Monday, 12 April 2021

Towards Perfection: the idea of an ideal version of an opera has not always played out in history, with composers being surprisingly willing to rewrite works to suit circumstances

Stuart Laing, Kristy Swift & ensemble - Beethoven's Leonore at the Buxton Festival, 2016 (photo Robert Workman)
Stuart Laing, Kristy Swift & ensemble - Beethoven's 1804 Leonore at the Buxton Festival, 2016 (photo Robert Workman)

Whilst it might seem a piece of 19th century Romanticism, the ideal of a composer straining to create the perfect version of an opera is one which still informs the way we think of many of the operas in the historical canon. But history shows that it was rarely thus, and operas were rarely final and even the great composers often showed a surprising willingness to tinker with works and adjust them.

On 23 May 1814, Beethoven premiered the opera we have come to know as Fidelio. It wasn't the work's first outing, originally the work premiered in 1804 with a revised version appearing in 1805. What was performed in 1814 was a further radical revision. In making the changes, Beethoven wasn't responding to changes of cast (all three versions featured soprano Anna Milder-Hauptman in the leading role of Leonore), nor radical changes of performance location, all three took place in commercial theatres in Vienna. Instead, Beethoven was working towards perfecting Fidelio as a work of art, though as his own musical personality had developed significantly in the years from 1805 to 1814, this meant that the final version of Fidelio had significantly different aims to the work which premiered in 1804 and which we now know as Leonore.

It is with Fidelio that the idea of opera as a perfect work of art would seem to come into being, a myth that would be continued by Richard Wagner. Not only do all of Wagner's mature operas exist in single, final versions but what constitutes mature Wagner was (and is) rigorously curated by the Bayreuth Festival so that his earliest three works are not included in the canon. In an ideal world Wagner would have kept The Ring and Parsifal as being performed only at Bayreuth but financial pressures forced him to sell the Ring copyrights. This is the creation of opera as a perfect work of art, controlled by the composer (and librettist); it assumes that an opera only exists in a single version and that is the one we should focus on.

Of course, it wasn't always thus and in fact is hardly ever thus, though the creation of traditional versions of some operas has led us to prize some music over other in a way which would have puzzled the works' creators.



In the 18th century, composers always reacted to circumstance. Work on the music was not commenced properly until the composer knew who the singers were, and if a work was revived then the composer would make revisions to suit the singers. Sometimes visiting singers would bring 'their' arias with them, and these had to be included. With a composer/impresario like Handel there is a tendency to divide Handel the composer from Handel the impresario and prize the original versions over later revisions. This can be because Handel's later changes could play havoc with any original conceptions and dramaturgy with the sole end of making the piece work for the current singers and the current audience. The result is that we tend to ignore Handel's revisions, even when they are interesting.

Handel: Giulio Cesare - Catherine Hopper (Cornelia), Heather Lowe (Sesto) - Opera North 2019 (Photo Alastair Muir)
Handel: Giulio Cesare - Catherine Hopper (Cornelia), Heather Lowe (Sesto) - Opera North 2019 (Photo Alastair Muir)

Handel's Giulio Cesare premiered in 1724, with the role of Sesto played by a soprano en travestie (Margherita Durastanti), but when Handel revived the work a year later he needed to find a role for the major Italian tenor Francesco Borosini (who would create the role of Bajazet in Tamerlano for Handel that same year) and so Handel made significant alterations to the role of Sesto, retaining one aria unchanged, altering another and providing three new tenor arias. Borosini would sing in 10 performances of the opera, and Handel's revivals in 1730 and 1732 would both use a tenor for Sesto so that in terms of primacy in the historical record the tenor version wins. But this is rarely if ever explored nowadays and we have created a canonic version of Giulio Cesare which has Sesto played by a mezzo-soprano. (It has to be admitted however that the transfer of the role from youth (sung by a woman) to mature man (sung by a tenor) turns the character's lack of effective action into something of a dramaturgical weakness.)

Handel's first version of Radamisto - Joanna Marie Skillett (Tigrane), Bianca Andrew (Radamisto) - Guildhall School of Music & Drama, 2017 (Photo Clive Barda)
Handel's first version of Radamisto - Joanna Marie Skillett (Tigrane), Bianca Andrew (Radamisto) - Guildhall School of Music & Drama, 2017 (Photo Clive Barda)

With one of Handel's earlier operas, we have the unusual situation of Handel creating almost a separate version of an opera which has a strong claim to rank above or with the original. This is Radamisto which was premiered in 1720 by the relatively new Royal Academy, in fact Radamisto was Handel's first opera for them. So Radamisto premiered in April 1720, but not all the company's stars had arrived and in December 1720 he revived the opera with major revisions to accomodate the castratos Senesino and Berselli. Senesino, the male star of the company, took the title role and this engendered a large-scale updating with three major characters changing voice type but Handel did not just transpose he re-wrote, creating a significantly different version. The changes are far more than what was needed, and represent a sense of thinking differently about the same material.

When Handel moved to oratorio, his control over his singers increased; he tended to work with the same group, and often they were performers whom he had trained or coached himself. There were still changes, but there were also fixed limits. A frequently performed work like Messiah tended to be adjusted for each outing, but when in 1750 Handel had an alto castrato available, Gaetano Guadagni, Handel did not let Guadagni sing the contralto arias, 'He was despised' was always sung by a female contralto and Handel crafted new solos for Guadagni. Thus, within all the changes we can detect the hints of a fixed work of art.

When Gluck and his librettist Calzabigi premiered Orfeo ed Euridice (1762), Alceste (67) and Paride ed Elena (1770) in Vienna they aimed to reform the style of Baroque opera practised by Handel and his contemporaries, simplifying, removing the reliance on virtuoso star singers and moving more drama from the recitative to the ensemble music. Yet Gluck was not completely wedded to this, he would write unReformed and Reformed operas simultaneously and, when in 1774 Gluck moved to Paris, he reconfigured two of his Reform operas for Paris, responding to new singers, new voice types (no castratos and high-voiced tenors) and new dramatic requirements. So we do not have a perfect version of the operas, and in fact it wasn't until the 19th century when Berlioz created the 'perfect' version of Gluck's Orpheus, synthesising the French and Italian versions. [See my article]

When the 14-year-old Mozart was commissioned to write a new opera, Mitridate, Re di Ponto, for the Teatro Regio Ducale, Milan in 1770, the singers were dubious and forced the teenager to replace six or seven arias. He did so, admirably, and would be asked back the next year (for Ascanio in Alba). His first thoughts survive for Mitridate, and were recorded by Ian Page and Classical Opera as an appendix to their recording of the opera [see my review]. Despite our wish to view Mozart's mature operas as perfect works of art, the composer was in fact always willing to adjust them and they were rarely static.

In 1781, he wrote Idomeneo for Munich. It was a not unproblematic premiere, the work was too long and Mozart cut material for the sake of the drama and the tenor in the title role was demanding. Mozart wished to return to the work, and had ideas for doing a version in German which moved it even closer to the works of Gluck (now that would have been interesting). But in fact the only changes he made were for a concert performance in Vienna in 1786 where he made cuts, added some music and transposed the role of Idamante down from castrato to tenor (leaving the opera with an unwieldy number of tenors). Idomeneo thus remains a tantalising piece, we know that Mozart was dissatisfied with the work yet he never gave posterity the perfect version.

Playbill for the Vienna premiere of Mozart's Don Giovanni in 1788
Playbill for the Vienna premiere of Mozart's Don Giovanni in 1788

With Mozart's three operas with Lorenzo da Ponte, we would seem to be on far firmer ground, after all the composer and librettist supervised the premieres. Le Nozze di Figaro premiered in Vienna in 1786, Don Giovanni premiered in Prague in 1787 and Cosi fan Tutte premiered in Vienna in 1790. But...

In 1788, Don Giovanni was performed in Vienna for the first time with a cast completely different to that which premiered the opera in Prague. Mozart wrote a replacement aria for Don Ottavio, an additional aria for Donna Elvira plus a new duet for Leporello and Zerlina, as well as cutting the finale to save running time. We have no reason to think of the Vienna version as Mozart's final thoughts, it is simply the last version he supervised before he died. And in fact, history has largely ignored it; the duet is rarely performed, and neither are the cuts often implemented. But instead, the two new arias are imported into the Prague version, thus creating a hybrid which was never performed in Mozart's lifetime. Similarly, with Le Nozze di Figaro, when the work was revived in Vienna in 1789 Mozart had to deal with a new Susanna (the original performer, Nancy Storace had returned to London) so he wrote two new arias (which again, are rarely performed). And when the works left the composer's control, managements played completely fast and loose with them, adjusting the pieces to suit local performance needs, there was no ideal version simply one that worked.

This flexibility of form continued in 19th century Italy, when the presence of star singers as co-creators, able to ornament the opera themselves, meant that there was a strong element of re-creation in each performance. Arias would be written for new performers, but you can detect elements of a desire for control. Gioacchino Rossini was happy for ornamentation to be different, and to vary between performers, but he preferred to be involved in its creation, and during his period in Italy he supervised a remarkable number of revivals of his operas, criss-crossing the country by coach, which would, of course, both ensure that he got a fee and that he supervised the tailoring of the music for the performers. Even Rossini's final opera, Giullaume Tell, written for the Paris Opera in 1829, would not be immune and in a remarkable pre-echo of Verdi's struggles with his grand opera Don Carlos, Rossini would work on a number of versions of the piece aimed at reducing it to a manageable size.

This idea of a perfect or ideal version was one that sometimes escaped even Richard Wagner. His opera Tannhäuser premiered in Dresden in 1845, and he continued to adjust the opera through performances in the 1850s then in 1861 Wagner produced a significant new revision which adjusted the work to the requirements of the Paris Opera. And Wagner made further changes for Vienna in 1875, but in 1883 his wife Cosima would write in her diary, three weeks before Richard died, "He says he still owes the world Tannhäuser." The work would not be performed at Bayreuth until 1891, when the Vienna version was used.

Wagner's exact contemporary Giuseppe Verdi composed in a rather different tradition. In his early years (what he later referred to as his galley years), the composer wrote a remarkable number of operas (14 operas in the years from 1839 to 1849) in a manner analagous to his bel canto predecessors, Rossini, Bellini and Donizetti. But in his middle period Verdi had enough income to give him the leisure to consider his works more, so that the middle period operas would be subject to more revision. So that, for instance La Traviata had revisions after the first run of performances, and we would never now dream of returning to the original version. We always perform Verdi's final La Traviata. With Verdi we come to the phenomenon of the publisher; it was during Verdi's lifetime that publishers, notably Verdi's publisher Ricordi, started to exercise significant control not only in terms of when something was performed, but controlling what was performed and in what way. We are approaching the idea of an ideal version which is then placed in aspic.

Verdi: Macbeth - premiere of the revised version in Paris in 1865
Verdi: Macbeth - premiere of the revised version in Paris in 1865

Except, that Verdi was not averse to changes. For a start, he allowed himself to be seduced by the idea of performances in Paris with the changes that might be required. Librettos were translated into French, and sometimes other significant changes were made. So for instance, Il trovatore was premiered in Italy 1853 and a French premiere followed in 1854, in Italian. But in 1856 it was given in a new French version at the Paris Opera as Le trouvère, not just translated into French but with new music including a ballet to turn it into a French grand opera. The temptation here was partially money, performances at the Paris Opera were lucrative for composers, but also the company's large degree of funding so that productions were luxurious and well funded. Verdi's operas for the Paris Opera included both works specially composed and new versions of existing operas, mostly just translating the libretto into French. But where he did add new music, this is rarely performed nowadays.

Sometimes he allowed himself to be more radical. Verdi's Macbeth premiered in Florence in 1847 and proved to be moderately popular in Italy up to the 1860s. But Verdi was seduced into the idea of revising the work for the Paris Opera. At first the idea was to add a ballet and a bit more, but it turned into quite a radical revision. Verdi replaced one of Lady Macbeth's arias, removing one in an old fashioned bel canto style, and performed major surgery on the finale. This is the mature Verdi revising one of his earlier works, not so much correcting it as creating a different ideal version thanks to his development as an artist, much in the way that Beethoven's concept of what his opera should be changed over nine years the opera developed. Verdi's revised version of Macbeth was not a great success in Paris, and did not really take in Italy so that the work effectively dropped out of circulation until the 1950s, when the French verson was preferred. This has become the standard version, though now performed in Italian and without the ballet music.

We sense from this that Verdi had an ideal in his head, and it was simply that this changed as he developed artistically. One work that he nagged away at until it was satisfactory is Simon Boccanegra. This premiered in Venice in 1857, but was not a critical success and Verdi had been unhappy with some of the librettists' work. Finally in 1881, having been introduced to Arrigo Boito by his publisher, Verdi and Boito collaborated on a radical new version which premiered at La Scala, Milan and became a critical success and part of the standard repertoire. It is here, that we see Verdi the artist straining at the ideal version and struggling with the shape of a work created in collaboration with a librettist, or in this case librettists.

Verdi's Don Carlo at La Scala in 1884, premiere of the four-act revised version
Verdi's Don Carlo at La Scala in 1884, premiere of the four-act revised version


The final opera that Verdi wrote for Paris was Don Carlos in 1867, a five act Grand opera in a genre popularised by Meyerbeer in the 1830s and starting to decline in popularity. Even before the premiere there were cuts, it was too long (the French version would only be performed complete in the 20th century after Andrew Porter discovered the cut sections had simply been pasted over in the original score). An Italian version was created (in fact premiered in London!) but the opera was simply too long for the average Italian opera house so unofficial and official cuts were made. Finally, Verdi decided to address the issue and in 1884 premiered a radical new version reduced to four acts, but also with significant amount of new and revised music. As with Simon Boccanegra, we see the later Verdi revising and earlier opera and moving it closer towards what he thinks of ideal at the time. But Don Carlo/Don Carlos would continue to have a curious half-life, and in 1886 Verdi allowed the original Act One to be added to the revised version to create the so-called Modena Version, a strange hybrid between old and new. And since the 1950s, directors and opera companies have played fast and loose with the opera. The Italian version is regarded as prime, even though for each revision Verdi worked with a French libretto (for the 1884 version, a new and more satisfactory Italian translation was commissioned), and the mix-and-match nature of the work's history means that productions often insert bits of the 1867 version into the 1884 revised version in a way which Verdi would not have countenanced. As with other instances, we have made our own ideal version and projected that onto the composer.

The idea of an artist creating the perfect work of art is something of a 19th century Romantic conception, and we have to be sympathetic to the way that 19th century composers often worked in this way. But we also need to recognise more often that the opera business is a practical one, and that the way a work exists reflects as much the circumstances of its creation, so there may be no prime version simply a selection of alternatives. 



The blog is free, but I'd be delighted if you were to show your appreciation by buying me a coffee.

Elsewhere on this blog
  • Go, not knowing where: I chat to pianist Elan Sicroff about Russian composer Thomas de Hartmann and the Thomas de Hartmann project - interview
  • A journey to Anatolia through the ears of The Turkish Five, pioneers of western classical music in Turkey - record review
  • Charmes: an alternative century of song from Olena Tokar and Igor Gryshyn with music by Alma Mahler-Werfel, Clara Schumann, Pauline Viardot-Garcia and Vitezslava Kapralova - record review 
  • 60th birthday celebration: Faroese composer Sunleif Rasmussen's works for recorder player Michala Petri survyed in this engaging and imaginative disc - record review
  • Music of sundrie sorts, and to content divers humours: Byrd's 1588 Psalmes, Sonets & songs of sadness and pietie in its first complete recording from Alamire - record review
  • Bringing audiences into closer contact with the poetry: tenor Ilker Arcayürek on the art of the song recital and his new disc of Schubert songs - interview
  • Scholarship and enjoyment combine in Il Gusto Barocco's lovely fresh account of Bach's Brandenburg Concertos - record review
  • Proud Songsters: a survey of English song by ten distinguished alumni of King's College, Cambridge on the college's own label  record review
  • A Healing Fire: Greek guitarist Smaro Gregoriadou in music by Bach, Britten, Gubaidulina & Hetu played on instruments by George Kertsopoulos  - record review
  • Super-excellent Gabrieli and RVW on viols: National Centre for Early Music's Awaken festival -  concert review
  • He can chisel a mood from just a few bars: pianist Peter Jablonski talks about his new disc of music by Alexey Stanchinsky, and about exploring the darker corners of the repertoire  - interview
  • A musical microcosm of 2020: Isolation Songbook from Helen Charlston, Michael Craddock and Alexander Soares on Delphian - record review
  • Home

No comments:

Post a Comment

Popular Posts this month