It is common knowledge that the musical text of 18th century operas was not inviolable, but varied from production to production, this being definitely a matter of course. The composers took an active part in these transformations; Christoph Willibald Gluck, for example, remodelled his opere serie Orfeo ed Euridice and Alceste, which were decisively important for the contemporary discussion of opera, into French Tragédies en musique. Joseph Haydn arranged the repertoire for performance at the princely court opera house at Eszterháza; he shortened, transposed, changed the instrumentation, the musical form, or the formal design of certain numbers, added new arias and exchanged others1. Wolfgang Amadeus Mozart composed insertion arias as well as ensembles for opere buffe by Pasquale Anfossi, Domenico Cimarosa, Martín y Soler, and Niccolò Piccinni, among others, and revised his own operas, testified for instance by the discussion about the Prague and Viennese versions of Don Giovanni.

In the years 1789 to 1792 Luigi Cherubini, eventually, adapted operatic works for the Parisian Théâtre de Monsieur. For these productions he composed new arias and ensembles2, and modified the musical text of individual arias3. In Cherubini’s Italian operas manifold forms of arrangement are manifest: some of the scores show traces of cuts like bars glued over4, where doubts remain, if the cuts were introduced by Cherubini himself or originate with some anonymous author; the same applies to an aria come down to us in two different instrumentations5. When performed with a new cast, Cherubini wrote alternative or insertion arias as in Paris in later years6.

All these aspects, though worthwhile of investigation, shall be excluded from the following discussion, instead of which correspondences and conformities in the musical substance in various musical contexts will be traced. More often than hitherto known, Cherubini used earlier works of his own for his operas. To begin with, this vigorous and productive practice of re-arranging will be described in more detail; secondly, the attempt will be made to link it provisionally to the institutional conditions, under which Cherubini brought his operas to performance.

Recourses and Revisions

The slow introduction of the sinfonia of Mesenzio, premièred in Cherubini’s home town of Florence in the autumn of 1782, sets it apart from other Italian operas. (Of the Italian operas only Giulio Sabino, which was premièred in London in 1786, also possesses a slow introduction.) Regarding these introductory measures there nevertheless is a close connection between the Mesenzio-overture and the following opera Quinto Fabio, composed by Cherubini for the Teatro Argentina in Rome in 1783: With the exception of the tempo instruction, the introduction of Mesenzio corresponds exactly to the first measures of Quinto Fabio7. But the correspondences are not exhausted with the adoption of the introduction, as Maximilian Hohenegger has recorded in the only study addressing Cherubini’s Italian opera sinfonias up to now:8 Cherubini rather appropriated extended portions of the sinfonia for the first movement of his new opera. In this connection it is striking that Cherubini did not adopt entire formal sections, but only particular elements.

Considering the main subjects this principle of patchwork gets even more lucid. Both overtures come to rest at a dominant six-five chord in m. 15 (Mesenzio) and m. 16 (Quinto Fabio), respectively, the cadences to each being identical. Likewise, the following ten measures are the same. The opening motif of the main subject, though, appears in a modified form. This, however, is no new idea for Quinto Fabio; Cherubini instead had recourse to the opera he composed before Mesenzio, Adriano in Siria, with which the Teatro dagli Armeni in Livorno was opened in spring 1782. At the beginning of Adriano there is also an introduction, which, however, is not set apart from the main subject by tempo. It consists of a broken chord in the principal key of D major, which in three starts of the full orchestra goes upwards over two octaves, and a descending line, broken by semiquaver upbeats, from d1 to g, the entire passage in unison. Cherubini’s sinfonia to Adriano in Siria, on the other hand, corresponds with the exception of some details to the first version of Quinto Fabio, produced in Alessandria in the autumn of 17809. In comparison with the original version, however, the instrumentation for the Livornese opera house was enlarged by trumpets, timpani, and a violoncello obbligato.

But this is not yet the end. Following the recourse to Mesenzio, in m. 27 Cherubini once more changes his source and again makes use of Adriano. And he begins, to be precise, almost exactly where he left off. At this point also, all four overtures overlap momentarily.

The overtures to the Roman Quinto Fabio as well as to Mesenzio differ with regard to the instrumentation, too. For Florence Cherubini had determined on an orchestra including one flute, two oboes, two French horns, two trumpets, timpani, strings and basso continuo, whereas in the first movement of the sinfonia of Quinto Fabio the flute and timpani are omitted. Not until the second movement Cherubini uses two flutes, although no oboes; in the third movement just the reverse is the case. That is why the four sinfonias still sound different, irrespectively of the musical correspondences:

It was not the only time that Cherubini had recourse to an earlier composition: The overture to Idalide, premièred carnival 1784, again at the Florentine Teatro della Pergola, he used largely unaltered for Alessandro nell’Indie, a production for Mantova only a few months later. The differences here concern primarily the scoring – conspicuous especially in the second movement – : For Florence the composer made use of one flute, two oboes, two French horns, two trumpets, timpani, strings and basso continuo, in Mantova he increased the scoring by two additional French horns and bassoon, but did in return without the flute and the timpani11.

This re-scoring was not without consequences, particularly for the slow movement. Here Cherubini introduces three concertante instruments at a time to an accompaniment of the violins, playing piano, and plucked basses. In Florence the three instruments comprehended the first oboe, a violoncello, and the flute, which obviously was not available to Cherubini in Mantova. Thus he adjusted the movement, at the same time including the bassoon, which constituted a distinctive feature of the Mantova orchestra: the flute’s part was assigned to the viola, the violoncello Cherubini retained, and the oboe was replaced by the bassoon. The exceedingly high position of the flute and oboe in the overture to Idalide Cherubini attenuated in Alessandro. For the instruments involved, namely viola and bassoon, the parts are, however, still in rather a high position. On occasion Cherubini accordingly modified the melodic line.

Twice Cherubini altered the melodic line: right in m. 9, in which the oboe rises up to b2, he remodelled the semiquaver motion in swelling sequences into a triad motif, which avoids on the one hand the topmost notes a1and b1, at the same time appearing adequate to the bassoon’s characteristic timbre. The semiquavers in m. 10 are modified, too, and reduced in pitch. (At first sight, the beginning of m. 11 seems to be different, but this is merely a variant of notation.)

Similar procedures of adoption are, of course, not limited to the overtures. The two Quinto Fabio settings mentioned above especially give rise to the question if and to what extend Cherubini could transfer parts of his first opera to the Roman production. A comparison between the librettos, however, shows that there are no identical numbers in both versions. The comparison of the surviving musical material confirms these results. Because of the fragmentary character of the first opera’s surviving material one nevertheless cannot rule out the possibility that the composer may have parodied one or other of the arias.

While the comparison between the two Quinto Fabio-librettos at last turns out as a quite fruitless track, the search for identical aria texts proves to be much more promising: an abandoned opera project of Cherubini possibly is connected to Giuseppe Sarti’s production of Giulio Sabino, for which Sarti and his pupil Cherubini presumably took residence in Venice. His plans are thus outlined:

“Dans le courant de 1781, il contracta l’engagement d’aller composer un opéra à Venise, où il resta peu de temps, attendu que pendant qu’il le composait et avant qu’il fût achevé, l’entrepreneur du théatre fit banqueroute, et Cherubini revint à Milan, près de Sarti”12.

When the impresario went bankrupt, Cherubini already had composed two musical numbers for the opera in the making. One is the aria “Caro padre, amato sposo”13, the other is the scene “Morte, morte fatal!” for the first act, consisting of a recitative and an aria. From the accompagnato’s text mentioning the name “Silla” one can infer that the opera was meant to be one Lucio Silla14.

An aria “Caro padre, amato sposo” is also to be found in the libretto of the Roman Quinto Fabio, where it is written for the role of the prima donna Emilia. The given text obviously reminded Cherubini of his earlier composition. For Quinto Fabio he had recourse to it. A comparison between both texts initially seems to show that only the beginnings are the same, but metre and rhyme pattern, these basic elements of composition, are identical, too15. Hence Cherubini could use his composition without major changes. Modifications turn up chiefly in the voice’s part and appear to serve primarily the simplification of the coloraturas.

In mm. 71–76 of the Venetian version Cherubini devised a thrice ascending scale, the voice rising to the peak tone c3 at the third time. In Rome the rising was shortened by one ascent and moreover lightened by descending steps in between. In the parallel passage the coloraturas were also diminished. A variant with several short semiquaver motifs takes the place of continuous coloraturas for two and a half measures. It closes with a long sustained note f2, which then drops to f1. Since this drop accompanies the words “Torna in calma”, one is inclined to presume that the arrangement was fostered to some part by tone painting reasons. Additionally, the cadenza was eliminated in the arrangement, a shortening typical for the late 18th century, which is also to be found in Cherubini’s Adriano16.

Eventually, recourses to motifs beyond the boundaries of Italian opera production also occur: the sinfonia of his first opera buffa Lo sposo di tre e marito di nessuna from 1783 Cherubini took up again in his opéra comique L’Hôtellerie portugaise from 179817, and his last Italian opera Ifigenia in Aulide, which Cherubini worked on already influenced by his first Parisian impressions, offered material for two French works for the stage.

The entire movement is run through by the pounding theme or its beginning motif, respectively. The theme itself is structured periodically, extending over 28 measures altogether, with the characteristic shift to the dominant in m. 16. The first phrase is split into four motifs of two measures each, the first motif being mirrored in the second. The third motif is a repetition of the first, but with a closing sixth, with which Cherubini disrupts the unison and initiates the concluding chords.

To what extent the beginning of the theme can change its character is already revealed in m. 17. While Cherubini at the outset presents the full orchestra including flutes, oboes, and clarinets, French horns, trumpets and timpani forte or fortissimo, respectively, he reduces the orchestration at this point to violins and violas, restating the beginning piano with soft dabbing staccati. Although the theme’s head returns identically in the second violins and violas, it appears freshly illuminated by the first violins’ embellishments and garnishes.

Cherubini pursues this line further in the second theme: the head motif or rather its mirrored form, modified in pitch, becomes part of the accompaniment in m. 45. Although the characteristic rhythm, consisting of six crotchets and one minim, remains clearly perceptible, it is enlivened by the violas’ quaver repetitions on d1. On this Cherubini superimposes the swinging melody of the solo flute.

The manifold usability of this theme becomes even more apparent in its subsequent “fate”: following Arnold Jacobshagen, Cherubini made next use of it in the unfinished opera project Koukourgi from 1792/9318. Large parts of this opéra comique were incorporated into Cherubini’s last opera Ali Baba ou Les Quarante voleurs. In this way the main theme of the Ifigenia-sinfonia found its place in the later fairy tale opera.

In Ali Baba the theme appears two times right away: at first likewise in the overture, namely in the presto-coda. The chief melodic modification Cherubini made in comparison with Ifigenia is the dissolution of the tone repetitions at the beginning into broken chords. Because of the tempo acceleration to presto and the extended dynamic intensification, increasing to fortissimo, the theme again appears in a new guise. In the orchestra Cherubini summoned up piccolo flutes, trombones, an ophicleïde and several percussion instruments. This point of culmination of the overture Cherubini then adopted in the ballet of the 2nd act19. The passage acquires an entirely new function as accompaniment for the dances of the female slaves of the rich oriental merchant Ali Baba20.

Concerning the institutional prerequisites of Cherubini’s Italian opera works

The differences in instrumentation as shown between the sinfonias of Quinto Fabio (1st version), Adriano, Mesenzio, and Quinto Fabio (2nd version) on the one hand, and between the overtures to Idalide and Alessando on the other hand, can probably be attributed to differing orchestral capacities. Not only in the overture of the Roman Quinto Fabio the combination of flutes and oboes is lacking, but throughout the score flutes and oboes do not appear at the same time, likewise the lacking timpani. The reason for this is probably that Cherubini could make use of only two players for flutes as well as oboes, and no timpani player at all. In Florence, on the other hand, he had an orchestra at his disposal that, according to the contract between the theatre society and the impresario, consisted of 38 musicians and was put together “da due Cimbali, da quattro Contrabbassi, da un Violoncello, da quattro Viole, da quattro Strumenti a ottone cioè due Corni e due Trombe, da due Obue, dai Timpani, e da venti Violini con facoltà all’Impresario di poterli ridurre a Sedici quando credesi necessario di sostituire altri strumenti di diversa natura, purchè tal variazione non porti diminuzione al numero totale degli Strumenti convenuto”21.

The list of the theatre’s personnel above falls slightly short of the society’s requests. Instead of four double bass players the list only mentions three; the same applies to the viola players. The projected 20 violins are reduced to 16 – a margin the contract fully permits – and at least partly replaced by other instruments: There were two violoncelli instead of one, after all, and the third player for flute or oboe, respectively, is also beyond the contract’s requests. Eventually the score of Mesenzio, requiring three players for trumpets and timpani in all, indicates that the surviving list possibly is not quite flawless23.

A kindred case is that of Idalide and Alessandro nell’Indie. In Mantova, too, there obviously were only two players for flutes and oboes; there is no musical number were both instruments are combined. Likewise timpani do not appear in Alessandro. Instead of these four horns come into play as well as a bassoon, which moreover plays a prominent solo part in the opera24. In one aria one of the horn players exchanged his instrument for a cor anglais25, and in the marcia of the ultimate scene Cherubini specifies two “Corni piccoli in bB”. The audience was to profit of the given means to the utmost. In a similar manner the vocal parts were modified, in order that the singers could excel to their best ability.

The general custom of having recourse to former works can in Cherubini’s case be defined even more precisely: in his oeuvre of overtures the sinfonias to Quinto Fabio (1780), Adriano in Siria, Mesenzio, and Quinto Fabio (1783) as well as to Idalide and Alessandro nell’Indie are closely linked. In addition Cherubini made use of motifs from the sinfonia to Lo sposo di tre in the French opera L’Hôtellerie portugaise as well as from Ifigenia in Koukourgi and later in Ali Baba. Among the arias the numbers with an identical textual base offered themselves for a rearrangement.

Disregarding his early intermezzi, the practice of transfer starts with Cherubini’s opera oeuvre, and ends with his last work for the stage. The crucial point here is the fact that these operas each were composed for different towns and cities: for Alessandria, Livorno, and Florence or Venice, respectively, and Rome, for Florence and Mantova, and eventually for Venice or Torino, respectively, and Paris26. The borrowings did not strike the audience. But the extensive transfers besides lend probability to the assumption that Cherubini did not count on his operas developing into repertoire. Instead of this he was conscious of the traditions of Italian opera enterprises which constantly demanded works never been heard before. This demand Cherubini was able to satisfy with his practice of borrowing. When burdened with a large amount of work these borrowings from earlier compositions gave the chance to reduce the expenditure of work and to present successful musical numbers again.

The procedures of patch-working as exemplified in the sinfonia of the Roman Quinto Fabio can not, or at least not entirely, be accounted for by the economical use of working powers, for Cherubini could have made himself even more comfortable by a simple re-rendering of the overture (apart from the necessary changes of instrumentation). Instead of this his method gives the impression of experimenting, of searching for various opportunities of combination and connections. In these instances Cherubini really did compose – not in the emphatic, but in the literal sense of the word. He joined already existing material – in his case coming from his own workshop – afresh, and incorporated it into new contexts.

Cherubini’s example herewith confirms, too, that the routine of arrangements is a matter of course in 18th century opera practice. Hence it follows in particular that more examples are certainly to be found in his oeuvre. The discovery of these will be clearly facilitated by the scholarly critical edition of Cherubini’s works27.

Translation: Silke Schloen