¹ Heartz, Daniel, ‘Constructing Le nozze di Figaro’, Journal of the Royal Musical Association, 112 (1987), 90–1, 93CrossRefGoogle Scholar; repr. in Heartz, Mozart's Operas, ed. with contributing essays by Thomas Bauman (Berkeley, 1990), 147, 148 (subsequent references will give page-numbers from the latter); Carter, Tim, Mozart, W. A. ‘Le nozze di Figav’, Cambridge Opera Handbooks (Cambridge, 1987), 119.Google Scholar

² This article presents in revised form a paper read at the National Meeting of the American Musicological Society in Minneapolis, Minn., in October 1994. It is part of a larger study, entitled ‘Tonal Organization in the Opera Buffa of Mozart's Time’, to appear in Mozart Studies 2, ed. Cliff Eisen (Oxford, 1996).Google Scholar

³ Webster, James, ‘Mozart's Operas and the Myth of Musical Unity’, this journal, 2 (1990), 200Google Scholar. A number of the issues (and specific examples) to be considered here have also been discussed in Webster's article. The importance of Lorenz (and Wagner himself) for Mozart operatic analysis is treated at greater length in Carolyn Abbate and Roger Parker, ‘Dismembering Mozart’(ibid., 187–95). Lorenz, also wrote on Mozart: see his ‘Das Finale in Mozarts Meisteropern’, Die Musik, 19 (1926–1927), 621–32.Google Scholar

⁴ Some writers, like Gerald Abraham, are content to explain the unity simply by reference to the keynote (the key shared by the overture and the last finale). ‘Whatever the genre to which Mozart's operas belong, he conceived them as unified musical compositions. … The unifying element in every case is tonality each opera ends in the key of the overture, however remote or complicated the key-system of the intervening numbers.’ (Abraham, , ‘The Operas’, in The Mozart Companion, ed. Landon, H. C. Robbins and Mitchell, Donald [New York, 1956], 291)Google Scholar. and Rosen, Charles, in The Classical Style (New York, 1971), 304–5Google Scholar, makes the same point in much the same way. For these scholars, at least, the keynote seems to be a sufficient condition of tonal unity in an opera.

⁵ Webster, ‘Mozart's Operas’; Abbate Parker, ‘Dismembering Mozart’.

⁶ The problem of ‘unity’ is complicated by the fact that writers who invoke this quality do not explain what they mean by it. I view unity in an eighteenth-century opera to lie not in any sort of uniformity – clearly the successive musical numbers of an opera are different from one another in all sorts of obvious ways – but in the sense of ‘an arrangement of parts or material that will produce a single harmonious design or effect’ (from Webster's New World Dictionay of the American Language, Second College Edition [New York, 1972], s.v. ‘Unity’, 6a)Google Scholar. This definition may be seen to be fulfilled by a reasonably coherent dramatic narrative, with or without music.

⁷ ‘For example, Hunter, Mary, in The Poetics of Entertainment Opera Buffa in Vienna 1770–1790 (forthcoming) cites a poster for a Burgtheater production (probably of 1778) that advertises card-tables for rent (Ch. 1)Google Scholar. And many contemporary reports mention – sometimes with amusement or horror – how little attention audiences paid most of the time to on-stage events.

⁸ While the concept of ‘authorial intention’is a complex one, and the perils of the ‘intentional fallacy’are never far off, it should none the less be possible at least in principle to talk quite simply about a composer's intentions. We may accept, in general terms, that a work of art has many characteristics. Surely some of these were chosen quite consciously by its creator, others chosen perhaps without conscious thought, and still others occur without being what we would call ‘chosen’at all. Whatever the difficulties of assigning any particular characteristic to one of these categories, we can at least acknowledge that the separate categories exist.

⁹ By this I mean that a listener with perfect pitch could have observed that, say, Susanna's ‘Deh vieni non tardar’ in Act IV of Firgaro was in the same key as Figaro's ‘Se vuol ballare’ (Act I). But it does not necessarily follow that such a listener did observe this, or derive any meaning from it.

¹⁰ As I have said, however, what listeners perceived (and perceive today) varies; in particular, for modern scholars and listeners analysis may make audible relationships that were not initially perceptible in the theatre.

¹¹ Levarie, , ‘Viewpoint: On Key Relations in Opera’, 19th-Century Music, 3 (1979–1980), 88–9Google Scholar. His piece is a response to an earlier ‘Viewpoint’by Kerman, Joseph in 19th-Century Music, 2 (1978–1979), 186–91CrossRefGoogle Scholar, that in turn responded to Levarie's original article, ‘Key Relations in Verdi's Un Ballo in Maschera’, ibid., 143–7.

¹² This may be seen both in ‘Key Relations in Un Ballo’ and in Levarie's book, Mozart's ‘Le Nozze di Figaro’: A Critical Analysis (Chicago, 1952)Google Scholar. See, for example, the argument in the former that there exists across the three acts of Ballo ‘the cadential progression A flat–A–B flat [which] can thus be understood as a well-balanced widened full cadence: second subdominant to (substitute) second dominant to tonic’(p. 144). Likewise in the latter, Levarie argues for a cadential progression of D–E flat–A–D across the four acts of the opera, despite his own demonstration that the third act is in C, with the opening A minor/major duet being the relative minor (pp. 233–45).

¹³ We may anticipate the objection that aspects of an opera not consciously perceived by listeners may still affect them at some subconscious level. But this potentially powerful argument has never been put forward in a convincing or systematic fashion. At least in principle, features of a musical work are either perceptible or not perceptible. One needs to make the case for any particular tonal relationship (or other feature) that it can be perceived by an audience.

¹⁴ A list of these works may be found in Platoff ‘Tonal Organization’(see n. 2), Appendix 1. Collectively they represent 36 per cent of the operas performed at the Burgtheater in these years (28 of 78). But because nearly all of the most popular operas of the decade are included, these operas account for 52 per cent of the operatic performances (626 of 1203). The performance data comes from Michtner, Otto, Das alte Burtheater als Opernbühne: Von der Einfuhning des deutschen Singspiels (1778) his zum Tad Kaiser Leopolds II. (1792) (Vienna, 1970), 473–511.Google Scholar

¹⁵ Heartz, , Mozart's Operas (see n. 1), 140.Google Scholar

¹⁶ In Don Giovanni the evidence for some degree of planning with respect to the uses of D major and minor is more compelling. See Rushton, Julian, Mozart, W. A.: ‘Don Giovanni’, Cambridge Opera Handbooks (Cambridge, 1981)Google Scholar Platoff, ‘Tonal Organization’(n. 2).

¹⁷ From Über das Leben and die Werke des Anton Salieri, k.k. Hofkapellmeisters (Vienna, 1827), 30–2.Google Scholar A passage is presented in Heartz, Mozart's Operas (see n. 1), 154–5 this quotation (translated by Heartz) is on p. 139.

¹⁸ See Steblin, Rita, A History of Key Characteristics in the Eighteenth and Early Nineteenth Centuries (Ann Arbor, Mich., 1983).Google Scholar

¹⁹ Rushton, , The New Grove Dictionary of Opera, ed. Sadie, Stanley, 4 vols. (London, 1992), s.v. ‘Mozart’, iii, 495.Google Scholar

²⁰ Only two characters – Donna Elvira and Guglielmo – have two arias in the same key in both cases this arose from changes in the score. Of course, it was Mozart who made these changes, suggesting that the avoidance of two arias in the same key for a particular character was preferable, but not an overwhelmingly important priority.

²¹ Carter, , ‘Le nozze di Figaro’ (see n. 1), 119.Google Scholar

²² There are nine such tonally closed acts out of 59. Since composers used seven tonic keys with roughly equal frequency (for details see Platoff, ‘Tonal Organization’), by pure chance the number of acts beginning and ending in the same key should be one-seventh of the total, or 8.4.

²³ Heartz, , Mozart's Operas (see n. 1), 140–2.Google Scholar

²⁴ Rushton, Julian, in a review of Andrew Steptoe's The Moart-Da Ponte Operas (Oxford, 1988)Google Scholar, Music & Letters, 70 (1989), 544.Google Scholar

²⁵ Webster, , ‘Mozart's Operas’(see n. 3), 208Google Scholar.

²⁶ Platoff, John, ‘How Original was Mozart? Evidence from opera buffa’, Early Music, 20 (1992), 107–9.Google Scholar

²⁷ There are twelve possible key-successions, since the tonic of a number can be the same as that of the preceding number or from one to eleven semitones higher (if one normalises the successions by octave transpositions). If the twelve possibilities occurred with equal frequency, then each would occur 8.25 per cent of the time. But the descending fifth (or rising fourth) is by far the most common succession: its frequency in the Viennese repertory I examined is 16.7 per cent, while the three next most common successions are the rising major second (10.6 per cent), the descending minor third (10.3 per cent) and the rising fifth (9.5 per cent). In Mozart's Da Ponte operas the descending fifth occurs 17.3 per cent of the time, followed by the rising minor third (16 per cent), the rising fifth (14.8 per cent) and the descending minor third (12.3 per cent). In short, while there are twelve possible key-successions, both Mozart and his contemporaries employed a single one of them – the descending fifth – more than one-sixth of the time.

²⁸ Here my figures are based on all of Mozart's operas. The descending-fifth succession leads to 18.2 per cent of Mozart's act-ending numbers, compared to the rising major third (15.9 per cent) and the rising fifth and descending minor third (both 11.3 per cent).

²⁹ By ‘introduzione’I simply mean the first vocal number following the overture.

³⁰ Abert's discussion, with additional remarks on Acts III and IV, is found in the preface to the Eulenberg full score of Mozart's, Le nozze di Fizgaro (London, n.d.), xiv–xvii.Google Scholar

³¹ These categories properly apply only to major keys but in fact pieces in minor are quite rare in the Viennese repertory: only 18 musical numbers out of 706 use a minor tonic wholly or in large part.

³² Platoff, John, ‘Tonal Organization in “buffo” Finales and the Act II Finale of “Le nozze di Figaro”’, Music & Letters, 72 (1991), 390–3.CrossRefGoogle Scholar

³³ For instance, in Act III of Figaro the Countess's ‘Dove sono’has little to do with the preceding number, the sextet, yet the key succession is a fifth, F–C. Conversely the first five numbers of Cosí fan tutte are related by descending thirds (G–E–C–A–f), though the first three are a discrete scene for the men, the fourth is a duet for the sisters alone following a change of location, and the fifth marks the beginning of the intrigue with Don Alfonso's false report that the men have been called away to the army.

³⁴ My thanks to Mary Hunter, Dan Lloyd, David Mauro, Leonard B. Meyer and Julian Rushton for many valuable suggestions.