to being in productions that last as long as what was standard in the 19th century, and as previously mentioned, performances over three hours incur overtime for the orchestra, crew, and sometimes also the chorus. One place where cuts are often made is in the recitative. This can often be done judiciously, and some cuts are so standard that the audience may not be aware that anything is missing. It also brings the bel canto operas more in line with Verdi’s operas where there is far less recitative.
Other cuts happen internally in numbers to make them shorter. Some of this is done for length and other times because of the difficulty in executing some passages. It is also quite common for repeated music to be excised from a production. In some cases, especially in Verdi, the entire cabeletta section of a two-part aria (like the cabaletta following Deh miei bollenti spiriti in Verdi’s La Traviata, is omitted. (In that case cut because many tenors struggled with it.)
Sometimes entire numbers are cut. This is in some ways part of the shift in dramaturgy from the 18th century into the 19th. In 18th century operas every character, including the servants, must have at least one aria to sing. By the mid-19th century such parts are cut to a bare minimum. It is therefore understandable why 19th century producers would find such moments in the opera unnecessary. This made earlier works sound a bit more contemporary to the audiences of the time.
I have greatly condensed the material covered in Divas and Scholars, but even scratching the surface means that there will be two more blog posts to cover the main points. This book really is a must read for anyone involved in the production of opera in any capacity.