7—
I compositori scenici :
Librettist and Composer

Marco Faustini's failure was intrinsic to the system of opera production that had nourished him. The institutionalization of opera had initiated a chain reaction. The increased demand for new works intensified the pressures on a limited pool of opera makers, whose expansion required the exploitation of additional (including foreign) suppliers, which in turn created difficulties of communication and coordination at every stage in the preparation of a production: at conception, composition, and rehearsal. Conception was impeded by the fact that the libretto was often written before the cast had been assembled, making it difficult for the poet to decide on the number and importance of the various characters. Composition suffered from the geographical distance between librettist and composer, who often resided in different cities and therefore could not work together;^[1] furthermore, like librettists, composers did not always know their casts in advance and were unable to tailor their music accordingly. Finally, the rehearsal period became increasingly difficult since, in addition to providing the opportunity for the normal ironing out of staging problems, it had to allow for many alterations involving text and music that would formerly have been taken care of earlier, at the stage of composition or even conception.

One of the significant results of institutionalization, then, was a change in the relationship among the makers of operas, among librettist, composer, and performers. Originally part of a single impetus, the three functions grew increasingly separate and independent of one another. This growing independence, even as it reflected the development of opera, in turn had its own impact on that development.

Collaborative Talents

In Ferrari's tiny troupe, whose Andromeda and La maga fulminata had been the first operas in Venice, librettist, composer, and performers were closely linked—by geographical proximity, background, and common goal. Ferrari, being a musician himself, although he did not actually compose the music for either opera, was certainly sensitive to the requirements of music, and the collaboration between him and Manelli must have been worked out on an intimate, daily basis. Both text and music must also have been precisely fashioned to suit the abilities of the singers in the company. The multiple talents and responsibilities of the various members of the troupe—Ferrari, as we know, played the theorbo in the orchestra, and Manelli sang—bridged the boundaries not only between text and music but between creation and performance as well.

The relationship between librettist and composer was of course closer still in those subsequent operas for which Ferrari served as both: Armida (1639), Il pastor regio (1640), La ninfa avara (1641), and Il principe giardiniere (1644). Unfortunately, we are unable to evaluate the results, since none of these scores have survived.^[2] Although a number of composers began their operatic careers as singers, including Manelli, Cesti, and Boretti,^[3] few of them combined Ferrari's publicly recognized "multiplici virtù," which suggested comparison to "un mostro diversamente simile alla chimera,"^[4] and made him an ideal maker of operas. But he considered himself primarily a musician rather than a poet, claiming that his most important attribute as a writer was knowing what kind of poetry was best suited to musical setting.^[5] To the extent that the two can be

separated, then, the composer in Ferrari clearly held sway over the poet. His self-evaluation reveals his attitude toward the relative importance and stature of music and text, at least in his own works. That attitude, the subservience of text to music, or of the librettist to the composer, is reflected more generally, though only fleetingly, in the operas of his time.

Indeed, several early composers of Venetian opera, although they may not have shared Ferrari's poetic skills, exercised control over their poets in various ways. The most striking example, of course, is Monteverdi, whose influence on the texts of his operas was fundamental. His famous lessons to Giulio Strozzi, revealed in letters he wrote concerning their collaboration on La finta pazza Licori , involved dramatic issues as fundamental as plot structure, characterization, and verisimilitude.^[6] Monteverdi also instructed his other poets in the art of libretto-writing, although our evidence for this is somewhat more indirect. One of them, the anonymous author of Le nozze d'Enea , confessed to having made a number of changes in his text in order to suit Monteverdi's style: "I have shunned remote thoughts and concepts and rather been attentive to the affections, as Signor Monteverdi wishes, and in order to please him I have also changed and omitted many of the things I had originally included" (Appendix I.9g). Another, Badoaro, gained his education in retrospect, claiming that he hardly recognized his Ritorno d'Ulisse in Monteverdi's setting because of the many changes made by the composer.^[7]

To be sure, Monteverdi, like Ferrari, represents a special case, though in a different way. His long experience and many successes had earned him a reputation as the greatest opera composer in Italy. He was credited by at least one seventeenth-century observer as having been the moving force behind the development of opera in Venice, although his participation in the Venetian opera scene was delayed somewhat, as we have already noted.^[8] Furthermore, Monteverdi's commitment to the ideals of the seconda prattica placed him in a special position with regard to words. In order to investigate fully the power of his music to communicate feeling through text, Monteverdi essentially had to create the text himself, often using what was provided by the poet as a skeleton to be fleshed out with repetitions, cuts, emphases, and so on (this will be examined further in chapter 9).^[9]

Close working relationships seem also to have existed between other early composers and their librettists. Cavalli, for example, exercising his prerogative as impresario in some cases and in others his position as the most important and respected opera composer in Venice after Monteverdi, personally commissioned or arranged for many poets to write for him: these included Persiani, Busenello, Faustini, Melosio, Minato, and Aureli.^[10] They, in turn, knowing the composer for whom they were writing, could fashion their texts to suit him. Indeed, the scores of a number of Cavalli's operas reveal that the music actually took shape along with the text—much as we imagine the Ferrari-Manelli creations to have emerged. Several, including Veremonda (1653), Xerse (1654), Statira (1655), and Arternisia (1656), preserve earlier versions of the text than those in the printed librettos, whose development out of those earlier versions can be traced in the scores themselves.^[11]

In a more general way, Cavalli's long-term working relationship with Giovanni Faustini, which produced at least ten operas, must have influenced the poet's style. Faustini's lengthy affective monologues in versi sciolti punctuated by refrains may well have been written to suit Cavalli's gifts as a composer of expressive recitative. Furthermore, Faustini's practice (though hardly his alone) of postponing resolution of his complex plots until the final scene, immediately preceded, even triggered, by a protagonist's lament, may have developed in response to Cavalli's skill at composing affective laments in a mixed recitative-aria style.

The librettos of Cavalli's only other steady collaborator, Minato--who may have been coerced by the composer into writing texts for him^[12] —were in various respects modeled on those of Faustini; he too punctuated lengthy monologues by refrains and wrote short strophic texts to be set as arias as well as extended lament texts. The composer indicated his appreciation of Minato in his response to Marco Faustini's request for new works in 1662. In a letter to the impresario of 8 August (Appendix IIIA. 3), he explained that he had decided to abandon operatic composition altogether, but that Minato's insistence had persuaded him to change his mind.^[13]

While Cavalli's impact on Minato's texts can only be inferred, his specific influence on a third librettist, Aureli, is documented. The young Aureli consciously modeled Erismena (1655), his second libretto, on the style of Cavalli's regular librettist, the recently deceased Faustini, even borrowing a lengthy passage from Faustini's Ormindo . The borrowing, from an opera Cavalli had set to music more than a decade earlier, was certainly the cornposer's idea, since he used his old music, a particularly moving lament dialogue, in the new context. Erismena became one of the most successful operas of the entire century.^[14]

Cavalli was the last (possibly the only) composer of the period to wield power as an impresario. The rapidly increasing complexity of the operatic scene must have made that role unbearable to him, for he gave it up quite early in his operatic career, choosing to work under contract, first with Bortolo Castoreo and others at S. Cassiano, then with Giovanni Faustini at S. Apollinare and with Marco Faustini at S. Cassiano and SS. Giovanni e Paolo, and finally with Mi-nato at S. Salvatore.^[15] Even after he renounced his managerial responsibilities, his reputation was sufficient to assure his continued aesthetic control over his subsequent works, but his early shift from employer to employee was one sign of more general adjustments in the operatic hierarchy, of the decline in the composer's comparative status with respect to the librettist—and soon the singer.

Three special individuals—Ferrari, the composer-librettist, Monteverdi, the "creator of opera," and Cavalli, the dean of Venetian opera composers and its leading businessman^[16] —maintained the power of the composer somewhat artificially. Competing forces gradually undermined that power, however. Perhaps this change was part of what Ferrari had in mind when he complained so poignantly, already in 1643, that he felt obsolete:

Although a veteran of scenic compositions, I am distressed by the refinement of the century, its factions torment me, but virtue, in the end, is a sun that, despite the clouds of passion surrounding it, never ceases to shine. These days musical theaters attract the most melodious swans; and sirens yearn to be angelic instead

of maritime. . . . With your usual kindness, forgive the defects of my Giardiniere , of my music, and of my theorbo. (Appendix I.4)

The death of Monteverdi, Cavalli's withdrawal from managerial duties, and Ferrari's lament about the changing order of things, all of which occurred around the same time, mark a shift. During the course of the 1640s, the various mechanisms of propaganda used to stimulate the growth of opera in Venice, in particular the printing press, focused with increasing exclusivity on the new professional librettist.

As printed scenarios, librettos, and reports of performances make increasingly clear, the librettist was considered "l'auttore." It was his text that was immortalized through print—usually, as we know, at his own expense (Appendix II.6bb)—and his name that was associated with the opera in the mind of the public. He was usually the recipient of any laudatory sonnets printed in the libretto, though an occasional singer or composer might be eulogized in the same way (often by the librettist himself). The librettist's preeminence, even as late as 1671, was recognized in the following praise of Nicolò Beregan:

The opera of Sig. Nicola [i.e., Heraclio ] is so beautiful that it never becomes tedious. So lofty is its governing idea that Sig. Nicola has been deified in that theater [SS. Giovanni e Paolo], where without his say-so no new production will be undertaken, and if he invents new worlds of exquisiteness, everyone will converge to admire him, since there is really no one comparable to him in invention, magnificence, and summit [Piramide ] of imagination. (Appendix IIIB.6)^[17]

It was also the poor librettist who had to bear the brunt of the public's fickleness, to respond to the taste for novelty, and to produce a constant stream of texts rapidly, on demand.^[18] Librettists often compounded their responsibility and control over operatic productions by assuming the duties of impresario—as both investors and organizers. Indeed, unlike the composers, Cavalli being a temporary exception, most of the early librettists, including Ferrari, Faustini, Minato, Aureli, Beregan, and Pietro Dolfin, were involved in theater management as well.^[19]

Obviously, composers did not completely lose their influence on the work, even in the eyes of the public. And their names were often included on the title pages of librettos or mentioned in prefaces, almost always in positive, if rather conventional, terms: the composer, variously called the Apollo (Monteverdi, Cavalli, Sartorio, Boretti) or the Amphion (Cavalli, Ziani) of the century, or else the Sun (Monteverdi and Cavalli) or the Moon (Sacrati), often covered or minimized the imperfections, enlivened the inanimate corpse, or sweetened the bitterness of the libretto;^[20] or he may have honored, enriched, ennobled, adorned, illustrated, or otherwise favored it.

Composers, as well as scene designers, costumers, and singers—the most obvious creators of pure spectacle—were presumably mentioned in the hope that their names would attract the crowds and thus guarantee the commercial success of an opera.^[21] To that end, various flattering biographical details were occasionally supplied, such as the fact that a composer served one or another princely patron. But in more than half the librettos printed between 1637 and 1675, composers' names did not appear at all. We have seen that Giovanni Faustini completed a number of librettos for which no specific productions had yet been planned. They were written for performances in the indefinite future, in a theater and with a cast and composer as yet unknown. Although Faustini was especially industrious, the dissociation between libretto and eventual setting was becoming increasingly characteristic of opera. Librettists, under pressure to produce works quickly, and assured of a steady market, began to write and print texts ahead of time, often before all of the other arrangements for a production could be concluded.

Librettists' Tribulations

Geographical separation, difficulties with last-minute arrangements—the hiring of singers, the completion of the score—all this placed a severe strain on everyone concerned in operatic production. If in many cases librettos were written before composers had been engaged to set them, in others librettists had to write their texts with composers breathing down their necks. Both situations were symptoms of the same disease: the lack of coordination between supply and demand that resulted from the institutionalization of opera.

Time was sometimes so short that composers were forced to begin setting librettos before they had been completed, presumably without even having read the whole text. A number of librettists remarked defensively on the conditions that forced such piecemeal settings. One (Francesco Sbarra) complained publicly to a friend that because of illness and the pressures of a deadline for performance, he had to send his text off immediately to be set (by Cesti):

Because I didn't have even the minimum time necessary to check it over, I had to allow it to be placed under the notes in the very same form in which I sketched it. Furthermore, since it ended up being too long for the music, it had to be shortened, and consequently it can only have been mangled; and they tell me that they had to print it in that form. (Appendix I.29c)

And he begged his friend to see to it that the libretto was corrected before being printed:

I appeal to Your Lordship to do me the favor of checking this work and altering it as you see fit before it is printed. . . . You will find many errors of language, numerous weaknesses and harshnesses in the poetry, and in the humor an infinity of our local idioms that are unsuitable elsewhere, in addition to many dissonances caused by the shortening of the work. (Appendix I.29d)^[22]

Perhaps the most vivid description of piecemeal composition was that by the librettist of Hipsicratea (1660), Giovanni Maria Milcetti:

This work was written in Badia delle Carceri and from there sent, page by page, to Murano, whereupon it was immediately structured into scenes. Anyone who knows the distances between the two places will not be surprised by the differences in poetic structure and words between the printed libretto and the opera performed in the theater, because, in effect, even before the poetry was finished and organized it was already being set to music, and the sinfonie were already being sung. (Appendix I.53a)^[23]

As usual, rhetoric mixes with reality in these complaints. Like the librettists' common disavowal of serious commitment to writing, the excuse of "l'angustia del tempo" was part of their characteristically apologetic preemptive strategy. In an amusing variation on this approach, a few authors admitted to inordinately long gestation periods for their librettos. Matteo Noris's first libretto, Zenobia (1666), for example, took him four years to write, while Be-

regan claimed that his second libretto, Tito (1666), took five.^[24] The seriousness of long-term commitment was evidently deemed as admirable as the sprezzatura of speedy composition.

Whether or not they wrote under pressure of time, librettists had to be ready to make changes in their texts, either at the request of the composer or in response to the requirements of a particular performance. It is quite clear from a variety of evidence that it was only during the rehearsal period that an opera took its final (though not necessarily permanent) shape. The librettist as well as the composer was expected to attend rehearsals in order to accommodate these alterations. In one instance the author of Adone (1639), Paolo Vendramin, wrote to the composer, Manelli, explaining that he could not be present for the rehearsals, and charged him with the responsibility of overseeing the production: "It will be up to you to lend it that study and that diligence that I could not, and which is required when producing an opera in Venice."^[25] In another instance, the librettist Francesco Piccoli "being unable to assist in the completion . . . nor in the alterations, indeed changes, required by the staging" of his L'incostanza trionfante overo il Theseo (1658), it was necessary to engage another librettist for the final changes.^[26]

In a note at the end of his Hipsicratea (1660), the author Milcetti asked the composer Don Pietro Molinari to add some arias to his text because he was unavailable for final alterations:

Since distance does not allow me to be around to see to the needs of my drama, I am happy that Your Lordship will insert the songs [canzonette ] that you describe. I beg you to mark them in the margins with a star or with double commas, as is the custom. I say this because I do not like to dress what is mine in others' ornaments; thus I will be extremely glad if the difference is clear. (Appendix I. 53b)^[27]

Composers must often have been required to add text, particularly for arias. In sending his libretto Lismaco to its composer, Giovanni Maria Pagliardi, Ivanovich encouraged him to make whatever changes he thought necessary:

If you find some emotion [affetto ] in the recitative that can be reduced to a cavata , don't hesitate to do so, since anything that stands out unexpectedly is pleasing. As far as the number of ariette is concerned, to be arranged with their sinfonie , use your discretion, with due attention to the brevity that is so important here [in Venice]. (Appendix II.5b)

The numerous textual alterations that occurred either at the time of setting to music or during rehearsals created a special problem at the publication stage. If printed librettos, like their less expensive predecessors the scenarios, were to be of use to the audience in the theater, they had to correspond to tire work being performed on stage. This was rendered difficult by the length of librettos—too long to be printed in the period, at most a matter of a few days, between rehearsals and premiere. As it was, librettos often did not come off the press until the very last minute.^[28] Printers compensated for this difficulty by devising various methods of incorporating last-minute alterations. In fact, the frequency and extent of such changes is confirmed by the development of standard techniques for indicating them: the use of virgolette , which could be added on the printer's copy of the manuscript even at the last minute before typesetting, and brief stop-press additions, which could be set at the bottom of pages that had already been printed.

At first, at least, virgolette gave the clearest message: they indicated text that was not sung. I have already mentioned one of the earliest uses, that in the 1640 print of Rinuccini's Arianna , which followed the revival of Monteverdi's opera in that year. Although they were unexplained, the virgolette were placed next to passages that were obviously cut in the revival (nearly all of the choruses, as well as some of the dialogue)—those passages, in fact, that would have been anachronistic in Venice. The printer—and probably Monteverdi too—had edited the text according to Venetian usage ("secondo l'uso di Venezia"), without acknowledging as much.^[29] At first it was usually recitative that was cut, but as the century wore on virgolette tended to appear next to second strophes of arias as well.^[30] Although they retained their original function throughout the cen-

tury, virgolette were also increasingly used to distinguish material added to a libretto by another hand.^[31]

For more complicated situations, and when time did not allow for virgolette , librettists used the preface to explain discrepancies or else resorted to stop-press additions. Since these affected only a single page of type, they could presumably be produced at the very last minute, conceivably even on the day of the performance, even though librettos were stitched (though not bound). A note in the preface to Aureli's Gl'amori infrutuosi di Pirro (1661), for example, explained that the libretto had been printed too soon to incorporate virgolette .^[32] The tightness of schedule is nicely illustrated by another Aureli libretto, La costanza di Rosmonda (1659), in which virgotette appear in the second and third acts only. A last-minute postscript explains the discrepancy: the first act had been printed before cuts were made so it appeared without virgolette , whereas the other two acts had been printed late enough to incorporate them (Appendix I.44a). Of course, we have no idea which lines of act I were affected; we know only that some were cut.

What postscripts lacked in specificity, however, they made up in flexibility, for they could convey many different kinds of information about a variety of last-minute changes, and not just cuts. Most important, they often justified such changes, thus providing valuable insight into aesthetic and practical considerations affecting opera production. At the end of his Delia (1639), for example, Strozzi informs the reader that his text has been altered by the composer, Manelli: "So that your eyes will agree with your ears, know, exquisite reader, that in representing the work more than three hundred lines have been omitted, in order not to abuse your courtesy" (Appendix I.15i). And then he proceeds to justify the cuts: "It is necessary that the poet abandon his ornaments, that is, his digressions and episodes, to make way for the singers' ornaments. Do not, therefore, fault the performers for doing what they have done the better to serve you" (Appendix I.15j). In one of his later librettos, La finta savia (1643), Strozzi again explains that the unusually lengthy text has been cut because it was too long for musical setting, having been written to be performed without music as well (Appendix I.18a).^[33]

Although some changes clearly occurred when the composer initially set the text to music and may have been justified by aesthetic considerations, many more must have been made at the last minute in response to specific performance conditions. In Veremonda (1653), for instance, several lines were added in a scene of the second act: "Since, owing to the distance of the Fort of Calpe, Zelamina and Zaida couldn't be understood, they have been made to come outside to speak, and then, having seen Delio, in respect for Giacutte, to go back in until he leaves. As a consequence, the following lines in the middle of act 2, scene 8, must be altered and added."^[34]

Whereas librettist and composer had to make these small adjustments on the spot, some changes were effected simply by a new direction to the singer (with perhaps a record of it left in the libretto), as in La costanza di Rosmonda (1659): "In act 1, scene 14, where it says 'Clitennestra recites in a balcony,' it was thought better to have her appear on the stage in order to make her visible to the eyes of everyone, especially those seated in the boxes" (Appendix I.44b).

The preface to Aureli's Claudio Cesare (1672), finally, summarizes quite effectively the kinds of changes that took place during rehearsals and the typical librettist's attitude toward them:

After the libretto was printed and the opera rehearsed on stage, it was decided to cut various superfluous parts of it; therefore, you are urged kindly to pass over various lines and certain scenes, which for greater brevity have been cut, since we did not have the opportunity of doing so, the libretto already having been printed. Below you will also find three aria texts that have been changed; all of this information is given so that only what is read will be sung, thus bringing greater delight to the audience. You will appreciate the difficulties that composers [compositori , meaning composers and librettists] face today in satisfying not only the strange tastes of this city, but also the extravagant whims of the performers [i Signori Musici recitanti ]. (Appendix I.50b-c)

Composers' Obligations

The kinds of alterations that so preoccupied librettists were of less concern to composers; or at least their concern was less obvious. One reason for this difference was sociological. Librettists, however inadequate, casual, or amateurish they claimed to be, were educated in and practiced lettere : they were writers, whose words laid traditional claim to immortality, especially since the invention of the printing press, and whose capabilities were judged by well-

established critical criteria. Composers, on the other hand, no matter how intelligent and well educated—and whatever the higher claims of music as theory—were essentially artisans, practitioners of a trade, for hire. Theirs was a service profession. Last-minute modifications determined by the demands of a patron, an occasion, or a particular performance site were a traditional part of their job. Such changes were, of course, more fundamental to the structure of librettos than of scores: they involved adding or subtracting characters, eliminating precious lines of poetry, metaphors, special turns of plot, and other such invenzioni . For composers the changes were minimal: a transposition here, an aria there.

Librettists might have been more relaxed had their works remained as ephemeral as opera scores or performances. It was the act of printing the libretto per se—a sign and consequence of institutionalization, as we have seen—that stimulated these writers' anxiety and caused them to express their concern so openly. Their professional identities (as well as their material profits) were embodied in the published work.

Librettos enjoyed an independent existence as printed texts. Scores did not; they were working documents, tied to the facts and moments of performance. Whereas it was to the librettist's advantage to distinguish between the composition and rehearsal stage in the development of his text, such a distinction was hardly relevant to the composer, for whom the two stages merged in the normal process of composition. Indeed, composition was not considered complete until the work had been readied for performance; the composer's presence at rehearsals was an integral part of his contract. This is clearly set forth in Cavalli's contract of 24 July 1658 with Marco Faustini at S. Cassiano, which stipulates that, in addition to providing one opera per year for three years himself ("con la diligenza et Virtù sua propria"), he "be present at all the rehearsals that are needed, and also to change parts, alter, cut, and add whatever is necessary in the music in the service of the opera" (Appendix IIIA.2).^[35]

Although composers' scores are never as explicit as printed librettos in distinguishing between the stages of composition—between what took place before and what took place during rehearsals—they often record something of the composer's process, his methods of shaping and continually modifying his musical setting as he moved toward performance or even from one perfor-

mance to the next. Printed librettos, by their very fixity, help to shape speculation about the meanings of the scores to which they are connected. Differences between libretto and score might indicate that the score either represents a modification for a subsequent performance or an earlier stage of the libretto, which later was revised for publication by the poet; or again, the score might represent a later version that the poet could not or would not adopt for his published text.

These differences could, of course, have originated at the composition stage. Even during the early years of intimate collaboration, composers did not always set the poet's text exactly. We have already referred to important differences between Monteverdi's and Badoaro's versions of Il ritorno d'Ulisse . The scores of Monteverdi's L'incoronazione di Poppea , too, diverge considerably not only from Busenello's printed text and the several extant manuscript versions of the libretto but from the scenario as well.^[36]

A number of Cavalli's scores are quite different from their printed librettos. Strophic texts are set non-strophically, both as arias and recitatives, refrain lines are added, and, of course, passages of text omitted. Such differences, however, probably more reflect the temporal relationship between extant score and printed libretto than any aesthetic question. These discrepancies would probably have disappeared if the libretto could have been printed late enough to include all last-minute alterations. On the other hand, they also reveal the kinds of changes Cavalli and, presumably, other composers normally made as they worked from a manuscript libretto, changes that might subsequently be incorporated in printed librettos. In the case of librettos printed before the works were performed, we can usually assume that most differences from the score reflect composers' changes. But if they were printed afterwards, as in Busenello's Delle hore ociose , which contains all of his librettos, we must also consider that the librettist may have revised his text in the interim, and that his original text, or one closer to it, is represented in the music.^[37]

Since composers presumably had less of a stake in a precisely finished product than did librettists, and since it was part of their job to make continual adjustments and changes up to the last minute, many of the indignities librettists complained about were simply matters of course to them. Piecemeal composition, for example, which might have prevented librettists from editing or polishing their works to their own literary satisfaction, was more the rule than the exception for composers. They reported the receipt and dispatching of

single acts or groups of scenes of librettos and scores with apparent equanimity. Thus Ziani, writing to Faustini on 2 August 1665 about the ill-fated Doriclea (whose last-minute cancellation we have already discussed), reported that he had just sent him the music of act 1, but was awaiting the arrival of act 2 of the libretto so that he could begin to set it.^[38] And Cesti exhibited similar nonchalance when he wrote on 21 June 1665 to Beregan, whose Tito he had been contracted to set, urging him to send act 1 as soon as it was ready, without waiting until the entire work was completed.^[39] He made the same request again on 12 July. He must have received the text (and set it) by 2 August, when he reacted enthusiastically to the arrival of act 2 in another letter to Beregan: "I received the second act; it is beautiful, beautiful, beautiful."^[40]

Composers' piecemeal settings routinely found their way to the singers (or impresario) in the same fashion. Cavalli sent the first act of his Hipermestra to Florence ahead of the others.^[41] Cesti and Ziani both refer often to having sent off a single act or several scenes with the promise of the rest in due course. In a letter of 20 December 1666, Cesti writes that he has just mailed half of act 2 of Tito and hopes that act 1 has arrived, although it lacks the opening sinfonia ("la Sinfonia avanti s'alzi la tenda"), which he will write after having finished the opera "because it isn't so urgent at the moment" ("per non essere presentemente di tanta necessità")^[42] Although piecemeal setting may not have bothered composers, it seems (reassuringly) to have had a negative effect on some singers, who were frustrated in their attempt to understand their parts. As we shall see, Nicola Coresi complained that he had received the wrong impression of his role in Meraspe from the excerpt he had been sent originally;^[43] and Catterin'Angela Botteghi asked to see the whole libretto of the same opera so that she could understand better how to interpret her role.^[44]

Composers may have taken piecemeal composition for granted, but they were not altogether insensitive to the time pressures that plagued librettists. Ziani lost no opportunity of reminding Faustini how he had saved the day by writing Annibale in Capua in five (or six) days.^[45] Speed was his trademark, a characteristic both he and, apparently, his critics emphasized, though obviously

for opposite purposes. Ziani regarded his facility ("solita facilità") as an inimitable quality,^[46] which he invoked ironically in a request to Faustini for payment ("I write quickly but am paid slowly") as well as to counter criticism of his works: "If my arias are, have been, and will have been good, for the theater as well as for church, all of them pass under the same rush in the making."^[47] He imputed to the envy of his competitors the accusation of his having written Alciade while in transit from Innsbruck to Venice, yet he himself described some music he was sending to Faustini as "another piece of an opera completed by me in transit."^[48]

Ziani was especially famous (or infamous) for his speed, but other composers could work just as quickly when necessary—though they and their supporters did not hesitate to exaggerate such feats. Sartorio, for example, was reported to have set Bussani's Massenzio in thirteen days, although it probably took him closer to a month.^[49] Whereas Sartorio survived the pressure and even profited from it in enhanced reputation, not all composers were so resilient. One of them, Giovanni Antonio Boretti, reportedly even died from the anxieties of readying an opera for production: "Poor Gio Ant Borretti . . . died yesterday after two weeks of sickness caused by readying his opera for production" (Appendix IIIB.11c).^[50]

Although they were often at the mercy of circumstances beyond their control, composers, like their librettist colleagues, recognized the difference between working at leisure and under pressure. Cesti, for example, who complained about having to compose five acts of Il potno d'oro in six weeks, wrote

appreciatively to his librettist Beregan about being able to take his time over Tito , working at the relaxed pace of one act per month: "When one can enjoy the benefit of time, things come out much better."^[51] In the event, however, he did not finish the opera any too soon—in fact, not until 17 January, less than a month before the premiere at SS. Giovanni e Paolo.^[52]

The main source of trouble for composers, as for librettists, was the time it took to negotiate with the singers. It is clear that casts were often not fully assembled until well after the librettist and composer would have had to begin work. Librettists could still manage with a fairly general notion of the cast; some idea as to the number and relative importance of the singers would suffice for an initial draft of the text, which could be modified later. Giovanni Faustini, we remember, had sketched a number of librettos without knowing who would perform (or even set) them. It was unthinkable, however, for a composer to begin work before he knew at least the ranges of the voices for which he was writing. In the early days of the stable, continuous companies of Ferrari-Manelli, Cavalli-Faustini, and Sacrati, this was not a problem, and there were even periods of stability later in the century, when casts carried over from one season to the next.

But last-minute hiring and cast changes were increasingly the rule.^[53] Composing Alciade (1666), Ziani had to make several decisions on his own ("in order not to lose time") because Faustini had not informed him of the cast in time. He wrote the role of the boatman (Nocchiero ) for the bass singer (who otherwise had a very small part and could evidently double here) and Lerilda for a soprano (of which Faustini had several to choose from). And he asked for information on the voice for the "Vecchia" Clipea in act 3, not indicated by Faustini, having assigned it, in the interim, to the soprano who also played Idiotea—and who, he said, had not appeared on stage for a while (Appendix IIIA.10). Even in parts that had been cast, however, he was forced to make a number of alterations in response to last-minute substitutions or even careless assumptions: octave

transposition of one role (Guerra) from baritone to alto, and transposition a fourth or fifth lower to suit the alto performer of the messenger (Nuntio Messo ) in Doriclea , which he had originally scored for soprano.^[54]

Cesti, too, repeatedly requested information about the intended singers for his Tito , at one point asking specifically who would sing the hunter (Cacciator ) in the final scene of act 2, "because there is no indication."^[55] At another point he made a revealing suggestion of his own, that instead of two sopranos and a contralto, it would be better to have a lower voice, either tenor or baritone, "because in the most beautiful scenes diversity of voices seems much better, and if the scene between Domitiano and Berenice, as it is, scored for two sopranos, were sung by two different voice parts, it would stand out marvelously, and I am taking into account that with the two Amorini there will be nine sopranos."^[56]

The information requested by both Ziani and Cesti evidently concerned minor roles, many of which involved decisions about doubling. Clearly the major roles would have been assigned earlier. Normally, casts were decided upon several months before the beginning of a season—usually by August—with singers scheduled to arrive in Venice, parts learned, by mid November at the latest—that is, about six weeks before the premiere, when rehearsals began.^[57] The planning of special stage effects and the construction of machinery

probably began before the singers arrived—in October.^[58] But there were often troublesome exceptions. Long-distance negotiations with at least one singer, the much-sought-after castrato Ciecolino (Antonio Rivani), were still in progress on 17 November 1670, a mere month before the season was to begin.^[59] Two years later the same difficult singer did not arrive in Venice until late December, after the season had started, "tardi per recitare," but not too late to appear in at least some performances.^[60]

Although negotiations with singers often concluded surprisingly close to the beginning of the season, there was apparently considerable flexibility with respect to the actual date of opening night for a work. Traditionally, of course, the season coincided with Carnival, which began the day after Christmas, and most opera houses opened as close to that time as possible. We noted earlier how Marco Faustini tried to move the opening day of SS. Giovanni e Paolo forward by some two weeks in 1666 in order to seize the advantage from his competition.^[61] Later in the century, openings in early December or even late November were not uncommon, particularly in years when Carnival was unusually short.^[62] Theater managers could hardly be blamed for wanting to squeeze more performances into a season.

In any case, since the opening date was not strictly fixed, it could be postponed if the opera were not ready yet; if the problem concerned the second opera of the season, the first could continue until the second was ready, or else an opera from a previous season could be substituted at the last minute. All of these situations actually occurred. Composers were announced and changed (Argiope );^[63] a new libretto and setting, or just a new setting, commissioned at the last minute (Eliogabalo, Massenzio );^[64] or an old opera substituted (Orontea for Doriclea ).^[65]

Naturally the most intense pressure occurred during the rehearsal period, when a libretto, along with the score, took its final shape. Many operas were reportedly revised and readied for performance in a matter of days: twelve days proved adequate to prepare Annibale in Capua for performance in 1661 (twice as long as the five [or six] it took Ziani to write it, as he never tired of saying), ten for a revival of Cesti's Orontea in 1666, and a record eight for his Dori in 1667.^[66] Printed librettos and manuscript scores both reveal the kinds of changes that would have taken place during rehearsals, or else during the run of the work. The cast of Faustini's and Cavalli's Calisto , for example, gained a character, a crude peasant (Bifotco ), at the last minute (or at least after the libretto had been printed and the score copied). Bifolco appears in three scenes, twice alone and once with another character, in acts 2 and 3. Faustini refers to these scenes in a note to the reader as "superimposed on the original structure of the story to delight you,"^[67] and Cavalli's score accounts for the addition with the rubric:

"Qui và la Scena del Bifolco," though no music is provided.^[68] Conversely, the cast of Cavalli's and Minato's Artemisia was modified by reduction sometime after the music was written, but before the libretto was printed. The score, mostly autograph, has music for a character called Cleante, who does not appear in the libretto, and it actually illustrates how the reduction was achieved. In three places, music for Cleante has been crossed out and the text distributed to other characters in the opera.^[69] The fact that Cleante never made it into the printed libretto and that the score is one of Cavalli's composing scores rather than a fair copy suggests a relatively early date for the modification.^[70]

Dramatic coherence, what there was of it, was increasingly at the mercy of such exigencies. A curious extreme was surely reached in 1676 when Giulio Cesare Corradi, the librettist of Germanico sul Reno , claimed to have added a scene for an Orfeo, not listed in the cast and totally extraneous to the plot, in order to exhibit the talents not of a singer but of a special violinist.^[71] But once again, the incorporation of such late arrivals, motivated by the usual desire to appeal to the audience, exacted a toll principally on the librettists. The required changes were often substantial enough to interfere with whatever dramatic continuity or balance their original texts may have aspiredto. To composers the cost was negligible. Musical modifications were usually relatively minor and had little if any effect on the opera as a whole.

We return then to the social distinction between librettist and composer and their different claims, aspirations, and functions with respect to dramma per musica . Librettists were deeply concerned with the dramatic coherence and literary integrity of their work, as we have seen in the elaborate strategies they developed for preserving it or compensating for its absence. Composers could not possibly have felt the same way. Their scores could never have aspired to the objective wholeness of librettos; their music was doubly contingent: on their librettists and on their performers. Musical coherence independent of el-

ther was an anachronism and an impossibility. Indeed, once the formerly unified activity of producing an opera had disintegrated into an amalgam of specialized tasks, the subservience of the composer in the operatic hierarchy became still clearer.

The kinds of alterations composers were called upon to make during rehearsals or in preparation for revivals, to adapt their works to new performers or to the requirements of a new theater or audience, remained central to their profession. But because of institutionalization and the consequent logistical problems raised by geographical separation and competing patronage, these alterations could not always be made by the original composers. Cesti, for example, instructed Beregan to arrange for Rovettino to fulfill his rehearsal obligations—namely, "to cut, add, change, or do whatever else was necessary in the music"—since he could not come to Venice for the premiere of his Tito , owing to commitments to his employer in Vienna.^[72] And Ziani, although he referred on several occasions to his obligatory presence in Venice for the rehearsals of his work, assuring Faustini at one point that he would not "leave Venice until the whole score has been furnished and accommodated,"^[73] later gave him rather grudging permission to engage someone else for the task.^[74]

Given the nature and extent of some of the changes that took place at rehearsals—new arias, cuts, transpositions, rearrangements, even whole new roles—which might have been supplied by someone other than the original composer, many of these works would certainly qualify as pasticcios. But they were not the first such cooperative ventures. Multiple authorship may not have become a necessity until the pressures of institutionalization made it so, but it had been a reality, even a positive feature considered worthy of advertising, well before then.

Some early Venetian operas were specifically designed and promoted as pasticcios. La finta savia (1643), for example, proudly boasted the music of four different composers: Laurenzi, Crivelli, Merula, and Ferrari (Appendix I.18b).^[75] Many of the operas we have already mentioned ended up as pasticcios, either by design or by accident. Argiope (1649), as we have seen, was touted as

a pasticcio manqué;^[76] Faustini's revival of Cesti's Orontea became a pasticcio in effect when it was prefaced by Ziani's new setting of the prologue to Doriclea .^[77] And Cavalli's Giasone became part of an overt pasticcio when it appeared in Rome as Il novello Giasone in 1671 with some music by Stradella. Indeed, nearly every work that was revived, either in Venice or elsewhere, in the original composer's absence, or after his death, was a pasticcio. The most famous of these, at least in the versions that have come down to us, is Monteverdi's L'incoronazione di Poppea , the only modern "classic" of Venetian opera. Both extant scores, which probably reflect the Naples revival of 1651, contain music that Monteverdi could not have written. The only disagreement in this case is over the amount of original music still remaining.^[78]

But that is not a meaningful question for operatic music of this period. Unlike that of the libretto, the sanctity of the operatic score was a concept virtually unknown and irrelevant in seicento Venice. Although it is difficult to judge from our distance, the replacement of one composer by another does not seem to have created a problem of stylistic incongruity either for contemporary audiences or for the composers themselves. Operas were increasingly bound by musico-dramatic conventions, making substitution of composers relatively easy.^[79]

Because their professional stature and livelihood depended on it, librettists emphatically claimed and numbered their texts, carefully keeping track of their careers and burnishing their reputations. Composers, however, as they continued, often anonymously, to fulfill their multiple functions, increasingly disappeared behind their creations, leaving the singers to represent them.