1 Paul Everett and Michael Talbot, ‘Homage to a French King: Two Serenatas by Vivaldi (Venice, 1725 and ca. 1726)‘, Antonio Vivaldi: Due serenate, facsimile edition, Drammaturgia musicale veneta, 15 (Milan, 1995), ix–lxxxvii (p. xlv); Michael Talbot, The Sacred Vocal Music of Antonio Vivaldi (Florence, 1995), 443–5.Google Scholar

2 This continuing edition, of which the first volumes appeared in 1982, is published by BMG Ricordi, Milan, for the Istituto Italiano Antonio Vivaldi. With very few exceptions, the 37 volumes containing cantatas are edited by Francesco Degrada.Google Scholar

3 The most recent catalogue to list the cantatas is Peter Ryom, Verzeichnis der Werke Antonio Vivaldis: Kleine Ausgabe (2nd edn, Leipzig, 1979). This catalogue gives first-movement incipits for all the works, but few supplementary details. Prendea con man di latte, RV 753, discussed in Colin Timms, ‘Prendea con man di latte: A Vivaldi Spuriosity?‘, Informazioni e studi vivaldiani, 6 (1985), 64–73, is also to be considered spurious on account of stylistic incompatibility.Google Scholar

4 On RV 796, see Heller, Karl, ‘Zu einigen Incerta im Werkbestand Vivaldis’, Vivaldi vero e falso: Problemi di attribuzione, ed. Antonio Fanna and Michael Talbot (Florence, 1992), 43–57 (pp. 52–7); on RV 799, see Olivier Fourés and Michael Talbot, ‘A New Vivaldi Cantata in Vienna’, Informazioni e studi vivaldiani, 21 (2000), 99–108.Google Scholar

5 Cantata plans ending with R are less common (they are wholly absent in Vivaldi). When they do occur, the last line or two of the recitative verse is most often set as a cavata (i.e. an aria-like section ‘extracted’ from verse ostensibly intended for recitative) in order to give the work a more lyrical and substantial conclusion. By this means, the recitative is allowed to perform its customary function of introduction – only in this instance the ‘aria’ is generated artificially from within. On cavatas in general, see Timms, Colin, ‘The Cavata at the Time of Vivaldi’, Nuovi studi vivaldiani: Edizione e cronologia critica delle opere, ed. Antonio Fanna and Giovanni Morelli (Florence, 1988), 451–77.Google Scholar

6 Affò, Ireneo, Dizionario precettivo, critico ed istorico della poesia volgare (Parma, 1777), 114. ‘Simple’ cantatas are very rare indeed. ‘Triple’ cantatas (RARARA or ARARA) are not uncommon in the late seventeenth century, but pass rapidly out of fashion subsequently as the average length of arias increases. By the middle of the eighteenth century, as we learn from a letter of 14 February 1755 from Pietro Metastasio to Raniero de' Calzabigi, solo cantatas with more than three arias had become a practical impossibility: tiring to the singer and tedious to the audience.Google Scholar

7 Quoted in Grout, Donald Jay, A Short History of Opera (2nd edn, New York, 1965), 187.Google Scholar

8 The reduction from RARA to ARA, which gathered force as the eighteenth century progressed, can be viewed as the counterpart in vocal chamber music of the reduction from SFSF (Slow–Fast–Slow–Fast) to FSF (Fast–Slow–Fast) that occurred at the same time in the sonata and even earlier in the concerto. There is no attempt here to equate ‘R’ with ‘S’ or ‘A’ with ‘F’; the significance of the comparison is that in each instance the structure becomes more compact and acquires an over-arching symmetry.Google Scholar

9 I shall employ throughout the convention of using upper-case roman numerals for major keys, lower-case roman numerals for minor ones.Google Scholar

10 This composer's cantatas are discussed in Michael Talbot, Benedetto Vinaccesi: A Musician in Brescia and Venice in the Age of Corelli (Oxford, 1994), 156–87.Google Scholar

11 Because recitatives commonly exist in a state of constant tonal flux, to describe their key structure can be problematic. In Table 1 and later tables I have identified their opening and closing chords. Where a superscript ‘V’ follows a pitch, I am referring to its dominant or dominant seventh chord (which is always major). A concluding tierce de Picardie is interpreted at face value, i.e. as a statement of major tonality.Google Scholar

12 Paris, Bibliothèque Nationale de France, fonds du Conservatoire, Rés. 1451.Google Scholar

13 In recitatives the option to return finally to the key of the first chord is not often taken in cantatas of Vivaldi's time and occurs mostly in opening recitatives. The relative frequency of tonally closed recitatives in Vinaccesi's cantatas (Table 1 has four examples, all of them first movements) probably reflects the composer's wish to stabilize the tonality of the work at its outset (as one might, for instance, in the slow opening movement of a sonata).Google Scholar

14 Michael Talbot, ‘Modal Shifts in the Sonatas of Domenico Scarlatti’, Chigiana, 40, n.s. 20 (1985), 25–43. One might indeed argue that baroque recitative was the cradle of radical techniques of modulation that did not find general application until the development sections and transitions of the Classical age.Google Scholar

15 There is less opportunity for violent modulations of this kind in cantatas than in operas, where the fluctuations of mood are more extreme, the recitatives tend to be longer and the tonal ‘distance’ between successive arias is potentially greater. A good example of the technique occurs in bars 7–8 of Vagaus's recitative ‘Iam non procul ab axe’ from Vivaldi‘s Juditha triumphans (shown as a music example in Talbot, The Sacred Vocal Music of Antonio Vivaldi, 439–40), where a root-position A minor chord preceded by its dominant is followed by a major chord of C# in first inversion acting as the dominant of F# minor. The elided chord is that of A major. This harmonic wrench arrives at the moment when Vagaus catches sight of his beheaded master Holofernes.Google Scholar

16 This is a principle breached – famously – by Handel on many occasions, but he is almost alone among the composers of baroque cantatas known to me.Google Scholar

17 For a detailed description of this process, see Paul Everett and Michael Talbot, ‘Homage to a French King’, xxi–xxxix.Google Scholar

18 Arie di baule, translatable idiomatically as ‘suitcase arias’, were favourite arias of singers inserted (or substituted for existing arias), often without textual modification, in productions of operas in which they sang. The metaphor refers to the trunks into which itinerant singers packed their belongings when travelling between cities.Google Scholar

19 The title-page is lost, but the entry for the composer in Johann Gottfried Walther's Musicalisches Lexicon oder musicalische Bibliothec (Leipzig, 1732) and the running titles acting as gathering signatures confirm the title, at least in its short form. The set has been published in an edition by the present author as vol. 31 of the series Recent Researches in the Music of the Baroque Era (A–R Editions, Madison, 1979).Google Scholar

20 Charles Burney, A General History of Music from the Earliest Ages to the Present Period (1776–89), ed. Frank Mercer, 2 vols. (London, 1935), ii, 637. Burney, knowing Vivaldi only from his instrumental music, was surprised to discover that these cantatas (and others that he perused elsewhere) were ‘very common and quiet’. No trace of them survives.Google Scholar

21 RV 649, 652, 653, 654, 658, 659, 663, 665, 680, 685, 686 and 799. Many different factors, bibliographical, textual and notational, have been used by investigators to establish the Mantuan connection of these works.Google Scholar

22 RV 655, 656, 657, 662, 663, 674, 676, 677, 678 and 796.Google Scholar

23 These key relationships between movements are highly unusual for the time.Google Scholar

24 I have counted 35 examples.Google Scholar

25 It is found in RV 28, 96, 105, 154, 235, 243, 320, 321, 326, 329, 498 and 757.Google Scholar

26 Strangely enough, the concertos and sonatas never use G minor for the internal slow movement of a work in E♭ major, although the reverse relationship is often encountered.Google Scholar

27 See, for example, the concertos RV 287, 294, 489, 490, 765/767 and 775.Google Scholar

28 This type of cadence is equated specifically with a question mark in the typology of recitative cadences offered in Johann Joseph Fux, Gradus ad Parnassum (Vienna, 1725), 277–8.Google Scholar

29 Note, however, the move to the relative major key following a Phrygian cadence in RV 679. This Corellian topos is surprisingly rare in Vivaldi's music.Google Scholar

30 Like most of his Italian contemporaries, Vivaldi figures his cantatas presented in score very lightly in the expectation that the accompanist will interpret most of the harmonies correctly by reference to the vocal line and bass. The absence of a flat over (or under) the final F is uncertain in its significance. Does it mean that the bass note should be harmonized diatonically, i.e. with A# – or has it been omitted as superfluous because of the unambiguously ‘minor’ character of the final vocal phrase?Google Scholar

31 It is noteworthy that even when Vivaldi borrows mere ‘snippets’ from earlier works, they are usually introduced in their original key.Google Scholar

32 Bella Brover-Lubovsky, ‘Between Modality and Tonality: Vivaldi's Harmony’, Informazioni e studi vivaldiani, 21 (2000), 95–108. Brover-Lubovsky considers Vivaldi's variable choice of key signatures (modern, ‘Dorian’, ‘Mixolydian’, etc.) important in this connection. I am not so sure of their relevance, but I share her perception that his musical language was in many respects unexpectedly conservative.Google Scholar

33 The juxtaposition of a major key and the minor key a tone lower is occasionally found at points of structural division within Vivaldi's instrumental movements.Google Scholar

34 Vivaldi's contemporaries would have understood by the word ‘movement’ (movimento, moto) not a structural component as such but a section or passage in a single tempo and metre. Vivaldi would probably have referred to the three components of an ARA cantata simply as ‘parts’.Google Scholar

35 For more on Vivaldi's unusual tonal plans, see Talbot, Michael, Vivaldi (2nd edn, London, 1993), 86.Google Scholar

36 I am grateful to Colin Timms for reading this article in draft and providing some very useful comments on it.Google Scholar